Cryptophane-xenon complexes in organic solvents observed through NMR spectroscopy

Article abstract of DOI:10.1021/jp807425t

The interaction of xenon with cryptophane derivatives is analyzed by NMR by using either thermal or hyperpolarized noble gas. Twelve hosts differing by their stereochemistry, cavity size, and the nature and the number of the substituents on the aromatic rings have been included in the study, in the aim of extracting some clues for the optimization of ¹²⁹Xe-NMR based biosensors derived from these cage molecules. Four important properties have been examined: xenon-host binding constant, in-out exchange rate of the noble gas, chemical shift, and relaxation of caged xenon. This work aims at understanding the main characteristics of the host-guest interaction in order to choose the best candidate for the biosensing approach. Moreover, rationalizing xenon chemical shift as a function of structural parameters would also help for setting up multiplexing applications. Xenon exhibits the highest affinity for the smallest cryptophane, namely cryptophane-111, and a long relaxation time inside it, convenient for conservation of its hyperpolarization. However, very slow in-out xenon exchange could represent a limitation for its future applicability for the biosensing approach, because the replenishment of the cage in laser-polarized xenon, enabling a further gain in sensitivity, cannot be fully exploited.

Full text of DOI:10.1021/jp807425t

J. Phys. Chem. A 2008, 112, 11363–11372
11363
Cryptophane-Xenon Complexes in Organic Solvents Observed through NMR Spectroscopy
Gaspard Huber,† Lætitia Beguin,† Herve´ Desvaux,† Thierry Brotin,‡ Heather A. Fogarty,‡
Jean-Pierre Dutasta,‡ and Patrick Berthault*,‡
CEA, IRAMIS, SerVice de Chimie Mole´culaire, Laboratoire Structure et Dynamique par Re´sonance
´
Magne´tique, URA CEA/CNRS 331, 91191Gif-sur-YVette, France, and CNRS, Ecole Normale Supe´rieure
de Lyon, Laboratoire de Chimie, 46 Alle´e d’Italie, F-69364 Lyon, France
ReceiVed: August 19, 2008; ReVised Manuscript ReceiVed: September 2, 2008
The interaction of xenon with cryptophane derivatives is analyzed by NMR by using either thermal or
hyperpolarized noble gas. Twelve hosts differing by their stereochemistry, cavity size, and the nature and the
number of the substituents on the aromatic rings have been included in the study, in the aim of extracting
some clues for the optimization of 129Xe-NMR based biosensors derived from these cage molecules. Four
important properties have been examined: xenon-host binding constant, in-out exchange rate of the noble
gas, chemical shift, and relaxation of caged xenon. This work aims at understanding the main characteristics
of the host-guest interaction in order to choose the best candidate for the biosensing approach. Moreover,
rationalizing xenon chemical shift as a function of structural parameters would also help for setting up
multiplexing applications. Xenon exhibits the highest affinity for the smallest cryptophane, namely cryptophane-
111, and a long relaxation time inside it, convenient for conservation of its hyperpolarization. However, very
slow in-out xenon exchange could represent a limitation for its future applicability for the biosensing approach,
because the replenishment of the cage in laser-polarized xenon, enabling a further gain in sensitivity, cannot
be fully exploited.
