A polyoxyethylene-substituted bimetallic europium helicate for luminescent staining of living cells

Article abstract of DOI:10.1002/chem.200700883

The homoditopic ligand H₂L^C3 has been designed to form neutral triple-stranded bimetallic helicates of overall composition [Ln₂(L^C3)₃]. The grafting of the polyoxyethylene fragments ensures water solubility and also favors cell penetration while being amenable to further derivatization. The ligand pK _a values have been determined by spectrophotometric titration and range from 3.5 (sum of the first two) to 10.3. The thermodynamic stability of the helicates is large at physiological pH (logβ₂₃ in the range 22-23). The ligand triplet state has an adequate energy (0-phonon transition at ≈20800 cm^-1) for sensitizing the luminescence of Eu^III (Q = 11%). Analysis of the Eu^III emission spectrum points to an overall pseudo D₃ symmetry for the metal environment. No significant effect of [Eu₂(L^C3)₃] is observed on the viability of several cancerous cell lines (MCF-7, HeLa, Jurkat, and 5D10). The cell imaging properties of the Eu^III helicate are demonstrated for the HeLa cell line by luminescence microscopy. Bright Eu^III emission is seen for helicate concentration > 50 μM and after 20-30 min loading time. The helicate stains the cytoplasm and the permeation mechanism is likely to be endocytosis.

Full text of DOI:10.1002/chem.200700883

FULL PAPER
DOI: 10.1002/chem.200700883
A Polyoxyethylene-Substituted Bimetallic Europium Helicate for
Luminescent Staining of Living Cells
Anne-Sophie Chauvin,* Steve Comby, Bo Song, Caroline D. B. Vandevyver,*
FrØdØric Thomas, and Jean-Claude G. Bünzli*[a]
Abstract: The homoditopic ligand
H2LC3 has been designed to form neu-
tral triple-stranded bimetallic helicates
logical pH (logb23 in the range 22–23).
The ligand triplet state has an adequate
viability of several cancerous cell lines
(MCF-7, HeLa, Jurkat, and 5D10). The
cell imaging properties of the EuIII heli-
cate are demonstrated for the HeLa
cell line by luminescence microscopy.
Bright EuIII emission is seen for heli-
cate concentration >50 mm and after
20–30 min loading time. The helicate
stains the cytoplasm and the permea-
tion mechanism is likely to be endocy-
tosis.
energy
(0–phonon
transition
at
of overall composition [Ln
A
ꢀ20800 cmꢁ1) for sensitizing the lumi-
nescence of EuIII (Q=11%). Analysis
of the EuIII emission spectrum points
to an overall pseudo D3 symmetry for
the metal environment. No significant
grafting of the polyoxyethylene frag-
ments ensures water solubility and also
favors cell penetration while being
amenable to further derivatization. The
ligand pKa values have been deter-
mined by spectrophotometric titration
and range from 3.5 (sum of the first
two) to 10.3. The thermodynamic sta-
bility of the helicates is large at physio-
effect of [Eu
2(LC3)3] is observed on the
G
Keywords: bimetallic compounds ·
helicates · luminescence
Introduction
knowledge, the first experiments dealing with the staining of
biological cells with lanthanide complexes date back to 1969
when Scaff et al. treated bacterial smears (E. coli cell walls)
with aqueous ethanolic solutions of europium thenoyltri-
fluoroacetonate and observed the red emission of EuIII in
continuous mode and under mercury lamp excitation.[9] The
usefulness of lanthanide probes for eliminating the auto-
fluorescence background by time-resolved detection was
demonstrated with a TbIII conjugate, formed from a diethy-
lenetriaminepentaacetate (dtpa) chelate sequentially reacted
with 4-aminosalicylate and dioleoylphosphatidylethanola-
mine; this probe has been used successfully as a membrane-
staining agent for morphological studies of 3T3 cultured
cells.[10] Brightly luminescent TbIII macrocyclic chelates have
then been proposed by Bornhop and co-workers as tissue-
selective markers.[11–13] Presently, novel highly luminescent
EuIII complexes are being tested for cell imaging,[14–16] and
in-cellular determination of analytes.[17,18] Increasing the
number of detectable probes on a single sample, for instance
by designing stains with tunable emission wavelengths and
simultaneously tunable excited-state lifetimes, is potentially
appealing. In this context, bimetallic functional molecular
edifices[19] or peptide-linked lanthanide tags[1] may therefore
combine two luminescent, two magnetic, or one magnetic
and one luminescent centers in a single dual probe.
Chemical tools are essential for the study of in-cellular phe-
nomena and for biomolecular imaging and sensing.[1,2] In
this perspective, the versatile luminescent properties of tri-
valent lanthanide ions are presently the subject of a re-
newed interest.[3–8] Indeed the sharp, characteristic emission
lines of the LnIII ions, which extend from UV to visible and
near infrared, are easily recognizable. In addition, the life-
times of the corresponding excited states are in the micro-
or millisecond range, so that time-resolved measurements
are feasible with unsophisticated instrumentation, which
allows one to eliminate autofluorescence of the biological
materials and/or to reach high signal-to-noise ratios. To our
[a] Dr. A.-S. Chauvin, S. Comby, Dr. B. Song, Dr. C. D. B. Vandevyver,
F. Thomas, Prof. Dr. J.-C. G. Bünzli
Laboratory of Lanthanide Supramolecular Chemistry
École Polytechnique FØdØrale de Lausanne (EPFL)
LCSL-BCH 1401 (Switzerland)
Fax : (+41)21-693-9825
E-mail: anne-sophie.chauvin@epfl.ch
caroline.vandevyver@epfl.ch
jean-claude.bunzli@epfl.ch
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2007, 13, 9515 – 9526
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9515

The methodology we pursue towards the latter strategy is
to self-assemble lanthanides ions into triple-stranded bimet-
allic helical structures,[20,21] amenable to functionalization
and biological coupling. The insertion of metal ions into hel-
icates results in their spectroscopic and magnetic properties
not only being fully retained due to the protective wrapping
of the ligand strands around the metal ions, but even en-
hanced and in any case tunable. In addition, the regular ar-
rangement of the metal ions along the same axis may result
in unusual properties such as directional energy transfer.[22]
During the past years, our laboratory has concentrated on
the building up of a library of hexadentate ditopic ligands
with bis(benzimidazole)pyridine cores. This molecular
design is tailored to induce nine-coordinate, tricapped trigo-
nal prismatic environments around the 4f ions similar to the
one found in aqua ions. One series of these potential hosts
are unsymmetrical and feature two slightly different triden-
tate coordinating units. They are therefore tailored for the
simultaneous recognition of a heteropair of lanthanide
ions,[23–25] with the aim of developing hetero dual probes. For
the time being, this approach is, however, limited to organic
solvents. The other series of building blocks are symmetrical
and bear two carboxylic acid functions. A first example is
ligand H2LC (Scheme 1) which self-assembles in water to
ence of the polyoxyethylene groups should also help avoid-
ing potential stacking of the luminescent tags.[8] We present
the adaptable synthetic procedure leading to the isolation of
the ditopic receptor, as well as a full thermodynamic and
photophysical study of the resulting helicates. Finally, the cy-
totoxicity of the [Eu
2(LC3)3] helicate towards four different
G
cancerous cell lines is tested, as well as the usefulness of this
emissive bimetallic tag for luminescence microscopy of
HeLa cells.
Results and Discussion
Ligand synthesis and properties: The main challenge of the
chemical synthesis is the sequential introduction of adequate
functional groups on the benzimidazole moieties (here the
polyoxyethylene chains) and of the carboxylic acid functions
in ortho position of the pyridines, so that several strategies
have been explored with various successes. The polyoxyeth-
ylene pendant arms can be grafted either by reacting the
brominated derivative with the amine functions of precursor
5, which beforehand are deprotonated with sodium hydride,
or through a Mitsunobu reaction with the alcohol derivative
of the arms in presence of azodicarboxylic acid diisopropyl
ester (DIAD) and triphenylphosphine. We also tested two
oxidative pathways for building the carboxylic acid func-
tions, using various oxidants, starting from either 5c fitted
with methyl alcohol groups or the methylated intermediate
5d.
The key reaction in the synthesis is the condensation of
tetramine 3 (see Scheme 2), obtained upon reduction of the
nitro groups of 3,3’-diamino-4,4’-bis(N-aminomethyldiphe-
nyl)methane (2), with a derivative of picolinic acid function-
alized in position 6. The resulting compound must be stable
under the harsh conditions provided by the phosphoric acid
medium at elevated temperature (2108C). In addition, de-
pending on the ortho substituent on the pyridines of inter-
mediate 5, further derivatization proved difficult. For in-
stance, condensation of picolinic acid 4a gave the un-substi-
tuted compound 5a onto which the grafting of two polyoxy-
ethylene pendants was possible, but further attempts to in-
troduce cyano or methyl ortho substituents on the pyridine
rings were unsuccessful. Other routes starting from 4 f (R =
CHO) or 4g (R = CN) led to the decomposition of the re-
action products 5. Moreover, some of the intermediates 5,
which bear two secondary amines, were far from being solu-
ble in organic or aqueous solvents preventing later modifica-
tions; this is exemplified by 5b obtained either from the car-
boxamide (4b) or ester derivative (4c) of picolinic acid. Fi-
nally, direct introduction of the methyl alcohol functions in
the ortho positions of the pyridine rings starting from 6-hy-
droxymethylpyridine-2-carboxylic acid (4d) or from its pro-
tected analogue 4e yielded only traces of the targeted inter-
mediate 5c.
Scheme 1. Symmetrical ditopic ligands for self-assembling bimetallic lan-
thanide helicates in water.
yield very stable neutral homobimetallic [Ln2(LC)3] helicates.
It is noteworthy that the Gibbs free energy associated with
the formation of these structures is able to overcome the
large hydration enthalpy of the lanthanide ions (DHo
ꢀ
hydr
ꢁ3300 to ꢁ3750 kJmolꢁ1). Additionally, the metal ions are
very well shielded from water interaction, as demonstrated
by the long Eu
(5D0) lifetime (2.43ms). [26]
G
We are now developing a new chemistry for the versatile
derivatization of this initial ligand allowing the grafting of
solubilizing, chromophoric or coupling groups on either the
pyridine or the benzimidazole moieties. We have recently
reported a pyridine-derivatized hexadentate ligand forming
a EuIII helicate which diffuses into HeLa cells and which
stains their cytoplasm.[27] Here we replace the ethyl substitu-
ents on the benzimidazole cores of LC by short polyoxyethy-
lene chains featuring three -(CH2)2O- units (H2LC3,
Scheme 1) with the aim of ensuring a better solubility for
the resulting helicates, compared with [Ln2(LC)3]. The pres-
The final strategy adopted was to start by grafting the
polyoxyethylene substituents with the help of a Mitsunobu
reaction on intermediate 5d synthesized from 6-methylpyri-
9516
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 9515 – 9526

