Cell uptake of a biosensor detected by hyperpolarized 129Xe NMR: The transferrin case

Article abstract of DOI:10.1016/j.bmc.2011.05.002

For detection of biological events in vitro, sensors using hyperpolarized ¹²⁹Xe NMR can become a powerful tool, provided the approach can bridge the gap in sensitivity. Here we propose constructs based on the non-selective grafting of cryptopha

Full text of DOI:10.1016/j.bmc.2011.05.002

Bioorganic & Medicinal Chemistry 19 (2011) 4135–4143
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Cell uptake of a biosensor detected by hyperpolarized 129Xe NMR:
The transferrin case
Céline Boutin a, Antoine Stopin b, Fatimazohra Lenda b, Thierry Brotin b, Jean-Pierre Dutasta b,
Nadège Jamin c, Alain Sanson c, Yves Boulard d, François Leteurtre d, Gaspard Huber a,
Aurore Bogaert-Buchmann a, Nawal Tassali a, Hervé Desvaux a, Marie Carrière a, Patrick Berthault a,
⇑
a CEA, IRAMIS, SIS2M, Laboratoire Structure et Dynamique par Résonance Magnétique, UMR CEA/CNRS 3299, 91191 Gif sur Yvette, France
b Laboratoire de Chimie de l’ENS Lyon, UMR 5182 CNRS-ENS Lyon, 46 Allée d’Italie, 69364 Lyon Cedex 07, France
c CEA, iBiTec-S, SB2SM, URA CEA/CNRS 2096, 91191 Gif sur Yvette, France
d CEA, iBiTec-S, SBIGeM, LBI, 91191 Gif sur Yvette, France
a r t i c l e i n f o
a b s t r a c t
For detection of biological events in vitro, sensors using hyperpolarized 129Xe NMR can become a pow-
erful tool, provided the approach can bridge the gap in sensitivity. Here we propose constructs based
on the non-selective grafting of cryptophane precursors on holo-transferrin. This biological system was
chosen because there are many receptors on the cell surface, and endocytosis further increases this den-
sity. The study of these biosensors with K562 cell suspensions via fluorescence microscopy and 129Xe
NMR indicates a strong interaction, as well as interesting features such as the capacity of xenon to enter
the cryptophane even when the biosensor is endocytosed, while keeping a high level of polarization.
Despite a lack of specificity for transferrin receptors, undoubtedly due to the hydrophobic character of
the cryptophane moiety that attracts the biosensor into the cell membrane, these biosensors allow the
first in-cell probing of biological events using hyperpolarized xenon.
Article history:
Received 24 February 2011
Revised 26 April 2011
Accepted 1 May 2011
Available online 6 May 2011
Keywords:
Biosensors
Cryptophane
Xenon
Cells
Ó 2011 Elsevier Ltd. All rights reserved.
NMR
Hyperpolarization
1. Introduction
Recently Dmochowski’s group proposed an original biosensor con-
struction based on the functionalization of cryptophane with a
There is a worldwide effort to propose hyperpolarized species
for molecular imaging applications based on magnetic resonance.
In this growing field, hyperpolarized xenon occupies a special
and well-recognized position as a powerful magnetic resonance
sensor of molecular or biological events. This is due to the huge sig-
nal obtained by the prior optical pumping1 or dynamic nuclear
polarization2 step, to its exogenous and gaseous nature, and also
to the wide chemical shift range spanned by this atom, which
has a big deformable electron cloud.3 Using these properties, an ap-
proach has been envisioned where the hyperpolarized noble gas is
targeted to biological receptors via functionalized host systems.4,5
When encapsulated in the core of transporters made by cage-mol-
ecules such as cryptophanes6 or nanoparticles,7 xenon has a spe-
cific resonance frequency which enables its discrimination from
free xenon in spectroscopic imaging.8
cell-penetrating peptide.14 Fluorescence tests on living cells were
performed but no hyperpolarized 129Xe NMR experiment was pre-
sented. The experimental implementation appears demanding.
Studying live cell suspensions by NMR requires a compromise
regarding cell density, which must be high enough to reach a con-
centration compatible with the monitoring of cellular biological
events, but not too high otherwise cell lifetime will be too short.
Also a contrast mechanism, where the free and bound biosensor
signals are discriminated, is needed to ensure that the biological
event is really detected. For approaches using hyperpolarization,
this means dealing with concentrations in the pico- to nanomolar
range and managing more efficiently the nuclear spin relaxation
induced by the presence of paramagnetic agents (ions, metallo-
proteins, dissolved oxygen, etc.).
Here the interaction of a biosensor of transferrin receptors (TfR)
is detected via hyperpolarized 129Xe NMR of cell suspensions, and
the NMR results are correlated with fluorescence microscopy. This
system has been chosen for the following reasons: (i) it is a well-
described system: transferrin (Tf), a naturally existing protein,
has already received considerable attention in the area of drug tar-
geting since it is biodegradable, non-toxic, and nonimmunogen-
ic;15 (ii) the biological targets are cell surface receptors and so
Whereas the proof-of-concept of this approach has been suc-
cessfully achieved in isotropic solution by several groups,9–13 it
has never been tested in vitro (or in vivo) by performing hyperpo-
larized 129Xe NMR experiments directly with cell suspensions.
⇑
Corresponding author. Tel.: +33 1 69 08 42 45; fax: +33 1 69 08 98 06.
E-mail address: patrick.berthault@cea.fr (P. Berthault).
0968-0896/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2011.05.002

4136
C. Boutin et al. / Bioorg. Med. Chem. 19 (2011) 4135–4143
are easily reachable by dissolved xenon; (iii) these receptors can be
abundant (up to 1.6 105 on the cell surface) for some cell lines;16,17
(iv) endocytosis due to Tf–TfR complex formation is likely to fur-
ther increase the local density of xenon receptors.
2. Results and discussion
2.1. Design and synthesis of the TfR sensors
It is known that the molecular systems for which xenon exhibits
the highest affinity are cryptophanes, aromatic cage-molecules
made of two cyclotriveratrylene bowls joined by aliphatic linkers.