TABLE 1: Structure of the Cryptophanes 1-12 Considered
Introduction
in this Studya
Cryptophanes are hollow molecules made of two cyclotriben-
zylene (CTB) units linked by alkyldioxy chains of various
lengths.1 They can form host/guest complexes with small neutral
molecules and form the most stable complexes with xenon
observed so far. The smallest cryptophane was recently syn-
thesized (1) and shows a tremendous affinity for xenon, thus
prompting a renewal of interest in cryptophane gas trapping
behavior.2 It is the latest member of a large family of molecular
hosts, the generic structure of which is depicted in Table 1,
that vary in the length of the alkoxy linkers (cryptophane-A
(2),3 cryptophane-223 (3),4 cryptophane-233 (4),4 cryptophane-
224 (5),4 and cryptophane-E (6)5), in the substituents on the
aromatic rings (7-10), and in the stereochemistry of the alkoxy
linkers (cryptophane-C (anti) (11)6 and cryptophane-D (syn)
(12).6
compound
n
m
R1
R2
usual name
reference
1
2
3
4
5
6
7
8
1
2
2
3
2
3
2
2
2
2
2
2
1
2
3
2
4
3
2
2
2
2
2
2
H
H
Cryptophane-111
2
3
4
4
4
5
OCH3 OCH3 Cryptophane-A
OCH3 OCH3 Cryptophane-223
OCH3 OCH3 Cryptophane-233
OCH3 OCH3 Cryptophane-224
OCH3 OCH3 Cryptophane-E
H
H
OBn
OBu
H
Three effects can be observed in the 129Xe NMR spectra of
xenon-cryptophane complexes. First, because of the high
polarizability of the xenon electronic cloud, the chemical shift
of xenon can span several hundreds ppm depending on its local
environment. Second and as a consequence, in the vast majority
of experimental conditions, a slow exchange on the xenon
chemical-shift time scale occurs, and two signals are observable,
corresponding to free xenon (approximately 200 ppm downfield
from the external reference given by the gas resonance at quasi-
null pressure) and xenon bound to the cryptophane (in the
30-80 ppm region). Third, the xenon signal can be enhanced
by 4-5 orders of magnitude through optical pumping. Because
of this set of favorable physical properties, xenon-cryptophane
complexes constitute a good starting point for biosensing
H
OBn
OBn
OBu
9
10
11
12
OCH3 Cryptophane-C
OCH3 Cryptophane-D
6
6
H
a Compound 12 is the syn-isomer of compound 11. Bn, benzyl
group; Bu, butyl group.
applications based on high-resolution NMR7 or chemical-shift
imaging.8 For such applications, the host is modified to bear a
tethered ligand that is designed to recognize specific biological
receptors and thus to drag xenon to these receptors.9 The
important benefits of this approach are clearly the host affinity
for xenon but also the multiplexing capability of 129Xe MRI
that can be offered by chemical modification of the cage. The
availability of xenon carriers bearing different recognition
* To whom correspondence should be addressed. E-mail:
Patrick.Berthault@cea.fr.
† CEA, IRAMIS.
‡
´
CNRS, Ecole Normale Supe´rieure de Lyon.
10.1021/jp807425t CCC: $40.75
2008 American Chemical Society
Published on Web 10/17/2008

11364 J. Phys. Chem. A, Vol. 112, No. 45, 2008
Huber et al.
SCHEME 1: Synthesis of Cryptophanes 9 and 10 from 15
SCHEME 2: Synthesis of Cryptophane 7 from
Cryptophane 11
antennae will enable the detection of correlated biological
events. With this approach, careful design of the host architec-
ture can provide different chemical shifts for caged xenon,
thereby giving each recognition antenna a unique signature.
Obviously, rational development would benefit from a
detailed study of the host-xenon interactions for families of
cryptophanes differing by size or the nature of the substituents.
In this paper, several key parameters are examined, including
(i) the affinity of xenon for the cryptophane cage, as defined by
the binding constant, (ii) the chemical shift of caged xenon for
various cryptophanes, with the objective of performing multi-
plexing experiments, (iii) the xenon in-out exchange rate, which
should be slow on the xenon chemical-shift time scale but fast
enough for a continuous refilling of the cage by laser-polarized
xenon, and (iv) the longitudinal relaxation time which will affect
the ability to maintain the polarization of xenon atoms in the
cage. Therefore, the influence of the cryptophane isomerism,
the cryptophane size, and the number and nature of the aromatic
ring substituents on some of these experimental parameters has
been examined for 12 cryptophanes, 1-12, which are soluble
in organic solvents and the structures of which are given in
Table 1.
Materials and Methods
efficiency of the reaction but rather the difficulty of extracting
and purifying 16, which exhibits a very low solubility in both
organic and aqueous media. This compound was then reacted
with a slight excess of triflic anhydride in the presence of
pyridine in CH2Cl2 at 0 °C to give cryptophane 17 with 91%
yield (Scheme 2). Reduction of the triflate moieties was achieved
by using a mixture of palladium dichloride diphenylphosphine,
1,3-diphenylphosphine propane (dppp), tris-n-butylamine (n-
NBu3), and formic acid in DMF at 110 °C according to a known
procedure.1 The procedure was very efficient for the reduction
of hindered triflate groups, as observed recently for a cryp-
tophane bearing a single triflate moiety.12 Cryptophane 7 was
thus obtained with a good yield after purification by chroma-
tography on silica gel (91%).