Luminescent Staining of Living Cells
FULL PAPER
oxide form being detected at trace levels only. Secondly, the
methyl groups of these products were converted into acetyl
functions by addition of acetic anhydride and then trans-
formed into alcohol in methanolic ammonia, which gave
better yields than aqueous NaOH. Ultimately, conversion of
the alcohol functions into carboxylic acid groups required 30
equivalents of potassium permanganate to ensure optimum
conversion, oxidation of the second alcohol function being
much more difficult than the first one. Under such experi-
mental conditions partial (5–10%) oxidation of the methyl-
ene bridge into a carbonyl moiety occurred, the latter prod-
uct being separated from H2LC3 by column chromatography
followed by preparative HPLC. In summary, the synthesis of
the ditopic hexadentate receptor H2LC3 necessitates nine
synthetic steps, all of them with a yield up to 60% except
the last one, which required extensive purification.
Ligand deprotonation constants were established by spec-
trophotometric titration of H2LC3 1.01”10ꢁ5 m with NaOH at
constant ionic strength (0.1m KCl). The large water solubili-
ty of LC3 resulting from the polyoxyethylene substituents al-
lowed working in the pH range 0.80–12.88 (Figure S1, Sup-
porting Information), which led to the determination of four
of the six pKa values, as well as the sum of the first two
ones. Data were successfully fitted to the following set of
Equations (1–5):
Scheme 2. Synthetic path for the synthesis of key intermediates (5).
dine-2-carboxylic acid (4h), itself obtained by mono-oxida-
tion of lutidine. Direct oxidation of the resulting intermedi-
ate 6 (Scheme 3) proved very difficult and despite many at-
tempts in which the solvent as well as the nature and the
amount of oxidant were varied, mono-aldehyde or mono-
carboxylic acid derivatives were mainly isolated while larger
amounts of oxidant led to decomposition. We therefore de-
cided on a more elaborate three-step route. The first reac-
tion implied activation of the pyridine functions by oxida-
tion with hydrogen peroxide in acetic acid leading to a mix-
ture of di- (7a) and tri- (7b) N-oxide products, the tetra-N-
4þ
2þ
ðH6LC3
Þ
Ð ðH4LC3
Þ
þ 2 Hþ ðpKa1 þ pKa2Þ ¼ 3:5ð1Þ
ð1Þ
2þ
ðH4LC3
Þ
Ð ðH3LC3Þþ þ Hþ
pKa3 ¼ 3:2ð1Þ
pKa4 ¼ 4:3ð1Þ
pKa5 ¼ 7:6ð1Þ
pKa6 ¼ 10:3ð1Þ
ð2Þ
ð3Þ
ð4Þ
ð5Þ
ðH3LC3Þþ Ð ðH2LC3
Þ
þ Hþ
ðH2LC3
Þ
Ð ðHLC3Þꢁ
þ
Hþ
ðHLC3Þꢁ Ð ðLC3
Þ
þ Hþ
2ꢁ
The corresponding distribution
diagram is shown in Figure 1.
For un-substituted H2LC, only
three protonation constants
could be determined because of
solubility
problems
below
pH 6;[28] they were assigned on
the basis of potentiometric,
electronic absorption, fluores-
cence, as well as 1H and
13C NMR data as follows [the
numbering is adapted to Eqs.
(1–5) above]: pKa6 =10.1(1) and
pKa5 =9.5(2), with a statistical
difference of 0.6, correspond to
deprotonation of the imidazolyl
groups, while pKa4 =6.0(2), and
calculated pKa3 =5.4(2) corre-
sponds to the deprotonation of
the pyridyl units. Large differ-
Scheme 3. Synthesis of ligand H2LC3
.
Chem. Eur. J. 2007, 13, 9515 – 9526
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9517

J.-C. G. Bünzli, A.-S. Chauvin, C. D. B. Vandevyver et al.
Figure 1. Distribution diagram for H2LC3 1.01”10ꢁ5 m versus pH, calculat-
ed with the deprotonation constants given in Equations (1–5).
ences with reported pKa values for benzimidazole, 2-(2’-pyri-
dyl)benzimidazole or 1,10-phenanthroline-2,9-dicarboxylic
acid were interpreted in terms of potential hydrogen bond-
ing taking place in the protonated subunits of the ditopic
ligand between the pyridinium proton and either the carbox-
ylate or imidazolium groups.[28] In turn, the differences be-
tween H2LC and H2LC3, DpKa5–3 ꢀ2 could be related to less
stabilization through hydrogen bonds in view of the larger
energy barrier for the trans–cis conformational change
needed for inducing these bonds, owing to the presence of
the bulky polyoxyethylene substituents in H2LC3.
Figure 2. Parts of the electrospray mass spectra of solutions containing
III
2:3stoichiometric amounts of Ln and H2LC3 in water/acetonitrile 90:10,
Ln=Eu (top) and Gd (bottom). Calculated spectra are shown above the
experimental data; [H2LC3]t =3”10ꢁ4 m.
Table 1. Mass spectrometric data for H2LC3 and stoichiometric 2:3solu-
tions (Ln:H2LC3, Ln=La, Eu, Gd, Tb, Lu) in water/acetonitrile 90:10.[a]
m/z (obsd)[b]
m/z (calcd)
D
A
MW [Da][c]
H2LC3
La
Eu
Gd
Tb
Lu
783.3356
1310.3040
1324.3270
1329.3300
1330.3704
1346.3691
783.3354
1310.3820
1324.3949
1329.3949
1330.4010
1346.4164
0.26
60
51
49
23
783.33
2618.74
2646.78
2656.78
2658.79
2690.82
Formation of the helicates in water: The formation of the
[Ln
2(LC3)3] helicates in water is ascertained by electrospray
G
mass spectrometry. Stoichiometric 2:3solutions of a perchlo-
rate salt (Ln=La, Eu, Gd, Tb, Lu) and of the ligand (3”
10ꢁ4 m in water) were prepared and analyzed in presence of
10% of acetonitrile to induce ionization. In all cases, a peak
35
[a] The observed and calculated values refer to the most abundant peak.
[b] [M+H]+ for H2LC3, {[M+2H]2+}/2 for the 2:3helicates. [c] Molecular
weight of the parent species, C123H132N18O30Ln2 for the complexes.
corresponding to a +2 charged species {[Ln
2(LC3)3]+ 2H}2+
C
was observed. Typical examples are shown on Figure 2 in
which the shape of this peak and its isotopic distribution are
shown to be in agreement with theoretical simulations. Nu-
merical data for molecular peaks corresponding to the 2:3
helicates for all investigated lanthanides are listed in
Table 1. Accuracy of the experimental m/z measurements is
in the range 20–60 ppm which is satisfying taking into ac-
count that the average mass accuracy achievable with the
spectrometer used is 5 ppm (over the mass range extending
from 400–900 Da). No species corresponding to the 1:3or
2:2 complexes could be evidenced, pointing to the 2:3heli-
cates being probably the major species in solution.
niques to the following set of equations (protons and charg-
es are omitted):
3 LC3 þ Ln
3 LC3 þ 2 Ln
LC3 þ 2 Ln
Ð
Ð
Ð
½LnðLC3Þ3Š logb13
½Ln2ðLC3Þ3Š logb23
½Ln2ðLC3ފ logb21
ð6Þ
ð7Þ
ð8Þ
Residuals were 1.4”10ꢁ4 (La), 5.7”10ꢁ4 (Eu), and 3.9”10ꢁ4
(Lu). Spectra are heavily correlated (Figure S2, Supporting
Information) so that the conditional constants collected in
Table 2 have to be taken with care. Comparing with the
logb23 values published for the helicates with H2LC, the in-
troduction of the polyoxyethylene substituents in H2LC3 is
clearly detrimental to the overall stability of the 2:3com-
plexes. However, the stability of the latter remains large
To quantify the LnIII interaction with H2LC3, we have per-
formed spectrophotometric titrations of the ligand 10ꢁ5 m by
lanthanide perchlorate solutions 5”10ꢁ3 m, up to a total con-
centration ratio R=[Ln] /
t [H2LC3]t =4. The UV/Vis spectra
A
display a well defined isobestic point at 322 nm (Figure 3,
top) and evolving factor analysis usually points to the pres-
ence of four absorbing species. Several models were tested,
but data were best fitted by non-linear least-squares tech-
enough to envisage in vitro applications. Indeed, as shown
III
in Figure 3(bottom) for Eu
and a total concentration of
9518
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 9515 – 9526