Consequently they constitute the core of our 129Xe NMR-based bio-
sensors. A tether comprising a polyethylene spacer ended by an
activated ester was first grafted on the cryptophane to produce
precursor 1 (Fig. 1). The choice of four O–CH2–CH2 groups to con-
stitute the spacer gave the cryptophane moiety flexibility, which
was expected to be sufficient to avoid broad lines for encapsulated
xenon caused by too fast transverse relaxation.18 This spacer was
also expected to be long enough to avoid multiplicity of the signals
of encapsulated xenon when using the cryptophane racemic mix-
ture.19 The activated group of the cryptophane moiety was chosen
to react with accessible primary amines, that is,
e-amine on lysine
residues and -amine groups at the N-termini of proteins. Cross-
a
linking reagents such as N-hydroxysuccinimide esters are cur-
rently used to investigate protein–protein interactions,20 but are
characterized by a fast hydrolysis.21 In order to minimize such
deactivation phenomena by hydrolysis, a pentafluorophenol ester
end22 was used for the grafting step on the protein. Such a con-
struction presents many advantages: (i) the same cryptophane-
PEG-activated ester precursor can be used for the design of various
biosensors, since it can be non-specifically grafted on any protein
with lysine residues on its surface; (ii) if more than one lysine is
present on the surface of the protein, several cryptophane cores
can be grafted, therefore increasing the number of xenon binding
sites and accordingly enhancing the intrinsic NMR sensitivity of
the biosensor; (iii) provided that lysine residues are distant from
the protein active site, due to the small size of cryptophane com-
pared with that of the protein, it is expected that the affinity of
the biosensor for the biological receptors is preserved.
Transferrin is a globular protein of molecular weight ꢀ80,000
Dalton. It possesses 54 Lys residues, 23 of which are located on its
external surface, and its holo form contains two iron atoms
(Fig. 2a). The cryptophane precursorwas grafted on the proteinusing
the procedure detailed in Section 4. The affinity of holo-transferrin
for its receptor being 800 times greater than that of apo-transferrin,
we decided to work with the former, despite the relaxation induced
on xenon by the paramagnetism of Fe3+ ions. An alternative would
have been, had the paramagnetic-induced relaxation reduced the
NMR signal of hyperpolarized xenon too much, to replace iron atoms
by gallium or indium atoms, which have similar affinities for TfR.23
In order to investigate the interaction of transferrin with its
receptor in the presence of cells, two types of constructs were used.
Figure 2. Example of a structure of the biosensors used and laser-polarized 129Xe
NMR spectra of the noble gas dissolved in a 1.3 M solution of B1 in PBS (277 K,
l
11.7 T). On the construct displayed in (a), two cryptophane moieties are grafted on
the protein. The lysine side chains are colored. Whereas the full spectrum in (b) was
obtained in one scan with a hard pulse of small flip angle, the sub-spectrum in (c)
was obtained by accumulation of 256 selective excitations with a repetition time of
122 ms. The 55–95 ppm frequency range, corresponding to the Xe@cryptophane
region, is displayed.
In the first, a 129Xe NMR-based biosensor made with several cryp-
tophanes per protein was built. As checked by 1H NMR, up to 10
cryptophanes could be grafted on transferrin before protein
unfolding. A safe solution of five cryptophanes per protein was fi-
nally chosen and identified as biosensor B1. The second construct
(biosensor B2) had a double modality: fluorescence and NMR. A
statistical ratio of 2:2:1 was used for the cryptophane, fluorescent
probe (rhodamine green) and protein, respectively. The fluoro-
phore was grafted on the protein using the same strategy as for
the cryptophane (see Section 4) and the folding of this protein con-
struct B2 was checked by 1H NMR. Finally, two other control con-
structs B3 and B4 were prepared. They corresponded to the
previous biosensors except that bovine serum albumin (BSA) was
used instead of transferrin. BSA has been widely used as a marker
of fluid-phase pinocytosis.24 Therefore, it is used to explore non-
specific interaction with cell membranes. For the first biosensor,
Figure 1. Chemical structure of the activated-ester cryptophane 1.

C. Boutin et al. / Bioorg. Med. Chem. 19 (2011) 4135–4143
4137
the cryptophane precursor was grafted on BSA with a ratio of 5
cryptophanes per protein (biosensor B3). The second probe (the
fluorescent biosensor B4) was built with a ratio of 2 cryptophanes
and 2 rhodamine green moieties per BSA protein.
140 million K562 cells were pretreated with pronase, a mixture
of proteolytic enzymes that disable the function of TfR on the cell
surface.15,25 Following exactly the same procedure as in (a) we got
samples sb and cb, respectively.
Figure 2 displays the 129Xe NMR spectra obtained at 11.7 T and
The present protocol allowed pairwise comparison of the cell
samples ca and cb through fluorescence microscopy when the bio-
sensor B2 was used, and through hyperpolarized 129Xe NMR with
biosensor B1, as well as comparison between the supernatant sam-
ples sa and sb.26 As exchange of the transferrin biosensor between
the intra- and extracellular compartments is driven by the laws
of thermodynamics, this procedure was a way to ‘freeze’ the situ-
ation encountered during the incubation of the cells. Considering
that the low temperature used for the washing steps and the con-
servation of the cells block the dynamics of the plasma membrane,
this amounts to considering that the fluorescence images and 129Xe
NMR spectra reflect the distribution of the biosensor at the end of
the incubation step.
277 K with B1 at 1.3 lM in PBS. Clearly the Xe@B1 line is unique
and not broad. This confirms that the spacer between the crypto-
phane part and the protein is long and flexible enough. From this
experiment, the detection threshold using the direct observation
method5 has been estimated as 80 nM, or about 50 pmol using only
1 mL of gaseous hyperpolarized xenon. Obviously in the presence
of the cells, the inherent loss of field homogeneity is likely to in-
crease these limit values. For samples of apo- and holo-transferrin
at 90
been measured, respectively. Therefore, at the concentration of
biosensor used in the presence of cells (5 M) relaxation due to
l
M, 129Xe relaxation times of 450 5 s and 147 10 s have
l
paramagnetism does not represent a limitation.
Figure 4 displays the fluorescence microscopy results. After
incubation of the cells with B2, the internalization of the biosensor
is clearly visible. The observation of clustered fluorescence inside
the cells could be due to the presence of concentrated pools of
B2 within intracellular compartments. B2 can also be found within
the plasma membrane as indicated by the presence of green spots
at the cell surface. After incubation of the K562 cells with pronase,
no fluorescence was detected in the cell.
Figure 5 shows the 55–95 ppm chemical shift region of the la-
ser-polarized 129Xe NMR spectra of both the cells incubated with
B1 and the supernatant samples. This chemical shift range exclu-
sively corresponds to the signal of the noble gas encapsulated in
the cryptophane.5 In the 129Xe NMR spectrum of sample ca, two
peaks were observed in this region; in addition to the peak at
68–69 ppm, which was also observed in the isotropic solution
(Fig. 2) or in the supernatant sa and corresponds to the biosensor
in an aqueous environment, a second strong broad resonance at
79 ppm was observed. These two peaks had similar intensities. A
quick calculation taking into account the number of cells present
in solution and the estimated number of transferrin receptors per
K562 cell revealed that the second peak cannot be assigned to xe-
non in the biosensor linked to the Tf receptor, whereas the first
peak corresponds to the biosensor free in solution. For the current
biosensor concentration, according to what is known for this bio-
logical system, the second peak should be much smaller. On the
other hand, Meldrum et al. have shown that a chemical shift differ-
ence of up to 10 ppm can occur between xenon in the cryptophane
2.2. Binding of the biosensors to K562 cells
The human erythroleukemic K562 cell line was chosen because
it has a huge number of transferrin receptors on the plasma mem-
brane: 1.6 105 receptors per cell (8 105 including the internalized
receptors).16,17 As our polarization level and our experimental pro-
tocol already enabled detection of 1012 xenon spins,5 a rough cal-
culation showed that 107 cells in the detection volume should be
enough.
