A series of tripodal cysteine derivatives as water-soluble chelators that are highly selective for copper(I)

Article abstract of DOI:10.1002/chem.201003613

A series of tripodal ligands derived from nitrilotriacetic acid and extended by three converging, metal-binding, cysteine chains was synthesised. Their ability to bind soft metal ions thanks to their three thiolate functions was investigated by means of c

Full text of DOI:10.1002/chem.201003613

DOI: 10.1002/chem.201003613
A Series of Tripodal Cysteine Derivatives as Water-Soluble Chelators that are
Highly Selective for Copper(I)
Anaꢀs M. Pujol, Christelle Gateau, Colette Lebrun, and Pascale Delangle*[a]
Abstract: A series of tripodal ligands
derived from nitrilotriacetic acid and
extended by three converging, metal-
binding, cysteine chains was synthe-
nuclear copper(I) complexes of low
stability. In contrast, the ester and
amide derivatives L1 and L2 are effi-
cient CuI chelators with very high affin-
ities, close to that reported for the
amide functions in the upper cavity of
the tripodal ligands stabilises these
mononuclear complexes and was evi-
denced by the very low chemical-shift
temperature coefficient of the secon-
dary amide protons. Moreover, L1 and
L2 display large selectivities for the tar-
geted metal ion that is, CuI, with re-
spect to bioavailable ZnII. Therefore
the two sulfur-based tripods L1 and L2
are of potential interest for intracellu-
lar copper detoxication in vivo, without
altering the homeostasis of the essen-
tial metal ion ZnII.
ACHTUNGTRENNUNGsised. Their ability to bind soft metal
ions thanks to their three thiolate func-
tions was investigated by means of
complementary analytical and spectro-
scopic methods. Three ligands that
differ by the nature of the carbonyl
group next to the coordinating thiolate
functions were studied: L1 (ester), L2
(amide) and L3 (carboxylate). The neg-
atively charged derivative L3, which
bears three carboxylate functions close
to the metal binding site, gives poly-
metal-sequestering
(logKꢀ19). Interestingly, these two
ligands form mononuclear copper com-
metallothioneins
AHCTUNGTRENNUNG
plexes with a unique MS3 coordination
in water solution. An intramolecular
hydrogen-bond network involving the
Keywords: bioinorganic chemistry ·
chelating agents · copper · ligands ·
selectivity
Introduction
neurodegenerative diseases like Alzheimerꢀs disease and
suspected to cause Ab precipitation and toxicity.[9,10] Chela-
tion therapy[8,9,11] is currently used or proposed for treating
these disorders and intoxication, therefore it is of major in-
terest to develop molecules able to efficiently and selective-
ly bind copper. In particular, intracellular copper chelation
would be an efficient tool to remove metal ions from organs
where it is accumulated.[12]
Metallic elements play major roles in biochemistry. The es-
sential transition-metal ions are used by cells in structurally
constrained binding sites in metalloproteins, in which they
can carry out structural, regulatory or catalytic roles. As
these metal ions can also catalyse cytotoxic reactions, sever-
al families of proteins are present in cells to control their
concentration and to confine them to vital roles.[1–4]
As the cytoplasm of most eukaryotic cells is a reducing
environment, the predominant oxidation state of copper in
cells is CuI,[7] which has a soft character and thus a high af-
finity for soft donors like thiolates. This preference for soft
sulfur ligands is exemplified in proteins involved in copper
homeostasis,[3] which mainly bind these ions with several thi-
olates of cysteine side chains. Many intracellular Cu trans-
porters contain a conserved N-terminal MxCxxC sequence
that binds metal ions with two cysteines.[13] In a previous
report, the model cyclodecapeptide PC, c(MTCGSCSRP),
incorporating the binding sequence (MTCSGC) of the yeast
copper chaperone Atx1 was found highly selective for CuI
and HgII with respect to ZnII, another essential metal ion
found in cells.[14,15] Therefore, a glycopeptide also containing
two cysteines and targeted at the hepatocytes was recently
designed and demonstrated to be able to chelate intracellu-
lar copper. This compound may be a good candidate to fight
copper overload in the liver.[12]
Among essential metallic elements, copper is used as co-
factor in many redox proteins involved in several vital pro-
cesses. Free copper can also promote Fenton-like reactions
and would thus be very toxic even at low concentration.
Therefore intracellular copper concentration needs to be
rigorously controlled so that it is only provided to the essen-
tial enzymes, but does not accumulate to toxic levels.[3–6]
Wilsonꢀs disease is one of the major genetic disorder of
copper metabolism in humans. Impairment of copper trans-
port in hepatocytes results in cytosolic copper accumulation
with associated cellular injury.[7,8] Copper is also involved in
[a] Dr. A. M. Pujol, Dr. C. Gateau, C. Lebrun, Dr. P. Delangle
INAC
Service de Chimie Inorganique et Biologique (UMR E3 CEA UJF)
Commissariat ꢁ l’Energie Atomique, 17 rue des martyrs
38054 Grenoble cedex (France)
Fax : (+33)438785090
Metallothioneins (MT) are other proteins that sequester
many metals, in particular Cu and Zn with a higher affinity
for CuI. When copper is in excess of physiological require-
E-mail: pascale.delangle@cea.fr
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201003613.
4418
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4418 – 4428

FULL PAPER
ments, cells induce the biosynthesis of these small cysteine-
rich proteins, which form metal clusters.[16,17] Indeed, MTs
interact very efficiently with CuI in thiolate-rich environ-
ments and the crystal structure of the yeast copper thionein
(Cu8-MT) shows that two copper ions are digonally coordi-
nated to cysteine residues, whereas the six other copper ions
are trigonally bound, in a Cu8-thiolate cluster.[18] This sug-
gests that the trithiolato trigonal coordination affords stable
complexes of CuI.
To design efficient and selective soft metal ion chelators,
we take advantage of the high affinity of cysteine sulfur
donors for CuI, evidenced in proteins trafficking or seques-
tering this metal in cells.[12,14,15,19] As trithiolato coordination
environments afford very stable CuI complexes in metallo-
thioneins, ligands with three cysteine residues that promote
a MS3 geometry are very attractive. Therefore, three cys-
teine moieties can be attached with peptide bonds to non-
biological scaffolds to get pseudopeptide ligands.[20] Indeed,
chemical scaffolds can serve as platforms for the design of
podands that have all their binding arms oriented in the
same direction to coordinate metal ions.[21] We have used
such a strategy to obtain polydentate metal complexing mol-
ecules by appending several chelating functions in the same
chemical architecture.[22] Polyaminocarboxylates provide a
series of chemical scaffolds with a range of carboxylic acids
numbers that can be easily functionalised with cysteines.
Following this strategy, we have recently demonstrated that
a first tripodal cysteine-based ligand, namely L1, anchored
on a nitrilotriacetic acid (NTA) moiety, chelates very effi-
ciently CuI and HgII.[19]
that the nature of the carbonyl
function next to the thiols influ-
ences the complexation proper-
ties. The ester and amide deriv-
atives L1 and L2 are efficient
CuI chelators with very high af-
finities. The formation of mono-
nuclear copper complexes with
a unique MS3 coordination in
water is especially attractive.
Moreover these ligands display
large selectivities for the target-
ed metal ion that is, CuI, with
respect to bioavailable ZnII.
Therefore these two compounds are of potential interest for
metal detoxication in vivo, without altering the homeostasis
of the essential metal ion ZnII.
Results
Syntheses and characterisation of the three ligands: The syn-
thetic procedures of the ligands L1–3 are summarised in
Scheme 1. The three pseudopeptides were synthesised from
nitrilotriacetic acid (NTA). The first step is the coupling of
the free amine group of the S-protected cysteine derivatives
H-CysACTHNUGTRNENUG(Trt)-OEt or H-CysCAHTUNGTRNEN(UGN Trt)-NH2 to the carboxylic acid
functions of NTA in presence of classical peptide-synthesis
coupling agents to afford the tripodal precursors 1 and 2 in
good yields. Then the thiol groups were deprotected under
acidic conditions to give L1 and L2. The ligand L3 was ob-
tained in two steps from compound 1, the ester functions of
which were first hydrolysed into acids with lithium hydrox-
ide. Then the thiol groups were deprotected under acidic
conditions to afford L3 as a white solid. The ligands L1–3
were purified by RP18 HPLC and obtained with overall
yields from NTA of 40, 34 and 60%, respectively. These
Here we report on a series of tripodal ligands extended
by three converging metal-binding cysteine chains. To test
the effect of the carbonyl function near to the coordinating
sulfur group, we synthesised the three derivatives L1–3 shown
here. The coordination properties of these ligands were in-
vestigated by means of complementary analytical and spec-
troscopic methods. The results presented in this article show
Scheme 1. Syntheses of the three ligands L1–3
.