Luminescent Staining of Living Cells
FULL PAPER
zole to pyridine and carboxylic groups, respectively, accord-
ing to CAChe calculations. Upon complexation, the main
band broadens and sustains a bathochromic shift of about
15 nm, to 326 nm (e=80000mꢁ1 cmꢁ1 for the 2:3species).
Upon excitation into the lowest energy band, emission from
1
the singlet pp* state is seen as a two-component band with
maxima at 376 and 403 nm (Figure 4 and Figure S3, Sup-
Figure 4. Photophysical properties of H2LC3 and its Ln/H2LC3 2:3solu-
tions with EuIII and TbIII at room temperature in Tris-HCl 0.1m, total
ligand concentration 4.5”10ꢁ5 m. Left: absorption (c) and excitation
(g) spectra; right: emission spectra (lex =310–326 nm) recorded with-
out delay, except for 3pp* (50 ms delay, 77 K). Vertical scale: arbitrary
units; spectra are not drawn to scale.
Figure 3. Top: UV/Vis spectra of H2LC3 10ꢁ5 m in water at pH 7.4 (Tris-
HCl 0.1m) in presence of increasing amounts of Eu(ClO4)3 (R=
A
[Eu]t/
[H2LC3]t =0–4). Bottom: distribution diagram drawn with the condi-
G
tional constants reported in Table 2 and for [H2LC3]t =10ꢁ4 m.
Table 2. Conditional stability constants (pH 7.4, Tris-HCl 0.1m; 258C) for
the systems Ln/H2LC3 compared to logb23 for Ln/H2LC (data from ref.
[26]).[a]
porting Information), which sustains a blue shift of about
40 nm upon decreasing the temperature to 77 K. At the
latter temperature, time-resolved measurement reveals a
broad band with a maximum at 478 nm (0-phonon compo-
nent at 453nm) and which extends up to 650 nm, with a vi-
bronic progression of about 1500 cmꢁ1 typical of aromatic
compounds (ring breathing mode). The emission spectra of
L
Constant
La
Eu
Lu
[b]
LC3
logb13
logb23
logb21
logb23
16.8(1)
23.4(1)
11.4(1)
26.1(4)
17.0[c]
23.1(1)
11.0(1)
27.3(6)
22.7(2)
10.9(1)
30(1)
LC
[a] Standard deviations are given within parentheses. [b] For La, a model
with three absorbing species gave the best results. [c] Value fixed, other-
wise convergence was not reached
III
the 2:3complexes with La and LuIII are very similar, with
minor shifts, while the phosphorescence spectrum of the
GdIII helicate displays a marked difference in the intensity
of the vibronic components.
10ꢁ4 m in ligand (i.e., similar to the one used for cell staining
experiments), the solution contains only 3% of free ligand
and 3% of 2:1 complex, while the helicate is the major spe-
cies in solution (82%). For a millimolar solution the latter
proportion reaches 91% and uncomplexed ligand accounts
for less than 0.5%. These findings are consistent with the
ESI-MS data reported above.
The triplet state luminescence decay at 77 K is bi-expo-
nential, in line with the presence of two aromatic systems on
the ligand. For unbound H2LC3, the corresponding lifetimes
are quite long, 50 ms (15% contribution to the decay) and
430 ms (85%). The long lifetime decreases to approximately
300 ms for the helicate with LaIII and to 200 ms for the other
systems (Ln=Gd, Lu). The shorter lifetime is not much af-
fected in the LaIII system but decreases to 5 and 17 ms for
GdIII and LuIII, respectively. These lifetimes are substantially
shorter than those reported previously for H2LC for which
only the longer lifetime could be extracted from the lumi-
nescence decays: 776ꢂ10 ms, and its helicates with LaIII
(246 ms) and LuIII (181 ms),[26] a fact which we trace back to
the presence of the polyoxyethylene substituents which
confer more fluxionality to the ligand. At room tempera-
Ligand-centered photophysical properties: The uncom-
plexed ligand H2LC3 displays two main absorption bands
(Figure S1, Supporting Information) centered at 245 nm and
310–315 nm, depending on pH, and with a shoulder at 330–
340 nm; molar absorption coefficients e for the deprotonat-
ed form (LC3)2ꢁ are 21550 and 31600mꢁ1 cmꢁ1, respectively.
These bands are essentially generated by p!p* transitions
involving intramolecular electron transfer from benzimida-
Chem. Eur. J. 2007, 13, 9515 – 9526
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9519

J.-C. G. Bünzli, A.-S. Chauvin, C. D. B. Vandevyver et al.
ture, the 3pp* emission disappears for the helicates with
4.0ꢂ0.1 ms. This clearly points to no water molecule inter-
acting in the first coordination sphere for the main species,
q=0 calculated using a known phenomenological equa-
tion,[35] much as for the helicate with H2LC (t=2.43and
4.66 ms).[26] Contribution from the long-lived species to the
overall luminescence emission amounts to ꢀ90%, in line
with the speciation determined by UV/Vis titration. The
second lifetime corresponds to a chemical species with a
population of ꢀ10% and for which, on average, q=1.4.
1
EuIII and TbIII, and pp* emission is considerably weakened
for EuIII, but not for TbIII (Figure 4). Conversely, a faint re-
sidual triplet state emission remains at 77 K while singlet
state emission is still sizeable for both helicates (Figures S3,
S4, Supporting Information).
Lanthanide-centered photophysical properties: As shown on
Figure 4, the EuIII and TbIII helicates mainly display the
characteristic metal-centered emission lines from the Eu-
5
7
The TbIII emission spectrum is dominated by the D4! F5
A
G
transition and the Tb
(5D4) lifetime is quite short at room
A
Energy transfer from the ligand is further established by the
excitation spectra matching the absorption spectrum of the
ligand.
temperature, 0.39ꢂ0.04 ms. Since this lifetime increases
considerably when the temperature is lowered, reaching
2.4ꢂ0.1 ms at 77 K, the non-radiative de-activation it re-
flects is not due to coordination of water molecules but,
rather, to a back transfer mechanism since the ligand triplet
state has an energy very close to the TbIII excited levels.
This phenomenon is, however, less pronounced than in the
The [Eu
A
arising from a main species with a single chemical environ-
5
7
ment displaying pseudo D3 symmetry: the D0! F0 transi-
tion appears as a very symmetrical, faint band (Figure S5,
7
top, Supporting Information) and the 5D0! FJ transitions
corresponding compound with H2LC for which the Tb(5D4)
A
display two, two and four main components for J=1, 2, and
4, in line with group-theoretical predictions for this symme-
try (0, 2, 2, and 4 allowed transitions for J=0, 1, 2, and 4).[29]
lifetime decreased from 2.10 ms at 100 K to 0.05 ms at
270 K, an effect caused by the slight hypsochromic shift of
the ligand energy levels upon functionalization with the
polyoxyethylene substituents.
7
The splitting of the F1 level is also in line with a trigonal
5
7
symmetry: the D0! F1 transition displays a sharp band cor-
Finally, the quantum yields of the EuIII and TbIII lumines-
cence have been measured upon ligand excitation in Tris-
HCl 0.1m by two different experimental techniques (see Ex-
perimental Section): QEu is equal to 11ꢂ1% and QTb to
7
responding to transition to the F1(A1) ligand-field sub-level
7
and a broader one, terminating on the F1(E) sublevel; the
A1–E separation amounts to 150 cmꢁ1, reflecting a B2 crys-
0
L
L
tal field parameter of about ꢁ600 cmꢁ1, using the correlation
given by Binnemans et al.[30] Moreover, the broadness of the
transition to the E component points to a probable splitting
of this sub-level, the extent of which is indicative of the de-
viation from the idealized symmetry;[31] it can be estimated
here to be ꢀ45–50 cmꢁ1. Comparable data for the helicate
with H2LC were 156 cmꢁ1 for the A1–E separation and
56 cmꢁ1 for the splitting of the E level (solid state, 10 k).
Therefore, we conclude that the metal ion environments are
very similar in the two cases. The only noticeable spectro-
0.34ꢂ0.04%, consistent with the back transfer phenomenon
ACHTREUNG
evidenced by the temperature dependence of the Tb
(5D4)
lifetime. The ligand efficiency for the sensitization of the
EuIII luminescence, hsens, can be estimated from:[36]
QEu ¼ hsens ꢃ QEu
¼ hsens ꢃ ðtobs=trad
Þ
ð9aÞ
L
Eu
where QEu is the intrinsic quantum yield (upon direct
Eu
metal excitation) and tobs and trad are the observed and radi-
ative lifetimes, respectively; the latter is calculated from:
5
scopic difference is the intensity of the hypersensitive D0!
7F2 transition: intensity ratios amount to 0.01, 1.0, 1.45, 0.11,
1.75, 0.06, and 0.23for J=0–6, whereas they were 0.01, 1.0,
0.89, 0.11, and 1.69 (J=0–4) for the helicate with H2LC.
Since the hypersensitive transition contains sizeable contri-
butions from vibronic components,[32] the larger intensity of
this transition in the presently reported compound is again
understandable in view of the presence of the polyoxyethy-
lene chains. Another argument in favor of a chemical envi-
ronment similar to the one in the H2LC helicate is the
1=trad ¼ 14:65 ꢃ ðItot=IMDÞ ꢃ n3 ½sꢁ1
Š
ð9bÞ
where Itot/IMD is the ratio of the total emitted intensity to the
5
7
intensity of the magnetic dipole transition D0! F1. In our
case it is equal to 4.69 and n=1.334, so that trad =6.13ms,
QEuEu =37% and therefore hsens =30%.
Some light on the sensitization process is shed by examin-
ing the ratios of the integrated total lanthanide lumines-
cence (J=0–6 for EuIII or 6–0 for TbIII) to the singlet and
triplet state emissions. For the europium helicate, the
5D0/1pp* ratio decreases from 17.4 at room temperature to
1.4 at 77 K; triplet state emission is only seen at the latter
7
energy of the 5D0! F0 transition, which amounts to
17235 cmꢁ1. Using a known equation[33] and nephelauxetic
parameters dcarb =ꢁ17.2 cmꢁ1 for the carboxylate and dbzp
=
ꢁ15.3cm ꢁ1 for heterocyclic nitrogen atoms,[34] we find
n˜calcd =17231 cmꢁ1, quite close to the experimental value.
When the luminescence decays are sampled with short
time intervals, the functions are bi-exponential reflecting
temperature with a D0/3pp* ratio equal to 2.6. This reflects
5
3
the moderately efficient pp*-to-5D0 energy transfer (hsens
=
30%), in line with an energy difference between the ligand
0-phonon component of the triplet state and the europium
emissive state of about 5000 cmꢁ1. For TbIII the gap is much
smaller (1500 cmꢁ1) and back transfer occurs so that the
overall ligand-to-5D4 energy transfer is much less important,
lifetimes for the Eu
A
is only slightly temperature dependent (2.6ꢂ0.1 ms at 77 K)
and 0.54ꢂ0.05 ms. In D2O, only one lifetime is evidenced,
9520
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 9515 – 9526