It appeared difficult to envisage a direct MR contrast in a cell
suspension, as our procedure includes shaking of the NMR tube
to introduce fresh hyperpolarized xenon into solution. Further-
more, even if a tiny modification of the cryptophane environment
is likely to give rise in the 129Xe NMR spectrum to a chemical shift
variation of the Xe@biosensor signal when the biosensor is inter-
acting with its biological target,10 here for cell suspensions the line
broadening due to susceptibility variation rendered such a detect-
able frequency separation unlikely. These difficulties led us to
imagine the following ‘batch’ procedure, which is summarized in
Figure 3. It is described here for biosensor B1, but strictly the same
procedure was used for the other biosensors. In (a), 140 million
K562 cells were incubated for one hour at 310 K with 3 mL of a
solution of B1 in PBS (final concentration 5 lM). After centrifuga-
tion at 277 K, the first supernatant gave rise to sample sa (superna-
tant a). After two further washings at 277 K, the cells were
separated, giving rise to sample ca (cells a). In (b) in parallel,
Figure 3. Synopsis of the batch experiments designed to study the interaction of
the biosensor B1 with the K562 cells.
Figure 4. Fluorescence (up) and bright field (down) images of K562 cells incubated
with the B2 biosensor, without (left) and with (right) previous pronase treatment.

4138
C. Boutin et al. / Bioorg. Med. Chem. 19 (2011) 4135–4143
in the aqueous phase and xenon in the cryptophane in a lipid
phase.27 It thus seems more likely that the peak at 79 ppm can
be assigned to the biosensor in the cell membrane. Whereas the
left peak represents xenon encapsulated in his host located in
the cell membrane, the peak at 69 ppm should contain the contri-
bution of xenon in the intracellular aqueous compartment and in
the free biosensor. Two facts substantiate this interpretation:
(i) as endocytosis passes through the formation of endosomes, it
could spread the biosensor—and therefore the cryptophane moi-
ety—in either an aqueous or a membrane environment, even in
the case of a selective interaction between the biosensor and the
transferrin receptors; (ii) in a recent paper, we have shown that xe-
non crosses the cell membrane in some tens of milliseconds and
keeps the major part of its hyperpolarization.28
As observed in Figure 5, only a very small and broad signal
around 80 ppm was detected in the 55–95 ppm region of the
129Xe NMR spectrum of the cb sample, that is, the cells pre-incu-
bated with pronase, although the xenon polarization level was
close to that of the other cell sample ca (Table 1). This is in perfect
agreement with the observations of fluorescence microscopy
(Fig. 4). The resemblance between the xenon spectra of the super-
natant samples (sa and sb) ensures that the supernatant contains
only the biosensor in aqueous environment.
caged xenon signal in the sample cb results from biological changes
in the sample and that the differences in peak integrals in this
spectral region is indicative of the biological events.
In this batch experiment, the net difference between the 129Xe
NMR spectra of the noble gas in the biosensor interacting with
the cells after treatment or not by pronase (samples cb vs ca) indi-
cates that the transferrin receptors are required for a strong inter-
action of the biosensor with the cells.
2.3. Further assessment of the specificity of the biosensor
Two other non-specific biosensors B3 and B4, based on bovine
serum albumin, were studied with K562 cells. The same procedure
as before was followed, giving cc and sc samples. Figure 6 displays
the fluorescence microscopy image obtained with biosensor B4
and the laser-polarized 129Xe NMR spectrum of the cc sample. The
fluorescence image is far less bright than that of the transferrin bio-
sensor B2. In particular, unlike the case of B2, no bright internal spot
can be observed inside the cells, a result in agreement with the ab-
sence of specific BSA receptor on the surface of the K562 cells en-
abling active internalization of the biosensor. In contrast,
surprisingly two peaks appear again on the 129Xe NMR spectrum
of B3 with K562 cells, tending to indicate an interaction between
BSA and the K562 cells. Again, the sum of the normalized integrals
(cc + sc) is comparable to (ca + sa) or (cb + sb), thus validating our
procedure.
To explain the propensity of B3 as well as of B1 to interact non-
specifically with the cells we could assume that the hydrophobicity
of the cryptophane moieties anchors both biosensors in the cell
membrane. To address this point, another experiment was per-
formed. Two new transferrin-based constructs were built. In the
first, the fluorescent probe was grafted (again via the lysine side
chains) on apo-transferrin (two rhodamine green moieties per pro-
tein). In the second, the fluorescent probe and the cryptophane pre-
cursor were grafted on apo-transferrin, at a ratio 2:2 for one protein.
These constructs were introduced in solutions containing giant uni-
lamellar vesicles. Comparison of the bright field and fluorescence
microscopy images obtained in both cases is displayed in Figure S1
of the Supplementary data. A strong difference was observed be-
tween the two images. Whereas the fluorescence was rather diffuse
in the case of the rhodamine green-labeled protein (no interaction is
expected with the vesicles), the location of the second construct in
the vesicle lipid bilayer was clearly visible. This interaction must
be due to the presence of the cryptophane moieties on the second
construct, proving that the hydrophobic character of the cage-mol-
ecule is likely to induce non-specific interaction with the cells via
their membrane.
A more quantitative examination of the relative area of the xe-
non peaks reveals many interesting features. The sum of the nor-
malized peak areas for caged xenon in a series (ca + sa or cb + sb)
is nearly constant. This means that our biochemical procedure
(incubation and washing steps) is robust, as the same amount of
biosensor is recovered. This validates the fact that the absence of
Table 1
Peak areas extracted from the 129Xe NMR spectra of the different cell or supernatant
samples
a
b
c
Sample
Xefree
Xe@Blipid
Xe@Bwater
Xe@Bwater/Xe@Blipid
ca
sa
cb
sb
cc
sc
4.68
2.70
1
2.48
2.94
4.13
169
0
9
0
157
0
100
408
0
615
44
0.59
—
0
—
0.28
—
497
a
The values given in this column are indicative of the xenon magnetization and
can be compared.
b
Corresponds to the signal at 79 ppm.