Chem. Eur. J. 2011, 17, 4418 – 4428
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4419

P. Delangle et al.
thiol-containing compounds are oxygen-sensitive and were
stored and manipulated in a glove box in an argon atmo-
ACHTUNGTRENNUNG
periments with this cation were conducted in the presence
of acetonitrile, which associates with CuI and overcomes the
disproportionation reaction.[25] Binding of CuI to L1–3 was
first monitored by UV spectroscopy. The addition of ali-
quots of tetrakis(acetonitrile) copper (I) hexafluorophos-
phate dissolved in acetonitrile over a peptide solution in a
9:1 (v/v) mixture of phosphate buffer (20 mm, pH 7.4) and
acetonitrile displays the appearance of a band centred at
267 nm that linearly increases with increasing CuI concentra-
tion up to two equivalents. This band is characteristic of
charge-transfer transitions (LMCT) of thiolate–CuI bonds,
as well as for divalent d10 metal ions like HgII or ZnII.[26,27]
The extinction coefficient of this LMCT bands are in the
range expected for the coordination of two CuI ions, found
in MT of ꢀ7000 per Cu bound.[26]
cate the presence of a single C3-symmetric species in which
ꢁ
the three arms are equivalent. The apical N CH2 methylene
protons give well-resolved AB doublets and the b-protons
of cysteines characteristic ABX systems.
Potentiometric studies were conducted in KCl 0.1m at
298 K to determine the protonation constants of the thiol
groups in the ligands. The protonation constants, KiH, are de-
fined in Equation (1) and could be obtained from the titra-
tions of the compounds with KOH and HCl. The results are
listed in Table 1.
Table 1. Protonation constants of the ligands L1–3 from potentiometric
measurements in water KCl 0.1m at 298 K.
Titrations with the ester and amide derivatives L1 and L2
give very similar results and the example of L2 is presented
in Figure 1. These titrations suggest the formation of two
Ligand
L1
L2
L3
pKa1
pKa2
pKa3
pKa4
pKa5
pKa6
9.4
9.0
8.1
<2.8
–
–
10.2
9.2
8.5
2.8
–
10.6
10.0
9.2
3.8
3.3
–
2.8
LHiꢁ1 þ H ! LHi
ð1Þ
KiH ¼ ½LHiꢂ=½LHiꢁ1ꢂ½Hꢂ
pKai ¼ log KiH
The pH titrations of L1 and L2 are indicative of three
weak acidic sites (Figure S1 in the Supporting Information),
corresponding to the deprotonation of the thiol functions
(pKa =8–10). A fourth protonation constant is necessary to
fit the titration of L2 and corresponds to the protonation of
the apical nitrogen at low pH. The fourth pKa of L1 could
not be fitted, probably because it is lower than the pH range
explored in the titration experiments (pH>2.8). The un-
Figure 1. UV titration of L2 (ꢀ50 mm) with CuI at pH 7.4 (20 mm phos-
phate buffer/CH3CN (v/v=9:1)). Spectra shown are difference spectra
[De=eACHTNUGRTNEUNG
(CuL2) - e(L2)] and correspond to samples with 0–3 equivalents of
CuI per L2.
ACHTUNGTRENNUNGusually low value of the apical amine pKa is consistent with
different types of complexes: a mononuclear complex when
less than one Cu equivalent is added and polynuclear com-
plexes when more than one Cu equivalent is added. Indeed
up to the addition of one equivalent of Cu, a high-energy
band develops with an absorption maximum below 230 nm.
No low-energy band, attributed to cluster-centred transi-
tions[26,27] around 340 nm is detected. CD titrations also evi-
dence the formation of a unique complex when less than
one equivalent of Cu is added, with intense dichroic bands
with positive and negative maxima (see Table 2), and no CD
signal characteristic of cluster-centred transitions. Besides,
three isodichroic points are evidenced in the CD titrations
between the addition of 0 and one equivalents of copper, as
seen in Figure 2 (left) demonstrating that the free ligand is
in equilibrium with a unique copper species. The mono-
the value published for another tripodal compound with an
apical nitrogen atom substituted with three amide groups
(N-(CH2CONH2)3, pKa =2.6).[23] As expected, the pH titra-
tion of L3 is indicative of six weak acidic sites. The three
highest pKaꢀs are in the range of thiol deprotonation, where-
as the three lowest are characteristic of carboxylic acid de-
protonations.
The values of the thiol protonation constants indicate that
the acid ligand L3 (ꢀpKaSH =29.8) is more basic than the
amide derivative L2 (ꢀpKaSH =27.9), which is also more basic
than the ester derivative L1 (ꢀpKaSH =26.5). This evolution
reflects the electron-withdrawing characters of the carbonyl
functions close to the thiol, which range in the order ester>
amide>carboxylate. The same tendencies were already ob-
served for small peptides containing cysteines.[24]
AHCTUNGTRENNUNG
nuclear complexes Cu-L1 and Cu-L2 are also nicely evi-
denced in the ES-MS spectra, which show intense peaks at
m/z=645 [CuL1H]ꢁ and 558 [CuL2H]ꢁ. These data are con-
sistent with the formation of mononuclear CuI complexes
with LMCT absorption bands at higher energy (ꢀ230 nm)
UV and CD titrations with CuI: Binding of CuI was investi-
gated by UV and CD spectroscopy. It is known that CuI dis-
proportionates in Cu0 and CuII in water. Therefore the ex-
4420
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4418 – 4428

Metal Chelating Agents
FULL PAPER
Table 2. UV and CD characteristics of CuI complexes at pH 7.4 (20 mm
phosphate buffer/CH3CN (v/v=9/1)).
the beginning of UV (Figure S2 in the Supporting Informa-
tion) and CD titrations (De(340nm) ꢀ1000 molꢁ1 Lcmꢁ1 for
1 equiv Cu). This exemplifies the formation of polynuclear
copper complexes even in excess of L3. Besides, the high-
energy band around 230 nm increases linearly with increas-
ing CuI concentration and does not show any endpoint at
the addition of one equivalent of Cu, as observed with L1
and L2. The presence of the carboxylate functions in L3 at
pH 7.4 disfavours the formation of the mononuclear copper
complex.
UV l [nm] (e [molꢁ1 Lcmꢁ1])
CD
Cu-L1
230 (14300)
267 (7200)
340 (40)
230 (15000)
267 (13500)
340 (2500)
230 (17600)
266 (7900)
340 (20)
230 (17600)
266 (13500)
340 (2600)
230 (17000)
267 (12200)
340 (1350)
237 (ꢁ)
262 (+), 288(ꢁ)
(Cu2-L1)3
235 (ꢁ)
259 (+), 284(ꢁ)
310 (ꢁ) 340(+)
222 (+)
Cu-L2
253 (ꢁ), 279(+)
(Cu2-L2)x
220 (+)
1H NMR spectra of Cu-L1,2: The CuI complexes of L3 give
258 (ꢁ), 291(ꢁ)
313 (ꢁ) 341(+)
226 (+)
1
poorly resolved H NMR spectra and were not extensively
ACHTUNGTRENNUNG
(Cu2-L3)x
1
studied. In contrast, the H NMR spectra of L1 and L2 in
256 (ꢁ), 278 (ꢁ)
presence of CuI show the formation of a C3-symmetric com-
plex for less than one metal equivalent (Figure 3). The equi-
librium between the free ligand and the complex is slow on
the NMR time scale at 500 MHz. In these complexes, the
chemical shift differences between the two b-protons of the
cysteines (ABX spin system) and of the apical NCH2 pro-
tons (AB spin system) are greatly enhanced with respect to
the free ligand, which suggests that the ligands adopt rigid
conformations in the metal complex with the three thiolate
arms wrapped around the metal ion. A closer look to the
ABX spin system of the b-protons of the cysteines indicates
343 (+)
ꢁ
that the conformation around the Ca Cb bonds of the cys-
Figure 2. CD titration of L1 (ꢀ50 mm) with CuI at pH 7.4 (20 mm phos-
phate buffer/CH3CN (v/v=9:1)). Left: 0–1 equivalents Cu. Right: 1–
2 equivalents of Cu.
than in copper clusters (ꢀ260 nm).[27] Besides, in solutions
of Cu-L1,2 in a 9:1 (v/v) mixture of phosphate buffer (20 mm,
pH 7.4) and acetonitrile, the cysteine free-thiol concentra-
tion, measured following Ellmanꢀs procedure[28] is insignifi-
cant, which demonstrates that the three thiolate functions of
the ligand are coordinated to the CuI ion in a mononuclear
CuS3 coordination environment.