2,7,12-Trimethoxy-10,15-dihydro-5H-tribenzo[a,d,g]cyclonon-
ene (13). 3-Methoxy benzyl alcohol (13.13 g, 95.1 mmol) was
added to a stirred mixture of P2O5 (21 g, 147 mmol) in dry
Et2O (450 mL); a weak reflux was observed. Stirring was
stopped, and the mixture was heated to 48 °C overnight. The
Et2O solution was poured into a separatory funnel, and the P2O5
was washed with CH2Cl2 (3 × 220 mL). The organic layers
were combined, and their volume was reduced by rotary
evaporation, then washed with H2O (2 × 220 mL), by adding
brine (80 mL) each time. The combined aqueous layers were
extracted with CH2Cl2, and the organic layers were combined
and dried over Na2SO4. The solvent was removed under reduced
pressure to provide a pale yellow wax. Column chromatography
(CH2Cl2/pentane 3:2), followed by rinsing the solid placed on
a fritted filter with Et2O, gave 13 as a white powder (6%
yield).11,13,14 1H NMR (200 MHz, CDCl3, 20 °C) δ 7.28 (d,
Xenon in natural isotopic proportions was obtained from Air
Liquide, and xenon enriched at 96% in isotope 129 was obtained
from Chemgas. 1,1,2,2-tetrachloroethane-d2 (99.60% deuterium),
dichloromethane-d2 (99.96% deuterium), chloroform-d (99.80%
deuterium), and 1,2-dichloroethane-d4 (99% deuterium) were
from Eurisotop. Carbon tetrachloride was from Janssen and 1,2-
dibromoethane from Fluka.
Synthesis of Cryptophanes. In contrast to the synthetic route
published recently,2 a novel approach is used for the synthesis
of cryptophane 1, which provides the cycloanisylene unit 13 in
a single step (6% yield) and cryptophane 1 in two further steps
(see Figure S1 in the Supporting Information). These two
synthetic steps require a removal of the methoxy moieties with
boron tribromide to give rise to the cyclotriphenolene, 14,
followed by the reaction of 14 with 1.5 equivs of bromochlo-
romethane to give cryptophane 1 in 19% yield. Cryptophanes
2-6 were prepared according to published procedures using
the template method.4,5,10 Cryptophane 8 was prepared from a
multistep synthesis according to a known experimental proce-
dure.11 New cryptophanes 9 and 10 were synthesized in good
yield from cryptophanol 15 by using benzylbromide or 1-bro-
mobutane in the presence of Cs2CO3 in DMF, respectively (see
Scheme 1). Cryptophanes 11 and 12 were both synthesized in
a single step by using scandium triflate as a catalyst.11
Cryptophane 7 was prepared by using a three-step synthetic
route from cryptophane 11 (32% overall yield). Cryptophane
11 was first allowed to react with lithium diphenylphosphide
in THF to give the cryptophanol 16 in moderate yield (nonop-
timized: 40%). This moderate yield does not reflect the

Cryptophane-Xenon Complexes in Organic Solvents
J. Phys. Chem. A, Vol. 112, No. 45, 2008 11365
portion of signal hidden by CDCl3 signal, 3 H, Ar), 6.86 (d,
as a white residue, which was purified through a short column
of silica gel (CH2Cl2/acetone: 90/10 then 80/20). Evaporation
of the solvent gives compound 16 as a white product, which
was washed on a frit with Et2O/CH2Cl2 (50/50), then with pure
Et2O. This procedure gives rise to compound 16 (0.09 g; 40%)
3
4J(H,H) ) 2.7 Hz, 3 H, Ar), 6.63 (d of d, J(H,H) ) 8.5 Hz,
2
4J(H,H) ) 2.7 Hz, 3 H, Ar), 4.73 (d, J(H,H) ) 13.5 Hz, 3 H,
2
Ha), 3.72 (s, 9 H; OCH3), 3.62 (d, J(H,H) ) 13.6 Hz, 3 H,
He).