Corresponds to the signal at 69 ppm. In each row the values of the second and
c
third columns are normalized with respect to the value of the first column (and the
value ‘100’ was chosen for the peak of Xe@Bwater for ca).
Even if these non-specific interactions prevent straightforward
analysis of the NMR results, various observations indicate that a
real biosensor has been prepared. The ratio between the caged xe-
non peaks is more in favor of the cryptophane in the aqueous part
for B1 than for B3 (0.59 vs 0.28, see Table 1). Logically the normal-
ized areas of xenon in the biosensors in a lipid environment are
similar for B1 and B3, as the same number of cells of the same line
is expected to offer the same number of membrane sites. The pres-
ence of a 129Xe NMR signal corresponding to B3 in the cells may be
explained by the non-specific internalization of the probe by pino-
cytosis.24 This hypothesis is in agreement with the diffuse inner
fluorescent signal observed in Figure 6.
Pairwise comparison of the sums of the peak areas in the 129Xe
NMR spectra for the biosensors in the presence of cells (ca vs cc) re-
veals that 33% more signal is present for B1 than for B3. This con-
firms that the specific interaction of B1 with its receptors induces
endocytosis, preventing the release of the biosensor at the washing
step. This observation proves that the biosensor is recognized by
the K562 cells, even if it is also subjected to nonspecific interaction
Figure 5. 129Xe NMR spectra obtained by selective excitation in the xenon@cryp-
tophane spectral region (293 K; 11.7 T). Upper left: with K562 cells after incubation
with the biosensor B1 for 1 h at 310 K; upper right: spectrum of the corresponding
supernatant; lower left: spectrum with K562 treated with pronase and then
incubated with the biosensor B1 for 1 h at 310 K; lower right: spectrum of the
corresponding supernatant.

C. Boutin et al. / Bioorg. Med. Chem. 19 (2011) 4135–4143
4139
one side and the oligonucleotide on the other side led to the forma-
tion of self-organized systems of the micelle or vesicle type. A solu-
tion will be to place hydrophilic groups closer to the aromatic rings
of the cryptophanes, while ensuring maintenance of the xenon in–
out exchange.29
This in vitro study demonstrates the feasibility of detecting a bio-
logical mechanism using hyperpolarized 129Xe NMR-based biosen-
sors in the presence of live cells and paves the way for the
conception of an improved biosensor addressing water solubility,
bio-distribution and toxicity issues that will be determinant for
the success of this approach in vivo.
4. Experimental
4.1. General
All the solvents were distilled before use (DMF from calcium
hydride under reduced pressure, acetone and CH2Cl2 from calcium
chloride, THF from benzophenone ketyl). Column chromatographic
separations were carried out over Merck silica gel 60 (0.040–
0.063 mm). Analytical thin-layer chromatography (TLC) was
performed on Merck silica gel TLC plates F254. Mass spectra were
recorded in LSIMS mode on a ThermoFinnigan MAT95XL spectrom-
eter. Exact mass measurements were obtained by ‘peak matching’.
1H and 13C NMR spectra were recorded on Bruker Avance 500 spec-
trometers. Chemical shift d (1H, 13C) are given relative to Me4Si.
Figure 6. Bright field (left) and fluorescence (right) images of K562 cells incubated
with the B4 biosensor, and 129Xe NMR spectrum of K562 cells incubated with the B3
biosensor, obtained by selective excitation in the xenon@cryptophane spectral
region (293 K; 11.7 T).
with the cells due to the hydrophobic character of the cryptophane
moieties.
3. Conclusion
In this article, a new strategy for the building of efficient 129Xe
NMR biosensors is developed. It is based on the nonspecific graft-
ing of cryptophane precursors on the primary amines of a protein
interacting with a specific receptor. This gives versatility to the ap-
proach and enables further gain in sensitivity by the multiplicity of
xenon hosts on the protein.
4.2. Synthesis of the cryptophane moieties
Derivative 1 was prepared using cryptophane-A (2) and a PEG
chain (3) as starting materials.
4.2.1. Synthesis of cryptophanol-A 4 (Scheme 1)
This method has been applied to the transferrin system in vitro.
Study through hyperpolarized 129Xe NMR of the biosensor interact-
ing with K562 cells reveals several features. Firstly, even for an
unmodified biological system containing a high level of paramag-
netism, NMR of hyperpolarized species in stable and reproducible
experiments is possible. At the concentrations of protein used
Cryptophanol 4 (whose synthesis has been independently re-
ported by Darzac et al.30,31 and Spence et al.9) was prepared in a
single step from cryptophane-A 2. Removal of a single methyl
group of 2 was easily achieved by using iodotrimethylsilane in
CH2Cl2 at room temperature. This synthetic route yielded cryp-
tophanol in 37% yield as well as unreacted cryptophane-A, which
could be reused for subsequent reactions:
(ꢀ5
lM), the observed xenon relaxation time remains long, which
enables full use to be made of the reservoir of polarization consti-
tuted by dissolved xenon. Moreover, given the large reservoir of
gaseous hyperpolarized xenon on top of the solution, if required
we can use additional spectra to increase the signal-to-noise ratio
simply by shaking the NMR tube between two acquisition series.
Secondly, fluorescence microscopy experiments prove that the
biosensor is recognized by the Tf receptors of the K562 cells and
endocytosed despite the presence of the cryptophane and fluoro-
phore moieties at the surface of the protein. The large differences
between hyperpolarized 129Xe NMR spectra of the biosensor alone,
of the biosensor in the presence of cells, intact and pre-treated
with pronase, confirm this result. For the first time, a specific
129Xe spectral signature is observed and correlated to a biological
process. Thanks to an original methodology leading to a quantita-
tive analysis, this biological interaction mediated by the transferrin
receptors is validated.
Pure iodotrimethylsilane (65 lL, 0.455 mmol) was added in one
portion using a syringe to a stirred solution of cryptophane-A 2
(400 mg, 0.45 mmol) in CH2Cl2 (8 mL). The solution was stirred
in the dark for 16 h under an argon atmosphere at room tempera-
ture. CH2Cl2 (10 mL) was added and the solution was acidified with
1 M HCl (6 mL). The organic layer was collected, washed with
water and then dried over sodium sulfate. After evaporating the
solvent under reduced pressure the residue was purified on a col-
umn chromatography (CH2Cl2/Et2O: 90/10) to give the cryptopha-
nol 4 as a white solid (144 mg, 37%). Spectroscopic data are
identical to those previously reported for this compound.30
4.2.2. Synthesis of the linker (Scheme 1)
The PEG linker used for the attachment of the biosensor mole-
cule to the protein was prepared in three steps from the mono-pro-
tected PEG 3, whose synthesis has been reported previously.32–34
An activated acid function was then introduced at one extremity
in order to allow coupling with the protein and a tosyl group at
the other extremity to allow coupling with cryptophanol 4.