When an excess of CuI is used with respect to L1 and L2,
the UV and CD features change dramatically and indicate
the formation of CuI polymetallic species. In the UV spec-
tra, the high-energy band around 230 nm attributed to the
CuS3 mononuclear centre reaches a plateau at the addition
of one to two equivalents of Cu and weak, low-energy con-
tributions (around 340 nm) appear. These weak low-energy
bands also appear in the CD spectra with maxima at 310
and 340 nm as seen in Figure 2 (right) and are attributed to
cluster-centred transitions.[26,27] Going from one to two
equivalents of CuI, several isodichroic points are detected in
the CD spectra obtained with both ligands L1 and L2, dem-
onstrating that the mononuclear complex Cu-L1,2 is trans-
formed into a single cluster complex (Cu2-L1,2)x. The cluster
(Cu2L1)3 was previously identified by ES-MS.[19]
Figure 3. 500 MHz 1H NMR spectra of L1 and L2 with 0.5 equivalents of
The titration of the acid derivative L3 with CuI show a dif-
ferent behaviour than those observed with L1 and L2.
Indeed, low-energy bands around 340 nm are detected at
Cu
(v/v=9:1) at 298 K. L and C mean the free ligand and the Cu-L complex,
respectively.
ACTHGNURTNE(NUNG CH3CN)4.PF6 at pD 7.4 (20 mm phosphate buffer in D2O/CD3CN
AHCTUNGTRENNUNG
Chem. Eur. J. 2011, 17, 4418 – 4428
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4421

P. Delangle et al.
teine residues is different for the two copper complexes.
tion (3) is the free ligand in the same solvent. As seen in
Table 3, these results confirm the formation of copper
3
Indeed for Cu-L1, the values of the J coupling constants are
4 and 2 Hz, characteristic of HaCCHb dihedral angles of
about 608 from the Karplus equation[29] (conformation I,
Figure 3). In contrast, the patterns are totally different for
monoACTHGNUTERNNUG
nuclear complexes Cu-L1 and Cu-L2.
As we suspected a network of hydrogen bonds involving
the amide functions of the “upper cavity” (NHarm(i)···O=
Carm(i+1)), we investigated the temperature coefficients of the
3
Cu-L2 with two very different J coupling constants (4 and
12 Hz) characteristic of HaCCHb dihedral angle values of
approximately 60 and 1808 (conformation II, Figure 3). Fur-
thermore, the a-proton is especially shielded in Cu-L2
(3.9 ppm instead of 4.5 ppm in the free ligand or 4.6 ppm in
Cu-L1), because the metal ion is closer to this proton in con-
formation II.
The diffusion coefficients were measured by pulsed gradi-
ent spin echo (PGSE) NMR to infer the species moleculari-
ty in solution.[30] This method has proved to be a useful tool
for probing the presence of unimolecular, bimolecular or
oligomeric species in solution.[31] Indeed, the diffusion con-
stant D can be related to the hydrodynamic radii of the mol-
ecules by the Stokes–Einstein equation [Eq. (2)] in which k
is the Boltzmann constant, T the absolute temperature, h
the viscosity and rH the hydrodynamic radius of the diffusing
species, considered as a hypothetical hard sphere that diffus-
es with the same speed as the particle under examination.
amide proton resonances in H2O/D2O (v/v=90:10). Indeed,
intermolecular hydrogen bonds and those to the solvent are
readily cleaved with increasing temperature, which leads to
large temperature coefficients. In contrast, it is generally as-
sumed that NHs with temperature gradients less negative
than ꢁ3 ppb8Cꢁ1 are solvent-shielded in stable peptide
structures, and indicate intramolecular hydrogen bonding.[32]
The temperature coefficients of the secondary amide proton
resonances of the two complexes Cu-L1, and Cu-L2 are re-
ported in Table 4, with those of the free ligands for compari-
Table 4. Temperature coefficients of the amide proton resonances, in
phosphate buffer (20 mm, pH 7.4)/CD3CN, 90/10 (v:v).
ꢁDd/DT [ppb per 8C]
L1
L2
Free L
CuL
6.8
0.8
6.5
0.1
kT
6phrH
ð2Þ
D ¼
son. As expected, the temperature coefficients of the amide
protons of the two ligands are large, because these protons
strongly interact with the solvent in the flexible structures of
the free tripods. In contrast, the chemical shift variations of
complexes are close to zero, which evidences intramolecular
hydrogen bonds involving the amide protons of the “upper
cavity”. The C3-symmetrical structures of the metal com-
plexes were built in Chem3D by imposing a trigonal planar
The determination of D by diffusional NMR spectroscopy
for shape-similar complexes and ligands is thus an efficient
tool for deducing the molecular mass of an unknown species
(i, molar mass Mi) in solution, when a reference compound
(r, molar mass Mr) is measured under the same conditions
[Eq (3)].[30]
I
[33]
ꢁ
coordination to Cu with S Cu distances of 2.25 ꢃ. As ex-
pected in these structures, the three peptide bonds of the
“upper cavity” are likely to form intramolecular H-bonds in-
volving the NH of one arm and the CO of the next arm
(Figure 4). The H···O, N···O distances are 3.1 and 3.9 ꢃ, re-
rffiffiffiffiffiffi
Di
Dr
Mr
Mi
3
ð3Þ
¼
The experimental diffusion coefficients and the calculated
mass of the complexes (Mdiff) are presented in Table 3. The
reference compound used for the mass calculation in Equa-
ꢁ
spectively and the N H···O angle is 1378, characteristic of
weak hydrogen bonds.[34] Of course, NMR experiments do
not allow us to discriminate between the C3-symmetrical
structure with three weak hydrogen bonds like in Figure 4
and a nonsymmetrical structure with only two strong hydro-
gen bonds, which would be dynamic on the NMR timescale
providing an average C3-symmetrical species spectrum. The
hydrogen-bond network probably contributes to the stabili-
sation of the mononuclear copper complexes of L1 and L2.
In an excess of CuI, the 1H NMR spectra of L1 and L2
complexes broaden, which was assigned to the formation of
higher nuclearity species, experiencing intramolecular dy-
namics associated with exchange of the sulfur ligands from
one Cu ion to another. Besides, these polynuclear complexes
were also evidenced above with UV and CD spectroscopy.
The diffusion coefficient of (Cu2-L1)x could be precisely
measured thanks to the 1H NMR resonances of the ethyl
groups of the ester functions. The diffusion coefficient,
D(Cu -L1) =2.0(1)ꢄ10ꢁ10 m2 sꢁ1, confirms the trimeric nature
Table 3. Translational diffusion coefficients measured by PFG-1H NMR
at 298 K in phosphate buffer (20 mm, D2O, pD 7.4)/CD3CN v/v=9:1. Ex-
perimental errors on the last character are indicated in parentheses.
D
Mdiff
[gmolꢁ1
Mformula
[a]
[b]
[m2 sꢁ1 molꢁ1]ꢄ1010
G
]
[gmolꢁ1
ACHTUNGTERNNUNG ]
L1
3.20(4)
3.02(4)
3.1(1)
2.80(4)
2.4(1)
3.40(4)
3.30(4)
3.20(5)
2.8(1)
–
585
645
789
647, 1292
1386
498
558
701
559, 1118
1212
Cu-L1
700(50)
650(90)
870(70)
1400
–
545(40)
600(50)
890
Pb-L1
Zn-L1, (Zn-L1)2
A
ACHTUNGTNER(NUNG 200)
L2
Cu-L2
Pb-L2
Zn-L2
A
A
2.5(1)
1250
ACHTUNGTRENNUNG
[a] Molecular mass calculated with Equation (3), reference compound is
the free ligand. [b] Molecular mass calculated with the chemical formula
of the species.
AHCTUNGTRENNUNG
2
x
4422
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4418 – 4428

Metal Chelating Agents
FULL PAPER
Table 5. Characteristics of the LMCT bands of Zn, Cd and Pb complexes
in phosphate buffer (20 mm, pH 7.4) for Zn and Cd and in Bis-Tris buffer
(20 mm, pH 7) for Pb. De=eACHTNUTGRNEUNG
(MLi)-e(Li).
UV l [nm] (De [molꢁ1 Lcmꢁ1])
L1
L2
L3
ZnII
CdII
PbII
210, (21500)
230 (20000)
350 (5500)
210 (22600)
230 (22000)
350 (4900)
210 (17250)
227 (21500)
350 (4900)[a]
[a] For the Pb/L3 system, higher energy bands develop in excess of PbII
with an absorption maximum at 215 nm.
complexes described by Pecoraro et al. show absorption
data very close to ours (lmax ꢀ230 nm and eꢀ20000–
22000 molꢁ1 Lcmꢁ1).[37] Besides, Vasak et al. determined an
ꢁ
averaged contribution per S Cd bond of 5500–
6500 molꢁ1 Lcmꢁ1.[38]
Figure 4. C3-symmetrical structure of Cu-L1 showing the hydrogen bonds
in the “upper cavity”. C: grey, H: white, O: red, N: blue, S: yellow, Cu:
green.