1
as a white solid. Mp: 296 °C. H NMR (500 MHz, CDCl3, 20
10,15-Dihydro-5H-tribenzo[a,d,g]cyclononene-2,7,12-triol
(14). Boron tribromide (37.0 mL of a 1 M solution of BBr3 in
CH2Cl2) was added dropwise to a solution of 2,7,12-trimethoxy-
10,15-dihydro-5H-tribenzo[a,d,g]cyclononene (2.5 g, 6.9 mmol)
in dry CH2Cl2 (64 mL) at -78 °C, and the solution was stirred
for 5 min, before being warmed to room temperature and stirred
overnight. The solution was poured onto an ice and H2O slurry.
H2O, then acetone (200 mL) and CH2Cl2 (200 mL), were added,
and the solution was transferred to a separatory funnel. The
organic layer was washed with H2O (1 × 170 mL). The
combined aqueous layers were extracted with CH2Cl2 (180 mL)
and acetone (30 mL). The combined organic layers were dried
over Na2SO4. The solvent was removed under reduced pressure
to provide a beige solid. Column chromatography (CH2Cl2/
acetone 3:2) provided an off-white powder (95% yield).13,14
3
°C): δ ) 7.10 (d, 3H, J(H,H) ) 8.5 Hz; Ar), 6.73 (d, 3H,
4J(H,H) ) 2.6 Hz; Ar), 6.67 (s, 3H; Ar), 6.52 (s, 3H; Ar), 6.37
(dd, 3H, 3J(H,H) ) 8.5 Hz, 4J(H,H) ) 2.6 Hz; Ar), 5.28 (s, 3H;
OH), 4.59 (d, 3H, 2J(H,H) ) 13.6 Hz; Ha), 4.52 (d, 3H, 2J(H,H)
) 13.6 Hz; Ha), 4.40-4.30 (m, 6H; CH2), 4.28-4.25 (m, 3H;
2
CH2), 3.86-3.82 (m, 3H; CH2), 3.51 (d, 3H, J(H,H) ) 13.6
Hz; He), 3.31 (d, 2J(H,H) ) 13.6 Hz, 3H, He). 13C NMR (126.7
MHz, CDCl3, 20 °C) δ ) 154.9 (3C), 144.6 (3C), 144.5 (3C),
140.8 (3C), 133.0 (3C), 131.1 (3C), 130.8 (3C), 130.7 (3C),
119.2 (3C), 115.0 (3C), 114.1 (3C), 109.5 (3C), 67.3 (3C;
OCH2), 64.5 (3C; OCH2), 35.8 (3C; Ca,e), 35.6 (3C; Ca,e). HRMS
(LSIMS), calcd for C48H42O9Na [M°+], 785.2727; found 785.2736.
Cryptophane 17. Triflic anhydride (0.25 g, 0.15 mL) was
added dropwise at 0 °C to a stirred solution of cryptophan-ol
16 (0.076 g, 0.1 mmol) in pyridine (2 mL) and CH2Cl2 (2 mL).
The orange solution was stirred overnight at 20 °C and then
poured in water. The product was extracted three times with
CH2Cl2. The combined organic layers were washed with brine
and dried over Na2SO4. Evaporation of the solvent under reduce
pressure leaves a red residue. Purification by chromatography
on silica gel (CH2Cl2/acetone: 90/10) gives cryptophane 17 as
a white glassy product (0.105 g, 91%). Decomposes above 300
1
Additionnal information: H NMR (200 MHz, acetone-d6, 20
°C) δ 7.21 (d, 3H, 3J(H,H) ) 8.0 Hz, Ar), 6.88 (d, 3 H, 4J(H,H)
) 2.6 Hz, Ar), 6.55 (dd, 3H, 3J(H,H) ) 8.0 Hz, 4J(H,H) ) 2.6
2
H, Ar), 4.78 (d, 3 H, J(H,H) ) 13.4 Hz, Ha), 3.55 (d, 3 H,
2J(H,H) ) 13.4 Hz, He).
Cryptophane 1. 10,15-Dihydro-5H-tribenzo[a,d,g]cyclonon-
ene-2,7,12-triol (1.01 g, 3.14 mmol) was placed in a flask
containing dry Cs2CO3 (9.65 g, 0.34 mmol); then, dry DMF
(40 mL) was cannulated into the flask, and the mixture was
heated to 60 °C under stirring. BrCH2Cl (0.73 g, 0.37 mL) in
DMF (100 mL) was placed in a dropping funnel, and 1-2 drop/
min of the solution was added to the mixture at 60 °C; stirring
was continued overnight after addition was complete. TLC (9:1
CH2Cl2/acetone) indicated no starting material remained. DMF
was removed while heating under reduced pressure. CH2Cl2 (220
mL) and H2O (300 mL) were added to the beige solid. The
aqueous layer was extracted with CH2Cl2 (2 × 240 mL), and
the organic layers were combined and washed with H2O (250
mL). The combined organic layers were dried over Na2SO4.