Luminescent Staining of Living Cells
FULL PAPER
as demonstrated by the 5D4/1pp* ratio which is small, in-
creasing from 0.40 to 2.0 between room temperature and
77 K. Moreover, the energy difference between the 0-
1
3
phonon lines of the pp* and pp* states (ꢀ6800 cmꢁ1) is
somewhat over the ideal 5000 cmꢁ1 gap for an efficient inter-
system crossing,[37] consistent with the ratio pp*/3pp*=1.9
1
for EuIII, and 2.2 for TbIII at 77 K. In fact, when the temper-
ature is lowered to 77 K, the band envelopes of the ligand
emission are noticeably displaced toward the blue with, as a
consequence, a dramatic reduction in the D0/1pp* emission
5
ratio. This is due to the overlap integral between the ligand
emission spectrum and the EuIII absorption spectrum being
5
reduced. The large decrease in the D0/1pp* ratio concomi-
1
tant with the bathochromic shift of the pp* emission may
be indicative that some singlet state energy transfer occurs
5
(e.g. to the L6 level located on the high energy side of the
1
ligand pp* emission). In the case of TbIII, a different trend
5
is observed, the D4/1pp* ratio increasing upon lowering the
5
temperature, while the D4/3pp* ratio is comparable to the
one observed for EuIII. We trace this back to an increase in
the energy gap between the ligand and the metal ion leading
to a substantial reduction in the Tb
(5D4)-to-3pp* back trans-
G
1
fer. As for EuIII, participation of the pp* level to the over-
all energy transfer onto the metal ion is plausible.
Figure 5. Top: WST-1 proliferation test of HeLa cells in absence or pres-
ence of different concentrations of [Eu2
(LC3)3]. Bottom: Cell viability
N
Cytotoxicity of the EuIII-containing solutions: The solutions
values (%), using the same test versus the helicate concentration, after
24 h incubation at 378C.
containing the [Eu
2(LC3)3] helicate stain living cells but
A
before describing luminescence imaging experiments, we ad-
dress two key questions: i) do the cells remain viable in
presence of these solutions over the period of examination
under the microscope, and ii) does the helicate survive un-
dissociated in the cells? Influence of the EuIII-containing
solutions on the proliferation of several malignant cell lines,
5D10 mouse hybridoma, Jurkat human T leukemia, MCF-7
human breast carcinoma and HeLa cervical adenocarcino-
ma, was determined by means of the WST-1 assay at time
intervals between 0.5 and 24 h. Results for the HeLa cell
line are shown on Figure 5 (top). No significant differences
could be observed in the proliferation of the cells in absence
or presence of the helicate, up to 500 mm. This observation is
confirmed by the viability of the cells after 24 h incubation
Table 3. Cell cytotoxicity values using the LDH assay at 228C and after
30 min incubation of the cells with different concentrations of [Eu2-
(LC3)3].
c [mm]
Jurkat
5D10
MCF-7
125
250
500
ꢁ1.8ꢂ0.1
ꢁ2.2ꢂ2.5
ꢁ0.1ꢂ1.8
ꢁ2.9ꢂ1.0
ꢁ5.9ꢂ1.8
ꢁ1.0ꢂ6.9
ꢁ2.6ꢂ0.7
ꢁ3.7ꢂ2.0
ꢁ2.2ꢂ1.9
both cases with n˜ =17235 cmꢁ1 and a full width at half
height of 22 cmꢁ1, confirming once more the presence of a
single major species in these solutions. Secondly, the crystal
field splitting of the species in water and in the HeLa cells is
5
7
at 378C with various concentrations of the 2:3(Ln/H LC3)
the same, and the intensity ratios of the D0! FJ transitions
are similar except for a slight reduction in the intensity of
the hypersensitive transition. Finally, the lifetimes of the Eu-
2
solutions (Figure 5, bottom) and similar results were ob-
tained for the other cell lines, with viabilities usually ꢀ90%
(Figure S6, Table S2, Supporting Information). In addition,
the cellular membrane damage assessed by measuring the
lactate dehydrogenase (LDH) leakage 30 min after incuba-
tion with the helicate solution is very small for the cell lines
tested (see Table 3). All these data show that the Eu species
in the solution, particularly the EuIII helicate, can be consid-
ered as non-cytotoxic under the experimental conditions
used for luminescence cell imaging.
A
aqueous solution: 2.0ꢂ0.2 and 0.58ꢂ0.06 ms, respectively.
The contribution of the species corresponding to the latter
lifetime is somewhat larger in the cells (ꢀ15%), but the
smaller emission intensity renders the bi-exponential fit
more difficult and the observed increase is in the experi-
mental error. From these results, we conclude that the speci-
ation hardly changes in the cells and, more specifically, that
the helicate remains un-dissociated in the HeLa cells under
the experimental conditions used.
Secondly, the emission spectrum of the [Eu
2(LC3)3] heli-
G
cate localized in HeLa cells is almost identical with the spec-
tra measured in the aqueous solution at pH 7.4 (Figure S5,
Supporting Information). Firstly, a high resolution scan of
Luminescence microscopy of HeLa cells: HeLa cells grown
on glass bottom cell culture dishes were incubated in cell
5
7
the D0! F0 transition reveals a single, symmetric band in
Chem. Eur. J. 2007, 13, 9515 – 9526
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9521