An additional experiment using a BSA biosensor has however
pointed out the presence of nonspecific interaction contributing
to the signal. It seems reasonable thinking that it results from
the strong hydrophobic character of the cryptophane. Without
leading to protein unfolding, it tends to attract the biosensor to
the cell membrane. This shows that the current approaches aimed
at solubilizing cryptophanes in water by introducing a hydrophilic
ligand far from the cavity can have some drawbacks, as already ob-
served with a cryptophane biosensor designed to recognize a com-
plementary DNA strand in solution.10 The amphiphilicity created
by the presence of the cage-molecule and the aliphatic spacer on
4.2.2.1. 3,6,9,12-Tetraoxatetradecanoic acid,14-[[tetrahydro-2H-
pyran-2-yl]oxy]phenylmethyl ester (5).
hydro-2H-pyran-2-yl)oxy]ethoxy]ethoxy]-ethoxy]ethanol
2-[2-[2-[2-[(tetra-
(5 g,
18 mmol, 1 equiv) in THF (60 mL) was added dropwise to a stirred
solution of NaH 60% (1.8 g, 45 mmol, 2.5 equiv) in THF (50 mL) un-
der an argon atmosphere. After complete addition the mixture was

4140
C. Boutin et al. / Bioorg. Med. Chem. 19 (2011) 4135–4143
pyridine (6 mL). The solution was stirred at this temperature for
an additional 2 h. Water (10 mL) and diethyl ether (40 mL) were
then added to the solution. The organic layer was collected and
the aqueous layer was extracted twice with diethyl ether
(2 Â 20 mL). The combined organic layers were washed with brine
(20 mL) and water (10 mL) and dried over sodium sulfate. Evapora-
tion of the solvent leaves a residue, which was purified by column
chromatography (AcOEt). Different fractions were collected and
the evaporation of the solvent gave 7 (2.32 g, 80%) as a colorless
oil. Spectroscopic data are identical to those previously reported
in the literature.36 Additional information 1H NMR (500 MHz,
CDCl3, 298 K) d 7.77 (d, 3J(H,H) = 8.3 Hz, 2H; Ar), 7.33–7.29 (m,
7H; Ar), 5.16 (s, 2H; OCH2Bn), 4.17 (s, 2H; OCH2COO), 4.13 (t,
3J(H,H) = 4.9 Hz, 2H; OCH2), 3.71–3.54 (m, 14H, OCH2), 2.42 (s,
3H; CH3).
4.2.3. Synthesis of cryptophane 1 (Scheme 2)
The coupling reaction between cryptophanol 4 and PEG 7 was
carried out in DMF in the presence of cesium carbonate. The reac-
tion gave rise to the monofunctionalized cryptophane 8 in 69%
yield. The deprotection of the carboxylic acid was then achieved
by hydrogenation under atmospheric pressure in the presence of
a Pd/C catalyst in CH2Cl2/EtOH solution to give 9 in 89% yield. Fi-
nally, the carboxylic acid group was activated by introducing a
pentafluorophenol moiety in the presence of 1-(3-dimethylamino-
propyl)-3-ethylcarbodiimide hydrochloride in CH2Cl2. to give the
desired cryptophane 1.
Scheme 1. Synthesis of cryptophanol-A (4) and synthesis of the linker (7).
4.2.3.1. Synthesis of cryptophane 8.
3,6,9,12-Tetraoxatetra-
decanoic acid-14-[[(4-methylphenyl) sulfonyl]oxy]phenylmethyl
ester (7) (254 mg, 0.51 mmol, 1.5 equiv) was introduced in a
three-neck flask containing cesium carbonate (167 mg,
0.51 mmol), cryptophanol 4 (300 mg, 0.34 mmol) in freshly dis-
tilled DMF (12 mL). The mixture was stirred for 16 h under argon
atmosphere at 333 K. CH2Cl2 (25 mL) and brine (10 mL) were
added to the mixture. The aqueous layer was extracted twice with
CH2Cl2 (10 mL) and the combined organic layers were washed with
brine (5 Â 10 mL) and then dried over sodium sulfate. The solvent
was removed under reduced pressure to give a residue, which was
then purified on silica gel (a gradient of solvent Et2O/AcOEt was
used: 100/0; 80/20; 50/50; 0/100). Evaporation of the solvent gave
compound 8 as a white amorphous solid (282 mg, 69%). 1H NMR
(500 MHz, CDCl3, 298 K) d 7.33–7.32 (m, 5H, Ar), 6.77 (s, 1H; Ar),
6.74 (s, 1H; Ar), 6.73 (s, 1H; Ar), 6.72 (s, 1H; Ar); 6.71 (s, 2H; Ar),
6.67 (s, 1H; Ar); 6.66 (s, 1H; Ar), 6.65 (s, 2H; Ar), 6.64 (s, 1H; Ar),
5.16 (s, 2H; OCH2Bn), 4.59–4.53 (m, 6H; Ha), 4.19–.10 (m, 15H,
OCH2, OCH2COO); 3.93 (m, 1H); 3.80–3.66 (m, 30 H; OCH2 and
OCH3), 3.39–3.35 (m, 6H; He). 13C NMR (125.7 MHz, CDCl3,
298 K) d 170.32, 162.50, 149.85, 149.73, 149.61, 149.56, 148.96,
147.20, 146.72, 146.67, 146.62, 146.59, 135.45, 134.31, 134.26,
134.12, 134.04, 133.99, 132.72, 131.83, 131.57, 131.54, 131.47,
131.36, 128.60, 128.42, 128.39, 122.30, 121.39, 121.26, 120.54,
120.47, 120.33, 117.03, 114.64, 113.70, 113.59, 70.94, 70.86,
70.75, 70.65, 69.88, 69.60, 69.53, 69.38, 69.29, 69.20, 68.73,
68.64, 68.48, 56.09, 55.73, 55.66, 55.64, 55.59, 36.46, 36.30. HRMS
stirred for an additional 1 h. A solution of bromoacetic acid (2.8 g,
20 mmol, 1.1 equiv) in THF was then added dropwise. After addi-
tion the solution was heated overnight under reflux conditions.
The solution was allowed to reach room temperature and benzyl
bromide (2.3 mL, 19.3 mmol, 1.1 equiv) was added in one portion.
The solution was then heated overnight under reflux conditions.