1
The H NMR spectra of these complexes are very similar
to those of CuI and indicate the formation of C3-symmetric
species (Figure S3 in the Supporting Information). They
show large NMR resonances for the addition of less than
one equivalent of metal, indicating that free ligand–complex
equilibria are more rapid on the NMR timescale with these
divalent ions than with CuI. Interestingly, the diffusion co-
of this complex and gives MACHTUNGRTEGN(NUN Cu L) ꢀ2200
(400) gmolꢁ1 and
ACHTUNGTRENNUNG
2
x
therefore x=3. The polymetallic complex Cu6-L3 can be as-
sociated with a Cu6S9 core as described in some metallothio-
neins, in which all of the CuI adopts nearly trigonal planar
coordination with exclusively bridging thiolates.[35]
AHCTUNGERTGeNNUN fficients indicate important differences for the coordination
complexes of the three metal ions (Table 3). Whereas PbII
forms mononuclear complexes Pb-L with a PbS3 environ-
ment, CdII forms dinuclear complexes Cd2-L2 with L1 and
L2. The formation of a dimeric cadmium complex has al-
Pb, Cd and Zn complexes of L1–3: The titrations of L1 and
L2 with PbII, CdII and ZnII show sharp endpoints for the ad-
dition of one equivalent of metal ion with the appearance of
ready been observed with a peptide modelling APP
charge-transfer transitions (LMCT) due to the thiolate M
(APP170–188) which contains also three cysteine residues.[39]
This dimer contains a dinuclear cluster in which two divalent
CdII are bridged by two thiolate ligands from cysteine resi-
dues. The tetrahedral coordination sites of each metal ion
are completed by nonbridging thiolate ligands as represent-
ed in Scheme 2. The formation
II
ꢁ
bonds. Typical titrations are presented in Figure 5 for ligand
L2. As seen in Table 5, the LMCT bands, in particular the
extinction molar coefficients are consistent with metal ions
coordinated with the three thiolates of the ligand. Indeed,
Godwin et al. characterised tristhiolato–lead complexes with
absorption
bands
at
about
335 nm
with
e
of a unique complex was con-
ꢀ4000 molꢁ1 Lcmꢁ1.[36] Well-defined tristhiolato cadmium
firmed by the CD titrations
with L1 and L2, which show the
transformation of the ligand
into one unique metal species
with
two
dichroic
bands
Scheme 2. Schematical repre-
sentation of the dimer Cd2S6.
(210(+) and 248(ꢁ)) and an
isodicroic point at 233 nm (Fig-
ure S4 in the Supporting Infor-
mation). According to diffusion
data, this unique complex was attributed to Cd2-L2. The dif-
fusion coefficients measured for the zinc complexes are in
between those expected for mono and dimeric species, sug-
gesting equilibrium between the two complexes Zn-L and
Zn2-L2. ES-MS titrations of L1 and L2 with these divalent
metal ions evidence mainly the monomolecular species M-
L, but weak signals corresponding to the bimolecular species
M2-L2 are also seen in the ES-MS spectra of cadmium com-
1
plexes. Surprisingly, all these complexes give nice H NMR
Figure 5. UV titration of L2. A) (65 mm) with PbII in Bis-Tris buffer
(20 mm, pH 7), B) (32 mm) with CdII in phosphate buffer (20 mm, pH 7.4)
and C) (32 mm) with ZnII in phosphate buffer (20 mm, pH 7.4) with 0–
2 equivalents of metal per L2.
spectra indicative of C3 symmetry, even though the species
M2-L2 has lower symmetry. This demonstrates that the intra-
molecular dynamics in the dimer is very rapid on the NMR
Chem. Eur. J. 2011, 17, 4418 – 4428
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4423

P. Delangle et al.
timescale and prevents the observation of the real symmetry
of the complex.
to form the mononuclear complex, with aliquots of the com-
petitor (BCS2ꢁ) solution could be fitted according to Equa-
tion (5) with very high conditional stability constants (logK
ꢀ19) reported in Table 6. The measurements were per-
The ligand L3 gives similar UV titrations to L1 and L2
with Cd and Zn. In contrast with PbII, the UV spectra
evolve significantly after one equivalent of Pb is added (Fig-
ure S5 in the Supporting Information). A first LMCT band
grows until the addition of one equivalent of Pb, with the
same characteristic as Pb-L1 and Pb-L2, indicative of a PbS3
environment. Then a second more intense LMCT band ap-
pears at lower energy (315 nm) characteristic of a lower
number of coordinated thiolate functions.[14] This singular
behaviour was assigned to the coordination of the non inno-
cent acetate function in a five-membered chelate ring which
is observed only for the large borderline ion PbII. These ti-
trations are perfectly fitted with the successive formation of
Table 6. Conditional stability constants at pH 7.4 and stability constants
of the metal complexes of L1 and L2 and 298 K. log K were calculated for
the formation of ML complexes.
CuI
ZnII
CdII
PbII
logKpH7.4
logK[a]
L1[19]
L2
19.2(1)
18.8(1)
23.6(1)
24.5(1)
10.3(2)
11.0(2)
14.6(2)
16.7(2)
11.8(3)
12.1(3)
16.1(3)
17.8(3)
10.1(1)
10.5(1)
14.4(1)
16.2(1)
L1
L2
[a] logK values were calculated from logKpH7.4 and pKa values according
to Equations (1) and (2) in the Supporting Information.
Pb-L3
(340 nm,
4900 molꢁ1 Lcmꢁ1),
Pb2L3(315 nm,
6800 molꢁ1 Lcmꢁ1) and Pb3-L3 (315 nm, 11 000 molꢁ1 Lcmꢁ1)
formed with a range of BCS concentrations and gave the
same constant according to Equation (5), which demon-
strates the absence of ternary species with BCS. It appears
clearly that the two ligands L1 and L2 are very efficient
copper complexing agents with affinities somewhat higher
than those reported for a model peptide of Atx1 (logK=
17.4 at pH 7.4)[14,15] or for the whole Atx1 protein.[40,41]
These very large conditional stability constants are of the
same order of magnitude as the one found in metallothio-
neins.[17]
as shown in the Supporting Information.
Affinities and selectivities: The affinity of the three tripods
for CuI was evaluated at physiological pH, by UV/Vis titra-
tions in presence of a chelator of known affinity. Bathocu-
proine disulfonate (Na2BCS) has been demonstrated to
form a stable 1:2 complex, CuACHTNUGRTNEUNG
(BCS)23ꢁ, according to Equa-
tion (4) with a constant, logb12 =19.8.[40]
3ꢁ
Cuþ þ 2 BCS2ꢁ ! ½CuðBCSÞ2ꢂ
ð4Þ
3ꢁ
Cu-Li þ 2 BCS2ꢁ ! Li þ ½CuðBCSÞ2ꢂ
ð5Þ
The complex [CuACHTUNGTRENNUNG
(BCS)2]3ꢁ exhibits an absorption band in
The conditional stability constants of L1 and L2 for di-
the visible region with its maximum at 483 nm (e=
13300mꢁ1 cmꢁ1). Addition of BCS to a solution of Cu-Li in a
9:1 (v/v) mixture of phosphate buffer (20 mm, pH 7.4) and
acetonitrile shows the appearance of the absorption band of
AHCTUNGTRENNUNG
[CuACHTUNGTRENNUNG
(BCS)2]3ꢁ, which corresponds to the transfer of the
metal cation from the tripod to BCS2ꢁ. Figure 6 displays the
quantity of BCS required to displace 50% of Cu from the
CuI complex of Li. Ligand L1 has the highest affinity for CuI
as 80 equivalents of BCS are required to remove 50% of
CuI from the complex. In comparison, L3 is a poor ligand as
only 2.5 equivalents of BCS displace 50% of the Cu. Titra-
tions of L1 and L2, preloaded with 0.8–0.9 equivalents of CuI
conditional stability constant at physiological pH (KpH7.4
)
and the stability constant (K) calculated from the condition-
al stability constants and the ligandꢀs pKa given in Table 1
are listed in Table 6. These data show rather similar selectiv-
ities to those obtained previously with thiol-rich ligands,
that is, cysteine-containing peptides: CuI @CdII >PbII ꢀ
ZnII.[12,14,15] The metal complexes of L1,2 are systematically
more stable than those of PC, which mimics Atx1 binding
loop (MXCXXC sequence). It can also be noted that the
stability constants of L2, logK in Table 6, which are not de-
pendent on pH, are larger than the same values for L1, be-
cause the electron-withdrawing effect of the ester group in
L1 is larger than the one of the amide group in L2. Finally,
the use of three thiolates in comparison to only two thio-
lates in the ligandꢀs architecture allowed us to obtain more
stable CuI complexes, with slightly lower Cu/Zn selectivities.
Discussion
Figure 6. Number of BCS equivalents per Cu, required to form 50% of
[Cu
20 mm phosphate buffer, pH 7.4 at 298 K, acetonitrile 9:1 v/v.