J.-C. G. Bünzli, A.-S. Chauvin, C. D. B. Vandevyver et al.
culture medium RPMI-1640 containing the EuIII helicate at
various concentrations. Cells were examined with a lumines-
cence microscope 7 h after complex loading. The [Eu
2(LC3)3]
G
helicate clearly permeates into the cells and stains their cy-
toplasm in a concentration-dependent manner (Figure 6 top
and S7, Supporting Information). Loading concentrations as
low as 50 mm are sufficient to generate reasonably bright
images. A counter staining experiment was undertaken in
order to confirm the localization of the helicate. Live cells
were loaded with the helicate and subsequently with acri-
dine orange, which is a nucleic-acid selective metachromic
stain useful for cell cycle determination. Acridine orange is
known to permeate into the cell nuclei.[38] The luminescence
images of Figure 6 (bottom) clearly confirm the localization
of the europium complex in the cytoplasm.
In order to define the mechanism by which the cells take
up the europium helicate, the time course and temperature
dependence of the complex loading were determined. HeLa
cells were incubated with a 500 mm solution of [Eu
2(LC3)3] at
U
378C and examined by luminescence microscopy from
15 min to 6 h after complex loading. The first bright spots in
the cytoplasm of the cells can be observed as early as 20–
30 min after incubation with the helicate (Figure 7, top, and
Figure S8, Supporting Information).
The presence of vesicular structures in the cytoplasm of
the cells could support an endocytotic uptake mechanism. It
is known that such a mechanism can be blocked at 48C. In a
Figure 7. Top: Time dependence of the luminescence intensity of HeLa
cells loaded with [Eu2
(LC3)3] 500 mm at 378C. Bottom: Comparison of the
A
luminescence intensity from HeLa cells loaded with [Eu2
(LC3)3] 250 mm at
G
4 and 378C.
separate experiment, the cells
were therefore loaded with a
250 mm solution of the [Eu2-
A
ture medium at 48C for 6 h. No
europium luminescence could
be observed at 48C despite the
good viability of the cells at this
temperature (Figure 7, bottom
and Figure S9, Supporting In-
formation), indicating that the
complex is likely to enter the
cells by this mechanism.
Conclusion
In this paper, we propose a suc-
cessful synthetic approach for
the derivatization of the benzi-
midazole moiety of the sym-
metric, ditopic dicarboxylic
ligand H2LC to yield H2LC3.
This strategy allows the grafting
of polyoxyethylene fragments
which are amenable to further
derivatization, for instance, for
tumor-targeting purposes. We
also demonstrate that the re-
sulting triple stranded helicates
Figure 6. Top: Luminescence images of HeLa cells loaded with different concentrations of [Eu2
ACHTREUNG
RPMI-1640 for 7 h at 378C (exposure time: 60 s). Bottom: Counter staining experiment with [Eu
ACHTREUNG
and acridine orange (AO, green); exposure time: 1 s for Eu and 15 ms for AO.
9522
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 9515 – 9526

Luminescent Staining of Living Cells
FULL PAPER
[Ln2
(LC3)3] are the main species in solutions with total
N
modeling was performed with the CAChe workpackage 7.5 (Fujitsu,
2000–2006). IR spectra were obtained on a Spectrum One Perkin Elmer
FT-IR spectrometer equipped with an ATR accessory. FIR spectra were
recorded with an IFS 13v FT-IR spectrometer from Bruker equipped
with a pyroelectric DTGS detector. Raman spectra were obtained by
means of a Spex 1403spectrometer with RCA photomultiplier and an
argon laser Stabilite 2017 from Spectra-Physics. Elemental analyses were
performed by Dr. Eder, Microchemical Laboratory of the University of
Geneva.
Preparation of ligands (see Schemes 2 and 3): H2LC used for comparison
purposes was prepared as previously described.[26] The syntheses of 2 and
3 were adapted from the procedure described for 3,3’-dinitro-4,4’-bis(N-
aminomethyldiphenyl)methane;[42] 6-methyl-pyridine-2-carboxylic acid
4b was prepared in 60% yield according to Black et al.[43]
ligand concentration 10ꢁ4 m. In addition, the EuIII-containing
2:3stoichiometric solution possesses the required character-
istics for its potential use as a luminescent marker for bio-
logical analyses and imaging. It is non-cytotoxic towards sev-
eral cancerous cell lines, the thermodynamic stability of the
main helicate species is large enough at physiological pH,
and [Eu
2(LC3)3] easily permeates into the cytoplasm of HeLa
G
cells. The helical edifice remains intact in the cells and the
required concentration and incubation time are compatible
with practical cell-imaging experiments. The presence of
other minor species in the solutions used for the latter ex-
periments (ꢀ10%) is not detrimental since free ligand and
metal ion are in trace quantities only and since the second
species of some importance is the 1:3complex which also
provides a protective chemical environment for the metal
ion. For the time being the only drawback of this lumines-
cent staining solution is its relatively short excitation wave-
length, 330 nm, that can be extended to 365 nm by choice of
a judicious excitation filter. After this proof of principle, ex-
periments are now in progress in our laboratory to extend
the excitation wavelength into the visible range by adequate
derivatization. We are also planning to stain other cell lines,
the use of time-resolved microscopy, and targeting experi-
ments.
3,3’-Dinitro-4,4’-diaminodiphenylmethane (2): Nitro-2-aminobenzene
(12.5 g, 9.05 mmol) and p-formaldehyde (1.35 g, 4.52 mmol) were heated
in 25% HCl (200 mL) for 5 h. The orange-yellow solution was then
poured in water (600 mL) and neutralized with 25% aqueous ammonia
(pH 10). The precipitate was filtered, washed with water, and dried under
vacuum to give a yellow-orange solid (23.4 g, 90%). 1H NMR (CDCl3):
d=7.93(s, 1H, H ar), 7.16 (d, 1H, Har), 6.75 (d, 1H, Har), 3.81 ppm (s, 1H,
-CH2-); ESI-MS: m/z: calcd for: 289.10; found: 289.53[ M+H]+; elemen-
tal analysis calcd (%) for C13H12N4O4·1.3H2O: C 50.10, H 4.72, N 17.98,
found: C 50.13, H 4.46, N 17.93.
3,3’,4,4’-Tetraaminodiphenylmethane (3): Compound 2 (3g, 10.4 mmol)
was mixed with zinc (10.9 g, 166 mmol) and refluxed in ethanol (120 mL)
under N2; 30 mL of aqueous NaOH 0.1m were then added dropwise
under vigorous stirring. After 1 h the orange solution should turn to pale
yellow, otherwise additional amounts of NaOH have to be added. After
additional 2 h refluxing, the solution was filtered and the filtrate evapo-
rated under reduced pressure. The crude product was poured into water
(100 mL), filtrated, rinsed with water until neutral pH and dried under
vacuum (2 mbar) at RT. The brownish solid was stored under vacuum
1
and should be used without delay (6 g, 70%). H NMR ([D6]DMSO): d=
Experimental Section
6.37 (d, 1H, Har), 6.77 (s, 1H, Har), 6.17 (d, 1H, Har), 3.39 ppm (s, 1H,
-CH2-); ESI-MS: m/z: calcd for: 229.15; found: 229.47 [M+H]+.
Starting materials and general procedures: Chemicals and solvents were
purchased from Fluka. Solvents were purified by a non-hazardous proce-
dure by passing them onto activated alumina columns (Innovative Tech-
nology Inc. system).[39] Stock solutions of lanthanides were prepared just
before use in freshly boiled, doubly distilled water from the correspond-
Compound 5d: Freshly prepared 3 (1.3g, 5.6 mmol) was introduced with
6-methylpyridine-2-carboxylic acid (4b; 1.6 g, 11.6 mmol) and polyphos-
phoric acid (20 mL) in a two-neck 50 mL flask equipped with a thermom-
eter and a distillation apparatus. The mixture was heated with a DrySyn
apparatus at 1808C, all the water removed by distillation and the temper-
ature then increased to 2108C. After stirring 6 h, the temperature was de-
creased to approx. 608C and the viscous brown residue poured into etha-
nol (150 mL). A green solid formed, which was filtrated, poured in 0.1m
NaOH (100 mL) and the pH was adjusted to 13–14 by addition of 5m
NaOH. The brown solid was filtrated, rinsed with water, and dried at
708C under vacuum (2 mbar) for 12 h (2 g, 82%). 1H NMR
([D6]DMSO): d=8.18 (d, 1H, Har), 7.77 (dd, 1H, Har), 7.48 (d, 1H, Har),
7.43(s, 1H, H ar), 7.25 (d, 1H, Har), 7.13(d, 1H, H ar), 3.17 (s, 1H, -CH2-),
2.55 ppm (s, 3H, -CH3); 13C NMR ([D6]DMSO): d=158.12, 152.83,
149,59, 140.92, 140.20, 137.70, 136.00, 123.70, 123.32, 118.78, 116.33,
115.45 ppm; elemental analysis calcd (%) for C27H22N6·1.5NaOH·2H2O:
C 61.59, H 5.59, N 15.96, found: C 61.54, H 5.40, N 16.46; MALDI-TOF-
MS: m/z: calcd for [M]+: 430.19; found: 430.19.
ing Ln(ClO4)3·xH2O salts (Ln=La, Lu, Gd, Tb, Eu, x=2.5–4.5). These
G
salts were prepared from their oxides (Rhône-Poulenc, 99.99%) in the
usual way.[40] Concentrations of the solutions were determined by com-
plexometric titrations using a standardized Na2H2EDTA solution in uro-
tropine buffered medium and with xylenol orange as indicator.[41]
Analytical measurements: NMR spectra were measured at 258C on
Bruker Avance DRX 400 (1H, 400 MHz) and AV 600 (13C, 99.8 MHz)
spectrometers. Spectra of organic compounds were recorded in CDCl3
(99.8%, Aldrich), MeOD (99.8%, Aldrich) or [D6]DMSO (99.8 Aldrich)
and those of the helicates in D2O (99.9%, Aldrich); deuterated solvents
were used as internal standards and chemical shifts are given with respect
to TMS. The ESI-MS spectra of the ligands were obtained on a Finningan
SSQ 710C spectrometer using 10ꢁ5–10ꢁ4 m solutions in acetonitrile/H2O/
acetic acid (50:50:1) or MeOH, capillary temperature 2008C and acceler-
ation potential 4.5 keV. The instrument was calibrated using the horse
myoglobin standard and the analyses were conducted in positive mode.
ESI-QTOf MS spectra of the complexes were measured on a Q-Tof
Ultima API mass spectrometer (Micromass, Manchester, UK) equipped
with a Z-spray type ESI source. Phosphoric acid was used for the positive
ion mass calibration range of 100–2000 m/z. Data were acquired and
processed using Masslynx version 4.0. Electrospray conditions were as
follows: capillary voltage, 2.3kV; source temperature, 80 8C; cone volt-
age, 35 V; and source block temperature, 1508C. The ESI nebulization
and drying gas was nitrogen. All experiments were performed in positive
ion mode. The sample was introduced via a syringe pump operating at
20 mLminꢁ1. Simulation of spectra was achieved with Molecular Weight
Calculator 6.42. UV-visible spectra were measured in 0.2 cm quartz Su-
prasil cuvettes on a Perkin Elmer Lambda 900 spectrometer. Molecular
Compound 6: Diamine 5d (2 g, 4.65 mmol) was dissolved in THF
(200 mL) in presence of triethyleneglycolmonomethylether (4.6 mL,
29.8 mmol), and triphenylphosphine (7.8 g, 29.8 mmol). Azodicarboxylic
acid diisopropyl ester (5.8 g, 29.8 mmol) was added dropwise at room
temperature. The orange solution was refluxed for 12 h then the solvent
was removed under reduced pressure, and the residue was re-dissolved in
dichloromethane (150 mL), washed with aqueous HCl 0.1m (100 mL)
and with water (2”100 mL). The organic phase was dried over Na2SO4
and evaporated and the resulting crude solid was purified by column
chromatography (silica gel; CH2Cl2/MeOH 100 ! 97:3) to give the dis-
ubstituted product 6 as an orange oil (2 g, 64%). 1H NMR (CDCl3): d=
8.16 (d, 1H, Har), 7.68 (m, 2H, Har), 7.43(d, 1H, H ar), 7.17 (m, 2H, Har),
4.93(dd, 2H, -OCH 2-), 4.28 (d, 1H, -OCH2-), 3.98 (m, 2H, -CH2-), 3.54–
3.42 (dm, 9H, -OCH2-), 3.31 (s, 3H, -OCH3), 2.60 ppm (s, 3H, -CH3);
Chem. Eur. J. 2007, 13, 9515 – 9526
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9523