The solvent was removed under reduced pressure to provide
an ochre solid. Column chromatography (CH2Cl2/acetone gradi-
ent 98:2 to 70:30), was followed by a second column (CHCl3:
100%). Then, the white solid was rinsed with Et2O onto a fritted
filter to provide a white powder. Yield 19%. Decomposes above
200 °C; 1H NMR (500 MHz, CDCl3, 20 °C) δ 6.967 (s broad,
6 H, Ar), 6.724 (s broad, 6 H; Ar), 6.570 (s broad, 6 H; Ar),
1
°C. H NMR (500 MHz, CDCl3, 20 °C): δ ) 7.10 (d, 3H,
3J(H,H) ) 7.0 Hz; Ar), 7.00 (s, 3H; Ar), 6.76 (s, 3H; Ar), 6.66
(s, 3H; Ar), 6.39 (d, 3H, 3J(H,H) ) 7.0 Hz; Ar), 4.60 (d, 2J(H,H)
2
) 14.0 Hz, 3H; Ha), 4.57 (d, 3H, J(H,H) ) 14.0 Hz, Ha),
4.40-4.25 (m, 9H; CH2), 3.97 (m, 3H; CH2), 3.50 (d, 3H,
2
2J(H,H) ) 14.0 Hz; He), 3.46 (d, 3H, J(H,H) ) 14.0 Hz; He).
13C NMR (126.7 MHz, CDCl3, 20 °C) δ ) 155.6 (3C), 149.4
(3C), 140.8 (3C), 139.5 (3C), 131.3 (3C), 131.2 (3C), 131.1
1
(3C), 123.4 (3C), 119.6 (3C), 118.9 (q, 3C, F(C,F) ) 324.0
Hz, CF3), 117.0 (3C), 110.3 (3C), 67.4 (3C; OCH2), 64.3 (3C;
OCH2), 36.1 (3C; Ca,e), 35.8 (3C; Ca,e). HRMS (LSIMS), calcd
for C51H39O15F9S3Na [M°+] 1181.1205; found 1181.1136.
Cryptophane 7. Cryptophane 17 (0.1 g, 0.086 mmol),
PdCl2(PPh3)2 (0.061 g, 0.086 mmol), dppp (0.071 g, 0.17 mmol),
DMF (1.5 mL), NBu3 (1 mL), and formic acid (0.1 mL) were
added in this order to a 5 mL round-bottom flask. The mixture
was stirred overnight at 110 °C under an argon atmosphere;
then, the mixture was poured in a mixture of CH2Cl2 and water.
The product was extracted three times with a small amount of
CH2Cl2. The combined organic layers were washed once with
brine and then dried over Na2SO4. Filtration and evaporation
of the solvent under reduced pressure leaves a dark red residue
(at the end, the vacuum pump is necessary to remove the NBu3).
Cryptophane 7 is isolated by column chromatography (CH2Cl2/
acetone: 90/10). The solid is washed on a frit with diethyl ether
to give compound 7 as white crystals (0.056 g, 91%). Decom-
poses above 300 °C. 1H NMR (500 MHz, CDCl3, 20 °C): δ )
7.09 (d, 6H, 3J(H,H) ) 8.5 Hz; Ar), 6.66 (d, 6H, 4J(H,H) ) 2.5
2
5.71 (s, 6 H; OCH2O), 4.48 (d, 6H, J(H,H) ) 13.5 Hz; Ha),
2
3.41 (d, 6H, J(H,H) ) 13.5 Hz; He); 13C NMR (CDCl3) δ )
157.77 (6 C, Ar), 140.76 (6 C, Ar), 132.13 (6 C, Ar), 129.85(6
C, Ar), 117.59 (6 C, Ar), 111.68 (6C, Ar), 86.17 (3C, OCH2O),
35.30 (6 C, Ca,e); IR (KBr pellet) 754, 827, 1039, 1149, 1232,
1300, 1321, 1421, 1441, 1498, 1579, 1604, 2920, 2983 cm-1
;
UV (hexanes) (λ) 242 nm (16 800 L mol-1 cm-1), 279 nm (5
500 L mol-1 cm-1), 288 nm (6 900 L mol-1 cm-1); HRMS
(TOF MS ES+) m/z calcd for C45H36O6Na 695.2410; found
695.2397.