ACHTUNGTRENNUNG
(BCS)2] from samples containing [Cu]=0.8 [Li] and [Li]=50 mm in
The objective of this work was the design of efficient chela-
tors of the essential metal ion copper in the +I oxidation
4424
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4418 – 4428

Metal Chelating Agents
FULL PAPER
state present in cells. Indeed, copper may become toxic in
case of dysfunction of its homeostasis like in Wilsonꢀs dis-
ease. We previously demonstrated that small peptide se-
quences with two cysteine residues chelate metal ions with
their two thiolate functions with a high selectivity for the
soft ions CuI and HgII with respect to the three divalent cat-
ions CdII, PbII and ZnII.[14,15] In particular, the selectivity
with respect to ZnII, which is a bioavailable ion present in
cells and thus a potential competitor, is a key parameter to
obtain efficient chelators in vivo.[12] We have shown here
that the introduction of three cysteine residues in a tripodal
architecture provides even more efficient CuI chelators with
large selectivities with respect to ZnII. Three tripodal mole-
cules derived from nitrilotriacetic acid were synthesised and
differ by the nature of the carbonyl function adjacent to the
thiol group. Our results demonstrate that the nature of the
carbonyl function has a great impact on the ligandsꢀ com-
plexation properties. In particular, charged residues de-
cule, which was clearly evidenced by the very low chemical-
shift temperature coefficient of the amide protons. Com-
plexes Cu-L1 and Cu-L2 exhibit rather similar conditional
stability constant at physiological pH and these very large
values are comparable to constants obtained with metallo-
thioneins.[17]
A key parameter in the design of metal chelator is their
selectivity with respect to bioavailable metal ions, the ho-
meostasis of which should not be disturbed. The affinity of
L1,2 for ZnII is higher than those obtained with peptides
bearing only two cysteines like PC,[14] as ZnII is known to
prefer sulfur-rich tetrahedral coordination sites.[42] The con-
ditional stability constants at pH 7.4 obtained with L1 and L2
are in the same range as the one reported by Faller et al.
(logK=9.5) with a peptide derived from APP and also con-
taining three cysteine residues.[39] Even if L1 and L2 have sig-
nificant affinities for ZnII, their selectivities for CuI with re-
spect to ZnII remain especially large, 8–9 orders of magni-
tude and could be used for the selective decorporation of
copper in cells.
ACHTUNGTRENNUNGstabilise the thiolate-based metal coordination, whereas neu-
tral groups tend to favour the formation of mononuclear
CuI complexes.
The negatively charged derivative L3, which has three car-
boxylate functions close to the metal binding site, forms
complexes of low stability, as seen from the BCS competi-
tion measurements. An even more striking effect is the in-
ability of L3 to bind CuI in a mononuclear CuS3 complex.
Indeed only polymetallic copper(I) thiolate species were evi-
denced with L3. The destabilisation of copper(I) complexes
by close carboxylate side chains of aspartate or glutamate
residues has also been observed in cysteine-containing cyclo-
decapeptides.[12] This effect can be attributed to the large
electrostatic repulsions generated by the proximate nega-
tively charged functions which are moved closer in the
metal complexes.
Conclusion
In conclusion, we have developed tripodal pseudopeptide
scaffolds extended by three converging cysteines, which
afford soft sulfur donors for the complexation of soft metal
ions. The nature of the carbonyl function next to the thiols
is determinant in the complexation properties of these novel
ligands. Indeed, the presence of carboxylate groups in L3
tends to lower the complexesꢀ stabilities and these functions
may also interfere in the coordination as it was evidenced
with the borderline cation PbII. In contrast, the ester and
amide derivatives L1 and L2 show interesting complexation
properties. In particular, they form mononuclear CuI com-
plexes with a unique MS3 coordination in water, which is
stabilised by a hydrogen-bonding network in the upper
cavity of the tripod. The CuS3 isolated centre is of particular
interest as it is proposed as an intermediate in copper trans-
fer in cells between copper chaperones and ATPases.[2,4,5,44]
Therefore Cu-L1 and Cu-L2 may be relevant models to study
this transfer, which is a key reaction in copper regulation
and distribution. Besides, the introduction of three cysteines
in L1 and L2 allowed us to obtain CuI chelators with very
high affinities, in the same range as metallothioneins (logK
ꢀ19).[17] Moreover, these ligands display large selectivities
for the targeted metal ion CuI with respect to bioavailable
ZnII, which is critical to obtain efficient decorporation
agents in vivo that do not alter essential ion homeostasis.
In contrast, the two neutral tripodal ligands L1 and L2,
which bear primary amide or ester groups, form very stable
mononuclear complexes of CuI with a CuS3 geometry. Such
a coordination geometry was described in the solid state in
CuI complexes with monodentate sulfur ligands that were
crystallised in organic solvents.[33,43] Nevertheless, to our
knowledge, such complexes with a CuI metal ion coordinat-
ed by three thiolate donor groups inducing a isolated CuS3
centre were never described in water. The complexes Cu-L1
and Cu-L2 were fully characterised by UV, CD, and
1H NMR spectroscopy and ES-MS and are especially stable
(logKꢀ19). These affinities may be compared to the one
previously determined with the cyclodecapeptide PC, mim-
icking the Atx1 binding loop, which was studied in the same
conditions (logK=17.4). The introduction of three cysteines
in this tripodal architecture provides CuI complexes of en-
hanced stability, nearly 1.5 orders of magnitude in compari-
son to peptides bearing two cysteines. The two tripods L1
and L2 afford three converging cysteine residues for metal
complexation, which promote the coordination of CuI in a
mononuclear CuS3 geometry. Furthermore, the whole com-
plex structure is stabilised by an intramolecular hydrogen-
bonding network in the upper cavity of the tripodal mole-
Experimental Section
General: Solvents and starting materials were purchased from Aldrich,
Acros, Fluka and Alfa Aesar and used without further purification.
HCys
solutions in water were prepared from ultrapure laboratory grade water
(Trt)OEt was synthesised according to a published procedure.[45] All
ACHTUNGTRENNUNG
Chem. Eur. J. 2011, 17, 4418 – 4428
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4425

P. Delangle et al.
that has been filtered and purified by reverse osmosis using Millipore
MilliQ reverse-osmosis cartridge system (resistivity 18 MWcm). Thin-
layer chromatography (TLC) was performed on silica gel 60 F254
(Merck). Flash chromatography was performed on silica gel 60 (40–
63 mm, Merck). Analytical and preparative HPLC was performed with a
VWR system fitted with a purosphere RP18 column (l=250 mm, Ø=
4.6 mm and p=5 mm for analytical column; l=250 mm, Ø=40 mm and
p=10 mm for preparative column). Solvent conditions were as follows:
solvent A=water/TFA (99.925:0.075), solvent B=CH3CN/water/TFA
(90:10:0.1). Flow rates of 1 mLminꢁ1 and 75 mLminꢁ1 were used for ana-
lytical and preparative column respectively with UV monitoring at
214 nm. 1H NMR and 13C NMR spectra were recorded on a Mercury
Varian 400 spectrometer or on a Bruker 500 spectrometer. Chemical
shifts are reported in ppm with the solvent as the internal reference.
Mass spectra were acquired with a Finigan LCQ-ion trap equipped with
an electrospray source. Elemental analyses were performed by the Ser-
vice Central d’Analyse (Solaize, France).
product (0.242 mg, 83%) was used without further purification. 1H NMR
([D6]DMSO, 400 MHz, 298 K): d=2.37–2.46 (m, 6H; CH2SC), 3.32 (s,
6H; CH2CO), 4.17–4.21 (m, 3H; CH), 7.20–7.37 (m, 45H; C
8.46 ppm (d, J=7.4 Hz, 3H; NH); 13C NMR (CD3CN, 100 MHz, 298 K):
d=34.03 (CH2S), 52.34 (CH), 60.68 (CH2CO), 130.00–127.67 ((C6H5)3),
145.16 (C(C6H5)3), 172.02 and 171.61 ppm (2CO); ES-MS: m/z: =1249.2
[M+Na]+.
Synthesis of ligand L1: Trifluoroacetic acid (1.81 mL, 24.4 mmol) and tri-
ethylsilane (0.47 mL, 2.9 mmol) were successively added to a solution of
ACHTUNGTRENNUNG(C6H5)3),
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
1 (0.640 g, 0.49 mmol) in CH2Cl2 (15 mL) under argon. After stirring for
30 min at room temperature, the mixture was evaporated. The resulting
crude product (0.627 mg) was purified by preparative HPLC (tR =
12.7 min ; linear gradient 50:50 to 0:100, A/B in 15 min; analytical
HPLC: tR =7.1 min with the same gradient). Ligand L1 was obtained as a
white oily solid (0.110 g, 49%). 1H NMR (CD3CN, 500 MHz, 298 K): d=
1.25 (t, J=7.1 Hz, 9H; CH3), 1.97 (t, J=8.8 Hz, 3H; SH), 2.95 and 3.00
(ABXY, JAX =4.6 Hz, JBX =6.1 Hz, JBY =9.0, JAY =9.3 Hz, JAB =14.0 Hz,
6H; CH2SH), 3.48 and 3.52 (AB, JAB =16.3 Hz, 6H; CH2CO), 4.18 and
4.22 (ABX3; JAX =7.1 Hz, JBX =7.1 Hz, JAB =10.8 Hz, 6H; CH2-CH3),
4.70 (ddd, J=4.7, 6.2, 8.0 Hz, 3H; CH), 7.71 ppm (d, J=8.0 Hz, 3H;
NH); 13C NMR (CD3CN, 100 MHz, 298 K): d=14.97 (CH3), 27.40
(CH2SH), 55.81 (CH), 59.75 (CH2CO), 63.02 (CH2CH3), 171.61 and
172.02 ppm (2CO); ES-MS: m/z: =585.0 [M+H]+, 607.3 [M+Na]+.