J.-C. G. Bünzli, A.-S. Chauvin, C. D. B. Vandevyver et al.
13C NMR (600 MHz, CDCl3): d=157.30, 150.19, 150.08, 149,99, 149.85,
142.73, 141.08, 137.42, 136.68, 136.18, 135.73, 124.55, 124.17, 124.10,
123.03, 122.99, 121.54, 121.50, 119.67, 110.65, 61.70, 58.99, 58.92, 45.53,
45.39, 42.83, 42.54, 24.40 ppm; ESI-MS: m/z: calcd for [M+H]+: 723.38;
found: 723.38, 362.33 [M+2H]2+/2; elemental analysis calcd (%) for
C41H50N6O6·H2O: C 66.47, H 7.07, N 11.34, found: C 66.16, H 7.33, N
11.78.
solution was measured by a KCl-saturated electrode and the UV/Vis ab-
sorption spectrum was recorded using a 1 cm Hellma optrode immersed
in the thermostated titration vessel and connected to the Tidas spectrom-
eter. Measurements were conducted in the pH range 0.80–12.88. Using
the same equipment, conditional stability constants were determined by
titration of H2LC3 by LnIII (Ln=La, Eu, Lu) at fixed pH 7.4 (0.1m Tris-
HCl buffer). Factor analysis[45] and mathematical treatment of the spec-
trophotometric data were performed with the Specfit software.[46,47]
Compounds 7a + 7b: Compound 6 (200 mg, 0.28 mmol) was dissolved
in acetic acid (10 mL) and hydrogen peroxide 30% (1.6 g) was added
dropwise. After refluxing 4 h, an additional amount of H2O2 was added
(1.6 g, 50 equiv) and the solution was further refluxed for 4 h. The reac-
tion was followed by mass spectrometry, giving a mixture of di- and tri-
N-oxide compounds; ESI MS: m/z: di-N-oxide: calcd for [M+H]+:
755.38; 755.32, 378.31 [M+2H]2+/2; tri-N-oxide: calcd for [M+H]+:
771.37; found: 771.34, 386.19 [M+2H]2+/2. The solvent was carefully re-
moved and the crude product was dried under vacuum for 2 d (210 mg).
CAUTION! Hydrogen peroxide combined with organic compounds is po-
tentially explosive and should be handled in small quantities and with ade-
quate precautions.
Luminescence data: Broad-band excited emission spectra were recorded
on
a Fluorolog FL-3-22 spectrometer from Horiba-Jobin-Yvon Ltd;
quartz cells with optical paths of 0.2 cm were used for room temperature
spectra, while 77 K measurements were carried out on samples put into
quartz Suprasil capillaries. All spectra are corrected for the instrumental
function. Quantum yields of the helicates were determined in aerated
water in two ways. Firstly with respect to a known standard, [Ln
(dpa)3]3ꢁ
G
(Ln=Eu, Q=24%, Ln=Tb, Q=22%).[48] Absorbance of the samples
and reference was usually kept near 0.2. Total ligand concentration was
3.75”10ꢁ5 m and the following equation was used:[49]
ꢀ
ꢁ
ꢀ
ꢁ
Ssample
Sref
Compound 8: The crude mixture of products 7a + 7b was dissolved in
acetic anhydride (10 mL) and stirred at 708C for 12 h; the solvent was re-
moved and the resulting residue was dried under vacuum (2 mbar) for
12 h. The completion of the reaction was monitored by mass spectrome-
try. ESI: m/z: calcd for [M+H]+: 839.95; found: 839.39, 420.34
[M+2H]2+/2; N-oxide derivative: m/z: calcd for: 420.21; found: 855.27
[M+H]+, 428.37 [M+2H]2+/2. This product could not be purified and
was directly used in the next step by dissolution in a methanolic solution
of 7n ammonia (20 mL) and stirring for 6 h at RT. The solvent was then
removed giving an orange oil (150 mg, overall yield: 71%). 1H NMR
(CDCl3): d=8.26 (d, 1H, Har), 7.79 (m, 1H, Har), 7.71 (d, 1H, Har), 7.53–
7.22 (m, 2H, Har), 7.21 (d, 1H, Har), 4.93(d, 2H, -OCH 2-), 4.27 (s, 1H,
-CH2OH), 3.96 (d, 2H, -OCH2-), 3.57–3.30 (m, 9H, -CH2-), 3.32 ppm (s,
3H, -OCH3); ESI MS: m/z: calcd for [M+H]+: 755.38; found: 755.32,
378.31 [M+2H]2+/2.
Aref
Asample
sample
abs
ref
abs
Q
¼
ꢃ
ꢃ Q
where A is the absorbance and S are integrated, corrected emission area.
Refractive index correction was not needed, all solutions having the
same n2D0. The obtained values were 10.0ꢂ1.5% for EuIII and 0.34ꢂ
0.04% for TbIII. Secondly, the quantum, yields were checked by an abso-
lute method[50] based on
a home-modified integrating sphere from
Oriel.[51] Solutions of the helicates (1.25”10ꢁ5 m) were introduced into
3mm OD quartz capillaries (volume needed: about 40 mL) and the in-
strumental function of the spectrometer was carefully established before
measurements; the results were Q=11.2ꢂ0.5 (EuIII
)
and 0.33ꢂ0.02
(TbIII) in very good agreement with the data obtained by the classical
method. All data reported are averages of at least three independent
measurements. High-resolution emission spectra and excited state life-
times were measured on a previously described instrumental set-up.[52] It
is noteworthy that lifetime determinations with the FL 3-22 spectrometer
always yielded perfectly single exponential decays but when measured on
the high resolution instruments (Ln=Eu only), which provides a shorter
sampling time interval, the best fits were obtained with a bi-exponential
function.
H2LC3: Compound 8 (150 mg, 0.2 mmol) was dissolved in 0.1m NaOH
(50 mL) and KMnO4 (92 mg, 0.59 mmol) was added under vigorous stir-
ring. The solution was heated at 708C for 12 h, filtrated, and the aqueous
basic phase was extracted twice with CH2Cl2 (50 mL). The pH of the
aqueous phase was decreased to 2 before extraction with CH2Cl2 (3”
75 mL). The combined organic phases were dried over Na2SO4, evaporat-
ed and the resulting crude solid was purified by column chromatography
(silica gel; MeCN/25% NH4OH 100!88:12) followed by preparative
HPLC (MeCN/H2O, TFA 0.09%) to give the final H2LC3 product as a
slight orange oil (48 mg, 31%). 1H NMR (CDCl3): d=8.67 (d, 1H, Har),
8.26 (d, 1H, Har), 8.09 (dd, 1H, Har), 7.85 (d, 1H, Har), 7.45 (d, 1H, Har),
7.27 (d, 1H, Har), 4.93(d, 2H, -OCH 2-), 4.27 (s, 1H, -CH2OH), 4.35 (d,
2H, -OCH2-), 4.13(d, 2H, -OCH 2-), 3.65–3.48 (m, 9H, -OCH2-),
3.69 ppm (s, 3H, -OCH3); 13C NMR (MeOD): d = 167.43, 150.15, 149.42,
140.09, 139.21, 137.88, 128.69, 128.60, 127.31, 126.89, 126.80, 119.09,
118.99, 118.68, 113.49, 113.28, 72.8, 71.67, 71.46, 71.388, 71.35, 71.22,
58.99 ppm; ESI MS: m/z: calcd for [M+H]+: 783.34; found: 783.36,
For luminescence measurement of the helicate permeated into HeLa
cells, the latter at 90% confluency were loaded with the helicate solution
(250 mm in RPMI-1640 culture medium) for 6 h, washed with PBS (phos-
phate buffer saline) ten times, and then harvested with trypsin. The mix-
ture was centrifuged (3000 rpm, 5 min), and the cell pellet was re-sus-
pended in PBS (0.8 mL).
Cell culture: The mouse hybridoma cell line 5D10, the human T leuke-
mia cell line Jurkat (ATCC TIB152), the human breast adenocarcinoma
cell line MCF-7 (ATCC HTB-22), and the human cervical adenocarcino-
ma cell line HeLa (ATCC CCL-2) were used in this study. Cells were cul-
tivated in 75 cm2 culture flasks using RPMI-1640 supplemented with 5%
fetal calf serum (FCS), 2 mm l-glutamine, 1 mm sodium pyruvate, 1%
non-essential amino acids, 1% 4-(2-hydroxyethyl) monosodium salt
(HEPES) (all from Gibco Cell Culture, Invitrogen, Basel, Switzerland).
Cultures were maintained at 378C under 5% CO2 and 95% air atmos-
phere. The growth medium was changed every other day until the time
of use of the cells. Cell density and viability, defined as the ratio of the
number of viable cells over the total number of cells, of the cultures were
determined by trypan blue staining and a Neubauer improved hemacy-
tometer (Blau Brand, Wertheim, Germany).
392.33
[M+2H]2+/2;
elemental
analysis
calcd
(%)
for
C41H46N6O10·NaCl·H2O: C 57.31, H 5.63, N 9.78, found: C 57.24, H 5.72,
N 9.48.
Preparation of the helicate solutions: The 2:3helicates
(C123H132N18O30Ln2) were synthesized in situ by mixing three equivalents
of H2LC3 with two eqs of Ln
(ClO4)3·xH2O (Ln=La, Lu, Gd, Tb, Eu, x=
A
2.5–4.5) in water. When needed, the pH was adjusted to 7.4 with a Tris-
HCl 0.1m buffer solution. CAUTION! Perchlorate salts combined with
organic compounds are potentially explosive and should be handled in
small quantities and with adequate precautions.[44]
Spectrophotometric titrations: Protonation constants of H2LC3 were de-
termined with the help of a J&M diode array spectrometer (Tidas series)
connected to an external computer. All titrations were performed in a
thermostated (25.0ꢂ0.18C) glass-jacketed vessel at m=0.1m (KCl). In a
typical experiment a 1.01”10ꢁ5 m solution of H2LC3 (50 mL) was titrated
with a freshly prepared sodium hydroxide solution at different concentra-
tions (10, 4, 1, 0.1 and 0.01m). After each addition of base, the pH of the
WST-1 cell proliferation assay: This assay is dependent on the cellular re-
duction of WST-1 (4-(3-(4-iodophenyl)-2-(4-nitrophenyl)-2H-5-tetrazo-
lio)-1,3-benzene disulfonate (cell proliferation reagent WST-1, Roche,
Germany) by the mitochondrial dehydrogenases in viable cells to give a
dark red formazan product. The viable cell number per well is directly
proportional to the production of red formazan, which can be measured
spectrophotometrically.[53,54] Cells were seeded in a 96-well tissue culture
microplate at a concentration between 1–5”105 cells per well in 100 mL
9524
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 9515 – 9526