After evaporating the solvent under reduced pressure, CH2Cl2
(250 mL) and water (100 mL) were added. The layers were sepa-
rated and the aqueous layer was extracted twice with CH2Cl2
(100 mL). The combined organic layers were then washed twice
with water (50 mL) and dried over sodium sulfate. After evapora-
tion of the solvent the crude product was purified by column chro-
matography (AcOEt) to give 5 as a colorless oil (5.75 g, 75%). 1H
NMR (500 MHz, CDCl3, 298 K) d 7.35–7.28 (m, 5H; Ar), 5.15 (s,
2H; OCH2Bn), 4.59 (m, 1H, OCHO), 4.17 (s, 2H, OCH2COO), 3.85–
3.80 (m, 2H; OCH2), 3.71–3.69 (m, 2H, OCH2), 3.66–3.60 (m, 12H,
OCH2), 3.59–3.54 (m, 1H, OCH2), 3.48–3.44 (m, 1H, CH2), 1.82–
1.76 (m, 1H, CH2), 1.71–1.65 (m, 1H; CH2), 1.60–1.44 (m, 4H;
CH2). 13C NMR (125.7 MHz, CDCl3, 298 K)
d 170.26, 135.38,
128.53, 128.35, 128.33, 98.85, 70.88, 70.57, 70.55, 70.53, 70.50,
70.46, 68.61, 66.57, 66.41, 62.12, 30.49, 25.36, 19.41. HRMS
[M+Na]+. Calcd 449.2117. Found 449.2142.
4.2.2.2. 3,6,9,12-Tetraoxatetradecanoic acid-14-hydroxyphe-
nylmethyl ester (6).
A stirred solution of 5 (4.98 g,
11.3 mmol, 1 equiv) and pyridinium toluene sulfonate (0.95 g,
3.78 mmol, 0.33 equiv) in ethanol was heated overnight at 318 K.
The solvent was removed under reduced pressure and the oily res-
idue purified by column chromatography (AcOEt). Evaporation of
the solvent gave 6 as a colorless oil (3.2 g; 83%). Spectroscopic data
are identical to those previously reported in the literature.35
[M+Na]+ Calcd 1227.4949. Found 1227.4935.
.
4.2.3.2. Synthesis of cryptophane 9.
H2 gas was introduced
into a 25 mL flask containing compound 8 (560 mg, 0.465 mmol),
CH2Cl2 (12 mL), ethanol (2 mL), and Pd/C (70 mg, 0.0658 mmol,
0.14 equiv). The mixture was stirred at room temperature for 5 h.
After completion of the reaction the mixture was filtered and the
solid residue was washed with CH2Cl2 (2 Â 10 mL). The solvent
was then removed under reduced pressure to give the carboxylic
acid derivative 9 (0.46 g; 89%). 1H NMR (500 MHz, CDCl3, 298 K)
d 6.77 (s, 1H; Ar), 6.74 (s, 1H; Ar), 6.73 (s, 3H; Ar), 6.725 (s, 1H;
4.2.2.3. 3,6,9,12-Tetraoxatetradecanoic acid-14-[[(4-methyl-
phenyl)sulfonyl]oxy]-phenylmethyl ester (7).
Para-chloro
toluene sulfonate (1.5 g, 7.87 mmol, 1.3 equiv) was added at
273 K in one portion to a stirred solution of 6 (2 g, 5.84 mmol) in

C. Boutin et al. / Bioorg. Med. Chem. 19 (2011) 4135–4143
4141
by preparative TLC (AcOEt/Et2O: 70/30). From this procedure
3 mg of the desired cryptophane moiety 1 was isolated and used
without further treatment for the coupling reaction with the
protein.
4.3. Preparation of the biosensors
The human transferrin and BSA were from Sigma–Aldrich.
Rhodamine green succinimide ester was from Molecular Probes.
4.3.1. Grafting of cryptophane
The proteins were incubated with the activated cryptophane for
1 h in phosphate buffered saline (PBS) pH 7.4 at room temperature
with stirring. The unreacted cryptophane was finally removed from
the protein solution using Sephadex G25 medium gel chromatog-
raphy. The elution was performed with PBS, pH 7.4.
4.3.2. Grafting of Rhodamine green
The proteins were incubated with the fluorescent dye in bicar-
bonate medium pH 8.4 for 1 h, at room temperature with stirring.
The unreacted dye was removed from the protein solution using
Sephadex G25 medium gel chromatography. The elution was per-
formed with PBS, pH 7.4. Labeling efficiency was checked using an
absorption spectrometer.
4.3.3. Construction of the various biosensors
B1 was obtained after the incubation of 5 molecules of crypto-
phane per molecule of transferrin; B2, after the incubation of two
molecules of cryptophane and two molecules of rhodamine green
per molecule of transferrin; B3, after the incubation of five mole-
cules of cryptophane per molecule of BSA; B4, after the incubation
of two molecules of cryptophane and 2 molecules of rhodamine
green per molecule of BSA.
4.4. Characterization of the constructs
The folding of the biosensors was checked by NMR on a
700 MHz Bruker spectrometer equipped with an HCN cryoprobe.
Solutions at 1 mM in 500 lL H2O:D2O 90:10 were analyzed in
Scheme 2. Synthesis of the activated cryptophane 1 from cryptophanol 4.
5 mm tubes. The presence of high field 1H resonances and the
chemical shift dispersion of the amide and aromatic 1H resonances
were used as an indicator of the preservation of the global fold of
the protein. A maximum of ten cryptophanes could be tethered
to transferrin. Beyond this value, the protein unfolds. A statistical
distribution of the number of cryptophanes and of their locations
on the protein was expected; however mass spectrometry data
indicated a narrow distribution of the number of grafted crypto-
phanes. In addition, 1H translational self-diffusion experiments
were performed on a 500 MHz Avance II Bruker spectrometer
equipped with a TBI HN-broadband probehead and a BGU II gradi-
ent unit. These diffusion experiments at different concentrations
were performed to confirm that cryptophanes were covalently
linked and not trapped inside anfractuosities of the protein.
Ar), 6.72 (s, 2H; Ar), 6.69 (s, 1H; Ar), 6.66 (s, 1H; Ar), 6.65 (s, 3H;
Ar); 4.59–4.53 (m, 6H; Ha); 4.21–4.12 (m, 15H; OCH2 and OCH2-
COO), 3.95 (m, 1H), 3.82 (t, 2H, 3 J(H,H) = 5.0 Hz), 3.78–3.73 (m,
27 H, OCH2–OCH3), 3.41–3.36 (m, 6H, He). 13C NMR (125.7 MHz,
CDCl3, 298 K) d 170.99, 149.83, 149.71, 149.63, 149.59, 149.55,
148.91, 147.21, 146.65, 146.60, 134.39, 134.26, 134.16, 134.09,
134.00, 132.75, 131.87, 131.59, 131.54, 131.40, 122.17, 121.28,
120.59, 120.54, 120.38, 117.12, 114.72, 113.72, 113.61, 71.59,
70.88, 70.76, 70.64, 70.37, 70.18, 69.85, 69.59, 69.53, 69.38,
69.31, 69.23, 69.13, 69.72, 56.16, 55.74, 55.66, 55.64, 55.60,
36.22, 36.20, 36.15. HRMS [M+Na]+. Calcd 1137.4460. Found
1137.4452.