Synthesis of compound 1: Nitrilotriacetic acid (0.196 g, 1.03 mmol) was
added to a solution of HCys
(20 mL). Then the mixture was cooled at 08C and N-ethyl-N’-(3-dimeth-
ylaminopropyl)carbodiimide (0.587 g, 3.06 mmol) and 1-hydroxybenzo-
triazole hydrate (0.414 g, 3.06 mmol) were successively added. The re-
action mixture was stirred at room temperature for 24 h under argon.
ACHTUNGTRNE(NUNG Trt)OEt (1.200 g, 3.06 mmol) in DMF
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
Synthesis of ligand L2: Trifluoroacetic acid (1.4 mL, 17.84 mmol) and tri-
AHCTUNGERTGeNNUN thylsilane (0.54 mL, 2.14 mmol) were successively added to a solution of
After evaporation of the solvent, the residue was dissolved in ethyl ace-
tate (100 mL). The organic layer was washed with water (2ꢄ50 mL), sa-
turated NaHCO3 solution (50 mL) and brine (2ꢄ50 mL). The organic
layer was dried over Na2SO4 and concentrated under reduced pressure.
The resulting crude product (1.391 g) was purified by column chromatog-
raphy on silica gel (100 mL, CH2Cl2/ethyl acetate, 80:20) to give com-
2 (0.437 g, 0.357 mmol) in CH2Cl2 (15 mL) under argon. After stirring for
50 min at room temperature, the mixture was evaporated. The resulting
crude product (0.6 mg) was purified by preparative HPLC (tR =18.9 min
linear gradient from 95:5 to 65:35, A/B in 25 min; analytical HPLC: tR =
16.4 min with the same gradient) and gave the compound L2 (0.065 g,
37%) as a white solid. 1H NMR (D2O, 500 MHz, 298 K): d=2.87 (ABX,
1
pound 1 (1.103 g, 82%) as a white powder. H NMR (CD3CN, 400 MHz,
298 K): d=1.05 (t, J=7.0 Hz, 9H; CH3), 2.39 and 2.68 (ABX, JBX
4.1 Hz, JAX =8.0 Hz, JAB =12.7 Hz, 6H; CH2S), 3.17 and 3.29 (AB, JAB
=
J
AX =4.9 Hz, JBX =7.5 Hz, JAB =14.2 Hz, 6H; CH2SH), 3.55 (s, 6H;
=
CH2CO), 4.45 ppm (dd, J=7.5, 4.9 Hz, 3H; CH); 13C NMR (D2O,
100 MHz, 298 K): d=28.18 (CH2SH) ; 57.93 (CH), 60.85 (CH2CO),
175.98 and 177.04 ppm (2CO); ES-MS: m/z: 498.1 [M+H]+.
15.0 Hz, 6H; CH2CO), 3.84 and 3.96 (ABX3, JAX =7.0 Hz, JBX =7.0 Hz,
AB =10.9 Hz, 6H; CH2CH3), 4.34 (td, J=4.0, 8.2 Hz, 3H; CH), 7.13–7.17
J
(m, 30H; SCACHTUNGTRENNUNG(C6H5)3), 7.22 (d, J=7.4 Hz, 15H; SCACHUTNGTREN(NUGN C6H5)3), 7.56 ppm (d,
Synthesis of ligand L3: Trifluoroacetic acid (1.77 mL, 23.8 mmol) and tri-
AHCTUNGERTGeNNUN thylsilane (0.456 mL, 2.85 mmol) were successively added to a solution
J=8.6 Hz, 3H; NH); 13C NMR (CD3CN, 100 MHz, 298 K): d=14.39
(CH3), 33.69 (CH2S), 52.11 (CH), 57.97 (CH2CO), 62.21 (CH2CH3),
of 3 (0.584 g, 0.476 mmol) in CH2Cl2 (21 mL), under argon. After stirring
for 30 min at room temperature, the mixture was evaporated. The result-
ing crude product (744.5 mg) was purified by preparative HPLC (tR =
13.75 min ; linear gradient 95:5 to 0:100, A/B in 15 min; analytical
HPLC: tR =12.5 min with the same gradient). Ligand L3 was obtained as
a white powder (0.209 g, 88%). 1H NMR (D2O, 400 MHz, 298 K): d=
129.90–127.16 ACHTUNGTRENNUNG(C6H5), 144.70 (CACHTUGNTERN(NUGN C6H5)3), 171.66 and 170.70 ppm (2CO);
ES-MS: m/z: 1310.8 [M+H]+; elemental analysis calcd (%) for
C78H78N4O9S2·2H2O (1347.70 gmolꢁ1): C 69.51, H 6.13, N 4.16; found: C
69.42, H 6.05, N 3.90.
Synthesis of compound 2: Nitrilotriacetic acid (0.068 g, 0.357 mmol) was
added to a solution of HCys
(10 mL). Then the mixture was cooled at 08C and N-ethyl-N’-(3-dimeth-
ylaminopropyl)carbodiimide (0.212 g, 1.10 mmol) and 1-hydroxybenzo-
triazole hydrate (0.150 g, 1.11 mmol) were successively added. The re-
action mixture was stirred at room temperature for 24 h under argon.
ACHUTNGTNERNUG(Trt)ACHTUNGTERN(NUGN NH2) (0.401 g, 1.10 mmol) in DMF
3.07 and 3.01 (ABX,
JAX =4.3 Hz, JBX =6.8 Hz, JAB =14.5 Hz, 6H;
CH2SH), 3.81–3.90 (m, 6H; CH2CO), 4.72 ppm (t, J=5.9 Hz, 3H; CH);
13C NMR (CD3CN, 100 MHz, 298 K): d=27.86 (CH2SH), 57.38 (CH),
60.07 (CH2CO), 173.46 (COOH), 175.77 ppm (NHCO); ES-MS: m/z:
499.0 [MꢁH]ꢁ.
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
After evaporation of the solvent, the residue was washed with water
(25 mL) and filtrated. Then the solid was dissolved in dichloromethane
(100 mL) and the organic layer was washed with water (3ꢄ50 mL) and
brine (1ꢄ50 mL). The organic layer was dried over Na2SO4 and concen-
trated under reduced pressure to give the compound 2 ( 0.404 g, 92%) as
a white powder. 1H NMR (CD3CN, 400 MHz, 298 K): d=2.37–2.44 (m,
6H; CH2S), 3.14 and 3.19 (AB, JAB =16.4, 6H; CH2CO), 4.00–4.06 (m,
3H; CH), 5.70 (s, 3H; NH2), 6.24 (s, 3H; NH2), 7.16–7.32 (m, 45H; SC-
Compounds L1, L2 and L3 are sensitive to air-oxidation. Therefore they
were stored and manipulated in a glove box (Argon, O2 <0.1 ppm).
Solution preparation: Since the cysteine residues in the chelators are sus-
ceptible to air oxidation, all the solutions were prepared in a glove box
under argon atmosphere. Fresh solutions of the ligand were prepared
before each experiment, using the appropriate buffer prepared with de-
oxygenated Milli-Qꢅ water (Millipore) and acetonitrile. Solutions of L1
were systematically prepared with 10% acetonitrile (vol), except for po-
tentiometry experiments for which only 2% acetonitrile (vol) were used.
Solutions of L2,3 were prepared with 10% acetonitrile (vol) for copper ti-
trations to avoid copper(I) disproportionation or for NMR samples to
use the residual signal of CD2HCN as an internal reference. For other ex-
periments solutions of L2,3 were prepared without acetonitrile.