Luminescent Staining of Living Cells
FULL PAPER
culture medium and incubated overnight at 378C and 5% CO2. The heli-
cate was dissolved in fresh RPMI-1640 medium at 378C, at a concentra-
tion of 500 mm. The medium was removed from the cell cultures and
100 mL per well of the helicate was added (final concentrations: 500 mm,
250 mm, and 125 mm); 10 mL of WST-1 reagent was added to each well
of the complexes, and Dr. J. Raus from the Biomedical Research Insti-
tute “Dr. L. Willems-Instituut”, University of Hasselt, Belgium, for
kindly providing us with the hybridoma cell line 5D10.
and the plate was shaken for 1 min on
a microtiter plate shaker
[1] L. J. Martin, M. J. Hahnke, M. Nitz, J. Wohnert, N. R. Silvaggi, K. N.
Allen, H. Schwalbe, B. Imperiali, J. Am. Chem. Soc. 2007, 129,
7106–7113.
[2] N. Johnsson, K. Johnsson, ACS Chem. Biol. 2007, 2, 31–38.
[3] J.-C. G. Bünzli, Acc. Chem. Res. 2006, 39, 53–61.
[4] J.-C. G. Bünzli, C. Piguet, Chem. Soc. Rev. 2005, 34, 1048–1077.
[5] S. Faulkner, S. J. A. Pope, B. P. Burton-Pye, Appl. Spectrosc. Rev.
2005, 40, 1–31.
(450 rpm). The plate was further incubated at 378C and 5% CO2 and the
absorbance of the formazan product was measured at 450 nm with an
ELISA reader (Spectra MAX 340, Molecular Devices, Sunnyvale, CA,
USA). Cell viability was calculated from the absorbance values as:
ðA450ꢁA650Þexp
viabilityWST ½%Š ¼
ꢃ 100
ðA450ꢁA650Þmedium
[6] P. R. Selvin, Annu. Rev. Biophys. Biomol. Struct. 2002, 31, 275–302.
[7] K. Matsumoto, J. G. Yuan, Lanthanide Chelates as Fluorescent
Labels for Diagnostics and Biotechnology, Metal Ions in Biological
Systems, Vol. 40 (Eds.: A. Sigel, H. Sigel), Marcel Dekker, New
York, 2003, Chapter 6.
[8] I. Hemmilä, V. M. Mukkala, Crit. Rev. Clin. Lab. Sci. 2001, 38, 441–
519.
[9] W. L. Scaff, D. L. Dyer, K. Mori, J. Bacteriol. 1969, 98, 246–248.
[10] S. Phimphivong, S. S. Saavedra, Bioconjugate Chem. 1998, 9, 350–
357.
[11] H. C. Manning, S. M. Smith, M. Sexton, S. Haviland, M. F. Bai, K.
Cederquist, N. Stella, D. J. Bornhop, Bioconjugate Chem. 2006, 17,
735–740.
[12] H. C. Manning, T. Goebel, R. C. Thompson, R. R. Price, H. Lee,
D. J. Bornhop, Bioconjugate Chem. 2004, 15, 1488–1495.
[13] M. P. Houlne, T. S. Agent, G. E. Kiefer, K. Mcmilian, D. J. Bornhop,
Appl. Spectrosc. 1996, 50, 1221–1228.
[14] R. A. Poole, G. Bobba, M. J. Cann, J. C. Frias, D. Parker, R. D. Pea-
cock, Org. Biomol. Chem. 2005, 3, 1013–1024.
[15] S. Pandya, J. H. Yu, D. Parker, Dalton Trans. 2006, 2757–2766.
[16] R. Pal, D. Parker, Chem. Commun. 2007, 474–476.
[17] K. Hanaoka, K. Kikuchi, H. Kojima, Y. Urano, T. Nagano, J. Am.
Chem. Soc. 2004, 126, 12470–12476.
with (A450ꢁA650) being the absorbance difference between 450 nm and
650 nm for the cells that were in contact with the complex (“exp”) and
medium only. The results are expressed as an average over three nomi-
nally identical measurements.
Homogeneous membrane integrity assay: The measurement of the leak-
age of components, such as lactate dehydrogenase (LDH), from the cyto-
plasm into the surrounding culture medium is a widely accepted method
to estimate the number of non-viable cells. LDH leakage was measured
for different cell lines by using the CytoTox 96 non-radioactive cytotoxic-
ity assay kit (Promega Corporation, Madison, WI, USA) in which the
LDH released from the cells catalyses the conversion of resazurin into
resorufin.[55,56] Cells were seeded in a 96-well tissue culture microplate as
for the WST-1 test. After addition of the helicate, the plate was further
incubated at 378C and 5% CO2 for 30 min, removed from the incubator
and equilibrated to 228C; 50 mL of substrate mix solution was then added
before final incubation at 228C for 10 min. Fluorescence values (lex
=
530 nm, lan =590 nm) were measured with microplate fluorimeter
a
SAFIR II, from Tecan Schweiz AG. The results are expressed as an aver-
age over 3nominally identical measurements. Maximum LDH release
after cell lysis by 9 vol% Triton X-100 and spontaneously released LDH
in cells not in contact with the helicate were also measured. The average
fluorescence values of the culture medium background were subtracted
from all fluorescence values of experimental wells. Cytotoxicity caused
by the complex is calculated as follows:
[18] B. Song, G. Wang, J. Yuan, Chem. Commun. 2005, 3553–3555.
[19] J.-C. G. Bünzli, C. Piguet, Chem. Rev. 2002, 102, 1897–1928.
[20] C. Piguet, G. Bernardinelli, G. Hopfgartner, Chem. Rev. 1997, 97,
2005–2062.
[21] C. Piguet, M. Borkovec, J. Hamacek, K. Zeckert, Coord. Chem. Rev.
2005, 249, 705–729.
IexpꢁImedium
cytotoxicityLDH ½%Š ¼
ꢃ 100
Iexp,maxꢁImedium
where Iexp, Imedium, and Iexp,max are fluorescence values for cells in contact
with the helicate, for the spontaneous release of LDH by cells not in con-
tact with the helicate, and for lysed cells, respectively. Results are report-
ed in Table 3.
[22] S. Torelli, D. Imbert, M. Cantuel, G. Bernardinelli, S. Delahaye, A.
Hauser, J.-C. G. Bünzli, C. Piguet, Chem. Eur. J. 2005, 11, 3228–
3242.
[23] N. AndrØ, T. B. Jensen, R. Scopelliti, D. Imbert, M. Elhabiri, G.
Hopfgartner, C. Piguet, J.-C. G. Bünzli, Inorg. Chem. 2004, 43, 515–
529.
[24] T. B. Jensen, R. Scopelliti, J.-C. G. Bünzli, Inorg. Chem. 2006, 45,
7806–7814.
[25] T. B. Jensen, R. Scopelliti, J.-C. G. Bünzli, Chem. Eur. J., DOI:
10.1002/chem.200700468.
[26] M. Elhabiri, R. Scopelliti, J.-C. G. Bünzli, C. Piguet, J. Am. Chem.
Soc. 1999, 121, 10747–10762.
[27] C. D. B. Vandevyver, A.-S. Chauvin, S. Comby, J.-C. G. Bünzli,
Chem. Commun. 2007, 1716–1718.
[28] M. Elhabiri, J. Hamacek, J.-C. G. Bünzli, A.-M. Albrecht-Gary, Eur.
J. Inorg. Chem. 2004, 51–62.
[29] J.-C. G. Bünzli, in Lanthanide Probes in Life, Chemical and Earth
Sciences. Theory and Practice (Eds.: J.-C. G. Bünzli, G. R. Choppin),
Elsevier, Amsterdam, 1989, Chapter 7, 219–293.
[30] K. Binnemans, C. Gçrller-Walrand, Chem. Phys. Lett. 1995, 245, 75–
78.
[31] C. Gçrller-Walrand, K. Binnemans, Rationalization of crystal field
parameterization, Handbook on the Physics and Chemistry of Rare
Earths, Vol. 23 (Eds.: K. A. Gschneidner, Jr., L. Eyring), Elsevier,
Amsterdam, 1996, Chapter 155.
Live cell imaging: Cells for fluorescence microscopy were seeded on
glass bottom cell culture dishes and loaded with the complex dissolved in
freshly prepared cell culture medium. They were incubated at 37 or 48C
for the indicated time and washed at least 5–6 times with PBS before
measurement. For the counter-staining experiment, cells were subse-
quently incubated at RT for 5 min with 100 mgmLꢁ1 of acridine orange
(from Invitrogen) and washed 10 times with PBS. Images were recorded
with a Zeiss fluorescence microscope Axiovert S100 fitted with a Plan-
Neofluar objective (20”), at lexc =330 nm (BP=80 nm) with an emission
filter LP 585 nm for EuIII and at lexc =470 nm (BP 40 nm) with an emis-
sion filter BP 515–565 nm for acridine orange. Exposure times varied
from 15 ms to 60 s.
Acknowledgements
This project is supported through a grant from the Swiss National Science
Foundation. Selected funding by ESF-COST Action D38 is acknowl-
edged. We thank FrØdØric Gumy for his help in luminescence measure-
ments and quantum yield determination, K. Lorentz for contributing to
the synthesis of LC3, Dr. A. Razaname for recording the ES-MS spectra
[32] J.-C. G. Bünzli, B. Klein, D. Wessner, N. W. Alcock, Inorg. Chim.
Acta 1982, 59, 269–274.
Chem. Eur. J. 2007, 13, 9515 – 9526
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9525