4.2.3.3. Cryptophane 1: activation of the carboxylic acid func-
4.5. Cells
tion of 9.
Cryptophane 9 (76 mg, 0.068 mmol, 1 equiv) penta-
fluorophenol (0.019 g, 0.1 mmol; 1.5 equiv) and CH2Cl2 (0.5 mL)
were introduced into a 10 mL three-neck flask. The mixture was
stirred for 15 min at room temperature and then 1-(3-dimethyl-
aminopropyl)-3-ethylcarbodiimide (13 mg, 0.068 mmol; 1 equiv)
was added in one portion. The reaction mixture was stirred over-
night at room temperature. CH2Cl2 (5 mL) and a saturated solution
of NaHCO3 (2 mL) were then added to the reaction mixture. The
two layers were separated and the organic layer was washed with
brine and then dried over sodium sulfate. Evaporation of the sol-
vent under reduced pressure leaves a residue, which was purified
4.5.1. Cell culture
Human erythroleukemic K562 cells from ATCC were grown in
IMDM medium (Sigma–Aldrich) supplemented with 10% heat-
inactivated fetal bovine serum, 1% of L-glutamine, 100 U/ml peni-
cillin and 100 g/ml streptomycin. The cells were maintained in
exponential growth in a 5% CO2 incubator at 310 K. Prior to the
experiment, the cells were washed three times in PBS and viable
cells were counted using the trypan blue exclusion test. The chosen
number of cells was re-suspended to a final volume of 1.5 mL.

4142
C. Boutin et al. / Bioorg. Med. Chem. 19 (2011) 4135–4143
4.5.2. Fluorescent labeling
columns can be directly compared. Due to the high xenon magne-
tization available in the gaseous phase, the values given in Table 1
correspond to the sum of two non-selective spectra for the first col-
umn, and six selective sub-spectra (two tube shakings) for the sec-
ond and third columns.
Cells were incubated with 200 nM of fluorescent probe for 1 h.
The cells were washed 3 times with PBS at 277 K to block the
membrane dynamics. The cells were then were fixed on the glass
substrate using 5% PFA for 30 min at 277 K. The observation was
made using an inverted microscope.
Acknowledgments
4.5.3. Preparation of the pronase sample
The cells were treated with a 2 mg/mL solution of pronase for
30 min at 310 K. The cells were then re-washed three times with
PBS and incubated with 5 lM of the biosensor for 1 h at 310 K. Fi-
Support from the French Ministry of Research (project ANR-06-
PCVI-0023 BULPOXI) is acknowledged.
nally, the cells were washed three times at 277 K to block plasma
membrane dynamics, and re-suspended in a final volume of
1.5 mL. Their viability was checked through the trypan blue exclu-
sion test.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2011.05.002.
4.6. 129Xe NMR experiments
References and notes
4.6.1. Laser-polarized xenon
1. Walker, T. G.; Happer, W. Rev. Mod. Phys. 1997, 69, 629.
2. Comment, A.; Jannin, S.; Hyacinthe, J.-N.; Miéville, P.; Sarkar, R.; Ahuja, P.;
Vasos, P. R.; Montet, X.; Lazeyras, F.; Vallée, J.-P.; Hautle, P.; Konter, J. A.; van
den Brandt, B.; Ansermet, J.-Ph.; Gruetter, R.; Bodenhausen, G. Phys. Rev. Lett.
2010, 105, 018104.
Xenon 86%-enriched in isotope 129 was from Cortecnet, France.
The hyperpolarized gas was prepared by the spin-exchange method,
using our batch apparatus described previously.37,5 The amount of
gas produced with this set-up using a titanium:sapphire laser, on
the order of 1 mL (in ꢀ10 minutes), was sufficient to fill the NMR
tube with a pressure of ꢀ1 bar and an average polarization of 40%.
The transfer of xenon between the cold finger designed to separate
xenon from nitrogen after optical pumping and the NMR tube of
interest was made through a vacuum line in the fringe field of the
NMR magnet in order to preserve polarization.
3. Brunner, E. Concepts Magn. Reson. 1999, 11, 313.
4. Spence, M. M.; Rubin, S. M.; Dimitrov, I. E.; Ruiz, E. J.; Wemmer, D. E.; Pines, A.;
Qin Yao, S.; Tian, F.; Schultz, P. G. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 10654.
5. Berthault, P.; Huber, G.; Desvaux, H. Prog. NMR Spectrosc. 2009, 55, 35.
6. Brotin, T.; Dutasta, J.-P. Chem. Rev. 2009, 109, 88.
7. Lerouge, F.; Melnyk, O.; Durand, J.-O.; Raehm, L.; Berthault, P.; Huber, G.;
Desvaux, H.; Constantinesco, A.; Choquet, P.; Detour, J.; Smaïhi, M. J. Mater.
Chem. 2009, 19, 379.
8. Berthault, P.; Bogaert-Buchmann, A.; Desvaux, H.; Huber, G.; Boulard, Y. J. Am.
Chem. Soc. 2008, 130, 16456.
9. Spence, M. M.; Ruiz, E. J.; Rubin, S. M.; Lowery, T. J.; Winssinger, N.; Schultz, P.
G.; Wemmer, D. E.; Pines, A. J. Am. Chem. Soc. 2004, 126, 15287.
10. Roy, V.; Brotin, T.; Dutasta, J.-P.; Charles, M.-H.; Delair, T.; Mallet, F.; Huber, G.;
Desvaux, H.; Boulard, Y.; Berthault, P. ChemPhysChem 2007, 8, 2082.
11. Wei, Q.; Seward, G. K.; Hill, P. A.; Patton, B.; Dimitrov, I. E.; Kuzma, N. N.;
Dmochowski, I. J. J. Am. Chem. Soc. 2006, 128, 13274.
12. Chambers, J. M.; Hill, P. A.; Aaron, J. A.; Han, Z.; Christianson, D. W.; Kuzma, N.
N.; Dmochowski, I. J. J. Am. Chem. Soc. 2009, 131, 563.
13. Schlundt, A.; Kilian, W.; Beyermann, M.; Sticht, J.; Günther, S.; Höpner, S.; Falk,
K.; Roetzschke, O.; Mitschang, L.; Freund, C. Angew. Chem., Int. Ed. 2009, 48,
4142.