ACHTUNGTRENNUNG
(ACHTUNGTRENNUNG(C6H5)3), 150.03 (CACHTUNGTRENNUNG(C6H5)3), 176.15 and 178.074 ppm (2CO); ES-MS:
m/z: 1223.8 [M+H]+; elemental analysis calcd (%) for
C72H69N7O6S3·H2O (1242.57 gmolꢁ1): C 69.60, H 5.76, N 7.89; found: C
69.60, H 5.72, N, 7.95
The final concentration of the ligand solutions were determined by mea-
Synthesis of compound 3: Compound 1 (0.310, 0.236 mmol) was dissolved
in ethanol (6 mL) and LiOH (1m, 0.95 mL, 0.95 mmol) was added. The
reaction mixture was stirred at room temperature for 1 h. Then, the re-
AHCTUNGTERGsNNUN uring the cysteine free-thiol concentration following the Ellmanꢀs proce-
dure.[28] This procedure uses 5,5’-dithiobis-2-nitrobenzoic acid (DTNB) as
an indicator: each free thiol group present in the peptide yields one
equivalent of TNB2ꢁ (e412 nm (TNB2ꢁ)=14150mꢁ1 cmꢁ1).
ACHTUNGTRENNUNGaction mixture was evaporated and the residue was dissolved in water
(6 mL) and HCl (1m) was added until pH 4–5. The aqueous layer was ex-
tracted with ethyl acetate (15 mL). The organic layer was dried over
Na2SO4 and concentrated under reduced pressure. The resulting crude
Solutions of CuI were prepared by dissolving the appropriate amount of
CuACHTUNRTGNEUNG(CH3CN)4PF6 in deoxygenated acetonitrile. The final concentration
4426
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4418 – 4428

Metal Chelating Agents
FULL PAPER
was determined by adding an excess of sodium bathocuproine disulfonate
dient pulses respectively. The length of the gradient pulse is d, D is the
(Na2BCS) and measuring the absorbance of [Cu
A
diffusion delay and g is the gyromagnetic ratio (for protons, gH
=26.7520ꢄ107 rad.Tꢁ1 sꢁ1). The values of D and d used in the diffusion
e=13300mꢁ1 cmꢁ1).
coefficient measurements were 100 ms and 2 ms respectively. In the ex-
The other metal solutions were prepared from the corresponding salt
(CdCl2, PbCl2 or ZnCl2) in water or in the appropriate buffer and titrated
by 5 mm volumetric EDTA (Fisher Chemicals) in presence of a colouri-
metric indicator.
periments g was incremented from 2.95 to 47.2 Gcmꢁ1
.
The temperature coefficients of the amide proton resonances of the
copper complexes were measured in phosphate buffer (20 mm, pH 7.4)/
CD3CN, 90/10 (v/v) using watergate water suppression, and the signal of
CD2HCN as an internal reference.
Potentiometry: All titrant solutions were prepared using water purified
by passing through a Millipore Milli-Q reverse-osmosis cartridge system
(resistivity 18 MWcm). Carbonate-free 0.1 molLꢁ1 KOH and 0.1 molLꢁ1
HCl were prepared from Fisher Chemicals concentrates. Potentiometric
titrations were performed in 0.1 molLꢁ1 aqueous KCl in the glove box to
avoid oxidation of the thiol groups into disulfide, the temperature was
controlled to ꢃ0.18C with a circulating water bath. The p[H] (p[H]=
ꢁlog[H+], concentration in molarity) was measured in each titration with
a combined pH glass electrode (Metrohm) filled with 3 molLꢁ1 KCl and
the titrant addition was automated by use of a 751 GPD titrino (Met-
rohm). The electrode was calibrated in hydrogen ion concentration by ti-
tration of HCl with KOH in 0.1 molLꢁ1 KCl.[46] A plot of meter reading
versus p[H] allows the determination of the electrode standard potential
(E8) and the slope factor (f). Continuous potentiometric titrations with
KOH 0.1 molLꢁ1 were conducted on 22 mL of aqueous solutions contain-
ing 0.5 10ꢁ3 molLꢁ1 of the ligand. Back titrations with HCl 0.1 molLꢁ1
were systematically performed after each experiment to check whether
equilibration had been achieved. In a typical experiment, 100 points were
measured with a 2 min delay between the measurements.
Affinity constants: See Supporting Information for details.
Acknowledgements
We thank Yves Chenavier for performing some of the titration experi-
ments and Pierre-Alain Bayle for his help in acquiring NMR data. The fi-
nancial support from the Cluster de recherche Chimie de la Rꢆgion
Rhꢇne-Alpes is duly acknowledged.
[1] a) R. H. Holm, P. Kennepohl, E. I. Solomon, Chem. Rev. 1996, 96,
2239–2314; b) S. Tottey, D. R. Harvie, N. J. Robinson, Acc. Chem.
Res. 2005, 38, 775–783; c) P. P. Kulkarni, Y. M. She, S. D. Smith,
E. A. Roberts, B. Sarkar, Chem. Eur. J. 2006, 12, 2410–2422.
[2] L. A. Finney, T. V. OꢀHalloran, Science 2003, 300, 931–936.
[3] a) N. J. Robinson, D. R. Winge, Annu. Rev. Biochem. 2010, 79, 537–
562; b) B. E. Kim, T. Nevitt, D. J. Thiele, Nat. Chem. Biol. 2008, 4,
176–185.
Experimental data were refined using the computer program Hyperquad
2000.[47] All equilibrium apparent constants are expressed as concentra-
tion ratio and not activities. The ionic product of water at 258C and
0.1 molLꢁ1 ionic strength is pKw =13.78.[48] The initial concentrations of
ligands were fixed.
[4] F. Arnesano, L. Banci, I. Bertini, S. Ciofi-Baffoni, Eur. J. Inorg.
Chem. 2004, 1583–1593.
UV and CD titrations: The UV/Vis spectra were recorded with a Varian
Cary50 spectrophotometer equipped with optical fibres connected to an
external cell holder in the glove box. The circular dichroism spectra were
acquired with an Applied Photophysics Chirascan spectrometer. CD
spectra are reported in molar ellipticity ([V] in units of degcm2 molꢁ1).
[V]=qobs/ACHTUNGTRENNUNG(10lc), in which qobs is the observed ellipticity in millidegrees, l
the optical path length of the cell in centimeters, c the peptide concentra-
tion in moles per litre.
[5] A. C. Rosenzweig, Acc. Chem. Res. 2001, 34, 119–128.
[6] S. Puig, D. J. Thiele, Curr. Opin. Chem. Biol. 2002, 6, 171–180.
[7] T. Y. Tao, J. A. Gitlin, Hepatology 2003, 37, 1241–1247.
[8] B. Sarkar, Chem. Rev. 1999, 99, 2535–2544.
[9] A. I. Bush, Trends Neurosci. 2003, 26, 207–214.
[10] a) E. Gaggelli, H. Kozlowski, D. Valensin, G. Valensin, Chem. Rev.
2006, 106, 1995–2044; b) P. S. Donnelly, Z. G. Xiao, A. G. Wedd,
Curr. Opin. Chem. Biol. 2007, 11, 128–133; c) P. Faller, C. Hureau,
Dalton Trans. 2009, 1080–1094.
[11] a) O. Andersen, Chem. Rev. 1999, 99, 2683–2710; b) K. L. Haas,
K. J. Franz, Chem. Rev. 2009, 109, 4921–4960; c) L. R. Perez, K. J.
Franz, Dalton Trans. 2010, 39, 2177–2187.
[12] 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.
[13] a) A. C. Rosenzweig, T. V. OꢀHalloran, Curr. Opin. Chem. Biol.
2000, 4, 140–147; b) S. J. Opella, T. M. DeSilva, G. Veglia, Curr.
Opin. Chem. Biol. 2002, 6, 217–223; c) F. Arnesano, L. Banci, I.
Bertini, S. Ciofi-Baffoni, E. Molteni, D. L. Huffman, T. V. OꢀHallor-
an, Genome Res. 2002, 12, 255–271.
[14] P. Rousselot-Pailley, O. Sꢆnꢈque, C. Lebrun, S. Crouzy, D. Boturyn,
P. Dumy, M. Ferrand, P. Delangle, Inorg. Chem. 2006, 45, 5510–
5520.
[15] O. Sꢆnꢈque, S. Crouzy, D. Boturyn, P. Dumy, M. Ferrand, P. Delan-
gle, Chem. Commun. 2004, 770–771.
[16] K. A. Koch, M. M. O. Pena, D. J. Thiele, Chem. Biol. 1997, 4, 549–
560; M. J. Stillman, Coord. Chem. Rev. 1995, 144, 461–511.
[17] P. Faller, Febs J. 2010, 277, 2921–2930.
A 2–2.5 mL portion of the ligand solution (ꢀ 50 mm) was transferred in a
cell (1 cm path) and aliquots of the metal solution were then added. The
buffer was phosphate (20 mm, pH 7.4) for all the titrations except for PbII
for which Bis-Tris (20 mm, pH 7) was used to prevent PbII hydrolysis and
the precipitation of Pb(OH)2.[42]
ESI-MS titrations: Mass spectra were acquired on a LXQ-linear ion trap
(THERMO Scientific, San Jose, USA) equipped with an electrospray
source. Electrospray full scan spectra in the range m/z=150–2000 amu
were obtained by infusion through a fused silica tubing at 2–10 mLminꢁ1
.