J.-C. G. Bünzli, A.-S. Chauvin, C. D. B. Vandevyver et al.
[33] S. T. Frey, W. D. Horrocks, Jr., Inorg. Chim. Acta 1995, 229, 383–
390.
[47] H. Gampp, M. Maeder, C. J. Meyer, A. D. Zuberbühler, Talanta
1986, 33, 943–951.
[34] C. Piguet, J.-C. G. Bünzli, G. Bernardinelli, G. Hopfgartner, S.
Petoud, O. Schaad, J. Am. Chem. Soc. 1996, 118, 6681–6697.
[35] R. M. Supkowski, W. D. Horrocks, Jr., Inorg. Chim. Acta 2002, 340,
44–48.
[36] M. H. V. Werts, R. T. F. Jukes, J. W. Verhoeven, Phys. Chem. Chem.
Phys. 2002, 4, 1542–1548.
[48] A.-S. Chauvin, F. Gumy, D. Imbert, J.-C. G. Bünzli, Spectrosc. Lett.
2004, 37, 517–532; erratum: A.-S. Chauvin, F. Gumy, D. Imbert, J.-
C. G. Bünzli, Spectrosc. Lett. 2006, 40, 193.
[49] B. Fanget, O. Devos, M. Draye, Anal. Chem. 2003, 75, 2790–2795.
[50] J. C. de Mello, H. F. Wittmann, R. H. Friend, Adv. Mater. 1997, 9,
230–232.
[37] F. J. Steemers, W. Verboom, D. N. Reinhoudt, E. B. Vandertol, J. W.
Verhoeven, J. Am. Chem. Soc. 1995, 117, 9408–9414.
[38] Z. Darzynkiewicz, Methods Cell Biol. 1990, 33, 285–298.
[39] A. B. Pangborn, M. A. Giardello, R. H. Grubbs, R. K. Rosen, F. J.
Timmers, Organometallics 1996, 15, 1518–1520.
[40] J.-C. G. Bünzli, C. Mabillard, Inorg. Chem. 1986, 25, 2750–2754.
[41] G. Schwarzenbach, Complexometric Titrations, Chapman and Hall,
London, 1957.
[51] F. Gumy, J.-C. G. Bünzli, unpublished work, 2006.
[52] R. Rodriguez-Cortinas, F. Avecilla, C. Platas-Iglesias, D. Imbert, J.-
C. G. Bünzli, A. de Blas, T. Rodriguez-Blas, Inorg. Chem. 2002, 41,
5336–5349.
[53] M. Ishiyama, M. Shiga, K. Sasamoto, M. Mizoguchi, P. G. He, Chem.
Pharm. Bull. 1993, 41, 1118–1122.
[54] M. Ishiyama, K. Sasamoto, M. Shiga, Y. Ohkura, K. Ueno, K. Nish-
iyama, I. Taniguchi, Analyst 1995, 120, 113–116.
[55] C. Korzeniewski, D. M. Callewaert, J. Immunol. Methods 1983, 64,
313–320.
[42] C. Piguet, G. Bernardinelli, B. Bocquet, A. Quattropani, A. F. Wil-
liams, J. Am. Chem. Soc. 1992, 114, 7440–7451.
[43] G. Black, E. Depp, B. B. Corson, J. Org. Chem. 1949, 14, 14–21.
[44] K. N. Raymond, Chem. Eng. News 1983, 61, 4–4.
[45] E. R. Malinowski, D. G. Howery, Factor Analysis in Chemistry,
Wiley, New York, 1991.
[56] T. Decker, M. L. Lohmannmatthes, J. Immunol. Methods 1988, 115,
61–69.
[46] H. Gampp, M. Maeder, C. J. Meyer, A. D. Zuberbühler, Talanta
1985, 32, 257–264.
Received: June 11, 2007
Published online: September 18, 2007
9526
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 9515 – 9526