14. Seward, G. K.; Wei, Q.; Dmochowski, I. J. Bioconjug. Chem. 2008, 19, 2129.
15. Du, X.-L.; Wang, K.; Ke, Y.; Yuan, L.; Li, R.-C.; Chang, Y. Z.; Ho, K. P.; Qian, Z. M. J.
Cell. Physiol. 2004, 201, 45.
16. Klausner, R. D.; Ashwell, G.; van Renswoude, J.; Harford, J. B.; Bridges, K. R. Proc.
Natl. Acad. Sci. U.S.A. 1983, 80, 2263.
4.6.2. 129Xe hyperpolarized NMR experiments
The experiments were run on a 500 MHz Avance II Bruker spec-
trometer equipped with a 129Xe/1H micro-imaging probehead (Mi-
cro-5) designed for tubes with a maximum outer diameter of
8 mm. Prior to the NMR experiments, the samples inside the
NMR tubes were degassed by the following procedure: helium
bubbling then static evacuation of the gas on top of the solutions
in a vacuum line. Each series of acquisitions was preceded by shak-
ing (to quickly dissolve the hyperpolarized gas) and a wait time of
ꢀ10 s to enable the elimination of bubbles.
For each experiment, that is, for each cell or supernatant sam-
ple, exactly the same procedure was employed: after the shaking
of the tube, a first 129Xe spectrum acquired with non-selective
pulse of small flip angle (ꢀ4°) was recorded in one scan. This
was immediately followed by three series of sub-spectra obtained
by selective excitation centered on the Xe@cryptophane region,
then detection.5 The selective excitation was performed using a
17. Klausner, R. D.; van Renswoude, J.; Ashwell, G.; Kempf, C.; Schechter, A. N.;
Dean, A.; Bridges, K. R. J. Biol. Chem. 1983, 258, 4715.
18. Lowery, T. J.; Garcia, S.; Chavez, L.; Ruiz, E. J.; Wu, T.; Brotin, T.; Dutasta, J.-P.;
King, D. S.; Schultz, P. G.; Pines, A.; Wemmer, D. E. ChemBioChem 2006, 7, 65.
19. The chirality of the resulting molecules arises from the chirality of the
cryptophane (see Ref. 6). As the grafting of the PEG group was made on a
racemic mixture of cryptophanes, the risk was to obtain two signals for caged
xenon. Putting the first asymmetric carbon far from the cavity where xenon
will reside, promises a non-significant chemical shift variation for xenon caged
in both diastereomers.
20. Partis, M. D.; Griffiths, D. G.; Roberts, G. C.; Beechey, R. B. J. Protein Chem. 1983,
2, 263.
21. Wang, W.; Ackermann, D.; Mehlich, A.-M.; König, S. Proteomics 2010, 10, 1525.
22. Kida, S.; Maeda, M.; Hojo, K.; Eto, Y.; Nakagawa, S.; Kawasaki, K. Chem. Pharm.
Bull. 2007, 55, 685.
1% truncated Gaussian pulse of 500 ls. Three series of 128 scans
differing by the repetition time (70 ms, 100 ms, 170 ms) allowed
us to eliminate the potentially different xenon in-out exchange
rates. This set of four experiments could be repeated many times
after shaking of the tube as long as the reservoir of gaseous xenon
gave significant polarization to the dissolved xenon.
23. Sun, H.; Li, H.; Sadler, P. J. Chem. Rev. 1999, 99, 2817.
24. Pinaud, F.; Clarke, S.; Sittner, A.; Dahan, M. Nature. Meth. 2010, 7, 275.
25. Qian, Z. M.; Morgan, E. H. Biochim. Biophys. Acta 1991, 1073, 456.
The spectra were processed as follows: a line broadening of
20 Hz (no zero-filling) was applied before Fourier transformation.
As exactly the same conditions (same receiver gain, in particular)
were used for the non-selective spectra, the integration values
can be directly compared (first column of Table 1). Except for the
repetition time, the sub-spectra were also acquired in the same
experimental conditions, but obviously here the receiver gain
was different from that of the non-selective experiments. Within
each series the three sub-spectra were summed, and the peak inte-
gral values were divided by the integral value of the peak in the
corresponding non-selective experiment, in order to eliminate
the polarization fluctuations. So the values of the second and third
26. Despite
a xenon polarization level not fully reproducible, quantitative
estimation of the xenon NMR signals has been rendered possible through
normalization of the integral of the peak of xenon free in solution, considering
that the same amount of xenon was each time introduced in the NMR tube
(checked by weighting of the tube before and after the experiment).
27. Meldrum, T.; Schröder, L.; Denger, P.; Wemmer, D. E.; Pines, A. J. Magn. Reson.
2010, 205, 242.
28. Boutin, C.; Desvaux, H.; Carrière, M.; Leteurtre, F.; Jamin, N.; Boulard, Y.;
Berthault, P. accepted in NMR Biomed. doi:10.1002/nbm.1686.
29. Berthault, P.; Desvaux, H.; Wendlinger, T.; Gyejacquot, M.; Stopin, A.; Brotin, T.;
Dutasta, J.-P.; Boulard, Y. Chem. Eur. J. 2010, 16, 12941.
30. Darzac, M.; Brotin, T.; Bouchu, D.; Dutasta, J.-P. Chem. Commun. 2002, 48.

C. Boutin et al. / Bioorg. Med. Chem. 19 (2011) 4135–4143
4143
31. Darzac, M.; Brotin, T.; Rousset-Arzel, L.; Bouchu, D.; Dutasta, J.-P. New J. Chem.
35. Marchi-Artzner, V.; Lorz, B.; Hellerer, U.; Kantlehner, M.; Kessler, H.;
Sackmann, E. Chem. Eur. J. 2001, 7, 1095.
2004, 28, 502.
32. Sakellariou, E. G.; Montalban, A. G.; Beall, S. L.; Henderson, D.; Meunier, H. G.;
36. Marchi-Artzner, V.; Jullien, L.; Gulik-Krzywicki, T.; Lehn, J.-M. Chem. Commun.
1997, 117.
Phillips, D.; Suhling, K.; Barrett, A. G. M.; Hoffman, B. M. Tetrahedron 2003, 59,
9083.
37. Desvaux, H.; Gautier, T.; Le Goff, G.; Pétro, M.; Berthault, P. Eur. Phys. J. D 2000,
12, 289.
33. Barluenga, J.; de Prado, A.; Santamaría, J.; Tomás, M. Organometallics 2005, 24,
3614.
34. Kaufmann, G. F.; Mejler, M. M.; Sun, C. Z.; Chen, D. W.; Kujawa, D. P.; Mee, J. M.;
Hoffman, T. Z.; Wirsching, P.; Lerner, R. A.; Janda, K. D. Angew. Chem., Int. Ed.
2005, 44, 2144.