The solutions were analysed in the negative and positive modes. The
LXQ calibration (m/z=50–2000) was achieved according to the standard
calibration procedure from the manufacturer (mixture of caffeine,
MRFA and Ultramark 1621). The temperature of the heated capillary for
the LXQ was set to 180–2008C, the ion-spray voltage was in the range 2–
4 kV and the injection time was 10–100 ms. The ligand solution (100 mm)
was prepared in ammonium acetate buffer (20 mm, pH 7)/acetonitrile
ACHTUNGTRENNUNG
a
500 MHz Bruker Avance spectrometer equipped with a BBI probe with
triple axis gradient field. 1H NMR spectra were recorded with 12 ppm
windows and 32 k data points in the time domain. Ligand samples pre-
pared in phosphate buffer 20 mm pH 7.4 in D2O/CD3CN (v/v=9/1) with
[18] V. Calderone, B. Dolderer, H. J. Hartmann, H. Echner, C. Luchinat,
C. Del Bianco, S. Mangani, U. Weser, Proc. Natl. Acad. Sci. USA
2005, 102, 51–56.
[19] A. M. Pujol, C. Gateau, C. Lebrun, P. Delangle, J. Am. Chem. Soc.
2009, 131, 6928–6929.
a ꢀ1 mm concentration, were titrated with aliquots of a Cu
ACHTNUGRTNE(NUGN CH3CN)4PF6
solution in CD3CN or a divalent metal ion solution in D2O.
Diffusion coefficient measurement were performed using the bipolar
[20] M. Gelinsky, R. Vogler, H. Vahrenkamp, Inorg. Chem. 2002, 41,
2560–2564.
[21] C. Jeunesse, D. Armspach, D. Matt, Chem. Commun. 2005, 5603–
5614.
stimulated spin echo sequence.[49] Diffusion coefficients were obtained
d
using, I
(d,D,g)=I0 exp
R
G
N
ties in the presence of gradient pulses of strength g and in absence of gra-
Chem. Eur. J. 2011, 17, 4418 – 4428
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4427

P. Delangle et al.
[22] a) M. Heitzmann, C. Gateau, L. Chareyre, M. Miguirditchian, M.-C.
Charbonnel, P. Delangle, New J. Chem. 2010, 34, 108–116; b) M.
Heitzmann, F. Bravard, C. Gateau, N. Boubals, C. Berthon, J.
Pecaut, M. C. Charbonnel, P. Delangle, Inorg. Chem. 2009, 48, 246–
256; c) A. Pellissier, Y. Bretonniere, N. Chatterton, J. Pecaut, P. De-
langle, M. Mazzanti, Inorg. Chem. 2007, 46, 3714–3725; d) F. Brav-
ard, C. Rosset, P. Delangle, Dalton Trans. 2004, 2012–2018; e) F.
Bravard, Y. Bretonniere, F. Wietzke, C. Gateau, M. Mazzanti, P. De-
langle, J. Pecaut, Inorg. Chem. 2003, 42, 7978–7989.
J. E. Penner-Hahn, H. A. Godwin, J. Am. Chem. Soc. 2005, 127,
9495–9505.
[37] M. Matzapetakis, B. T. Farrer, T.-C. Weng, L. Hemmingsen, J. E.
Penner-Hahn, V. L. Pecoraro, J. Am. Chem. Soc. 2002, 124, 8042–
8054.
[38] C. J. Henehan, D. L. Poutney, O. Zerbe, M. Vasak, Protein Sci. 1993,
2, 1756–1764.
[39] E.-D. Ciuculescu, Y. Mekmouche, P. Faller, Chem. Eur. J. 2005, 11,
903–909.
[23] C. J. Siddons, R. D. Hancock, Chem. Commun. 2004, 1632–1633.
[24] R. Vogler, M. Gelinsky, L. F. Guo, H. Vahrenkamp, Inorg. Chim.
Acta 2002, 339, 1–8.
[25] P. Kamau, R. B. Jordan, Inorg. Chem. 2001, 40, 3879–3883.
[26] D. L. Pountney, I. Schauwecker, J. Zarn, M. Vasak, Biochemistry
1994, 33, 9699–9705.
[40] Z. Xiao, F. Loughlin, G. N. George, G. J. Howlett, A. G. Wedd, J.
Am. Chem. Soc. 2004, 126, 3081–3090.
[41] R. Miras, I. Morin, O. Jacquin, M. Cuillel, F. Guillain, E. Mintz, J.
Biol. Inorg. Chem. 2008, 13, 195–205.
[42] J. C. Payne, M. A. terHorst, H. A. Godwin, J. Am. Chem. Soc. 1999,
121, 6850–6855.
[27] P. Faller, M. Vasak, Biochemistry 1997, 36, 13341–13348.
[28] P. W. Riddles, R. L. Blakeley, B. Zerner, Methods Enzymol. 1983,
91, 49–60.
[43] D. Coucouvanis, C. N. Murphy, S. K. Kanodia, Inorg. Chem. 1980,
19, 2993–2998; S. Zeevi, E. Y. Tshuva, Eur. J. Inorg. Chem. 2007,
5369–5376.
[29] M. Karplus, J. Am. Chem. Soc. 1963, 85, 2870–2871.
[30] A. R. Waldeck, P. W. Kuchel, A. J. Lennon, B. E. Chapman, Prog.
Nucl. Magn. Reson. Spectrosc. 1997, 30, 39–68.
[31] a) M. Alajarꢉn, A. Pastor, R.-A. Orenes, E. Martinez-Viviente, P. S.
Pregosin, Chem. Eur. J. 2006, 12, 877–886; b) E. Terazzi, S. Torelli,
G. Bernardinelli, J.-P. Rivera, J.-M. Bꢆnech, C. Bourgogne, B.
Donnio, D. Guillon, D. Imbert, J.-C. G. Bꢊnzli, A. Pinto, D. Jeanner-
at, C. Piguet, J. Am. Chem. Soc. 2005, 127, 888–903.
[32] a) E. Schievano, A. Bisello, M. Chorev, A. Bisol, S. Mammi, E. Peg-
gion, J. Am. Chem. Soc. 2001, 123, 2743–2751; b) H. Kessler,
Angew. Chem. 1982, 94, 509–520; Angew. Chem. Int. Ed. Engl.
1982, 21, 512–523; c) H. J. Dyson, M. Rance, R. A. Houghten, R. A.
Lerner, P. E. Wright, J. Mol. Biol. 1988, 201, 161–200.
[44] a) R. A. Pufahl, C. P. Singer, K. L. Peariso, S.-J. Lin, P. J. Schmidt,
C. J. Fahrni, V. C. Culotta, J. E. Penner-Hahn, T. V. OꢀHalloran, Sci-
ence 1997, 278, 853–856; b) A. K. Wernimont, D. L. Huffman, A. L.
Lamb, T. V. OꢀHalloran, A. C. Rosenzweig, Nat. Struct. Biol. 2000,
7, 766–771; c) A. Rodriguez-Granillo, A. Crespo, D. A. Estrin, P.
Wittung-Stafshede, J. Phys. Chem. B 2010, 114, 3698–3706.
[45] C. Bolzati, A. Mahmood, E. Malago, L. Uccelli, A. Boschi, A. G.
Jones, F. Refosco, A. Duatti, F. Tisato, Bioconjugate Chem. 2003, 14,
1231–1242.
[46] A. E. Martell and R. J. Motekaitis, Determination and Use of Stabili-
ty Constants, VCH, Weinheim, 1992.
[47] a) P. Gans, A. Sabatini, A. Vacca, Talanta 1996, 43, 1739–1753; b) L.
Alderighi, P. Gans, A. Ienco, D. Peters, A. Sabatini, A. Vacca,
Coord. Chem. Rev. 1999, 184, 311–318.
[33] C. D. Garner, J. R. Nicholson, W. Clegg, Inorg. Chem. 1984, 23,
2148–2150.
[34] G. A. Jeffrey, An Introduction to Hydrogen Bonding, Oxford Uni-
versity Press, New York, 1997.
[48] R. M. Smith, A. E. Martell, R. J. Motekaitis, NIST Critically Select-
ed Stability Constants of Metal Complexes Database, NIST Stan-
dard Reference Database 46, 2001.
[35] a) K. B. Nielson, C. L. Atkin, D. R. Winge, J. Biol. Chem. 1985, 260,
5342–5350; b) A. Presta, A. R. Green, A. Zelazowski, M. J. Still-
man, Eur. J. Biochem. 1995, 227, 226–240.
[49] A. Jerschow, N. Mꢊller, J. Magn. Reson. 1997, 125, 372–375.
[36] J. S. Magyar, T.-C. Weng, C. M. Stern, D. F. Dye, B. W. Rous, J. C.
Payne, B. Bridgewater, A. M. Mijovilovich, G. Parkin, J. M. Zaleski,
Received: December 14, 2010
Published online: March 17, 2011
4428
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4418 – 4428