Cryptophan-ol (16). Lithium diphenyphosphide 1 M (3.4 mL,
3.35 mmol) was added dropwise to a stirred mixture of
cryptophane 11 (0.25 g, 0.3 mmol) in dry THF (5 mL). The
resulting red mixture was stirred overnight at 60 °C under an
argon atmosphere. The mixture was then poured in water and
extracted three times with CH2Cl2. The combined organic layers
were washed three times with brine and dried over Na2SO4.
Filtration and evaporation of the solvent leaves compound 16
3
4
Hz; Ar), 6.48 (dd, 6H, J(H,H) ) 8.5 Hz, J(H,H) ) 2.5 Hz;
Ar), 4.62 (d, 6H, 2J(H,H) ) 13.5 Hz; Ha), 4.17 (m, 12H; OCH2),
3.48 (d, 6H, J(H,H) ) 13.5 Hz, He). 13C NMR (126.7 MHz,
2
CDCl3, 20 °C) δ ) 156.6 (6C), 140.7 (6C), 131.4 (6C), 130.6
(6C), 116.8 (6C), 113.3 (6C), 65.2 (6C; OCH2), 36.2 (6C;
OCH2). HRMS (LSIMS), calcd for C48H42O6Na [M°+] 737.2879;
found 737.2883.

11366 J. Phys. Chem. A, Vol. 112, No. 45, 2008
Huber et al.
Cryptophane 9. Benzyl bromide (0.22 g, 0.2 mL, 1.98 mmol)
was added to a stirred solution of cryptophan-ol 15 (0.1 g, 0.12
mmol), Cs2CO3 (0.28 g, 0.86 mmol) in DMF (3.5 mL). The
mixture was stirred for 48 h at 80 °C under an argon atmosphere.
The mixture was poured in water and extracted three times with
CH2Cl2. The combined organic layers were washed three times
with brine and then dried over Na2SO4. Evaporation under
reduced pressure left a residue, which was passed through a
column of silica gel (CH2Cl2/acetone: 90/10). Several fractions
containing the product were evaporated under reduced pressure
to give cryptophane 9 as a white glassy product. Recrystalli-
zation in a mixture of CHCl3 and EtOH gave 9 (0.1 g; 62%) as
exchange method,16,17 to approximately 30% via our home-built
apparatus.18 Briefly, the D1 electronic transition of rubidium was
excited by using a tunable titanium:sapphire laser, producing
approximately 6 W continuous power when pumped by a 25
W argon ionized laser. The light crossed a beam expander, a
beam splitter cube and a λ/4 plate, ensuring purely circularly
polarized light. Then, the beam reached the pumping cell
containing 0.13-2.7 kPa xenon, 27 kPa nitrogen, and some
droplets of rubidium, partially vaporized by heating to ap-
proximately 368 K. After approximately 10 min of optical
pumping, hyperpolarized xenon was condensed in a tube
immersed in liquid nitrogen, in a magnetic field of 0.3 T created
by a solenoid, whereas gaseous nitrogen was evacuated toward
a pump. The xenon was then transported to the NMR
spectrometer.
Sample Preparation and NMR Spectroscopy. Unless
otherwise specified, individual cryptophanes or mixtures of
cryptophanes were dissolved in 1,1,2,2-tetrachloroethane-d2 in
the 1-25 mM range, depending on the parameter to be
measured. Samples were degassed by several freeze-pump-
thaw cycles. Polarized or thermal xenon was then added to the
NMR tube by condensation (by means of a hollow spinner filled
with liquid nitrogen) or by expansion, according to the desired
final xenon pressure in the NMR tube. NMR spectra dedicated
to the characterization of xenon-cryptophane binding were
recorded on a 11.7 T Bruker Avance II spectrometer equipped
with either a Bruker broadband inverse 5 mm probehead or a
Nalorac broadband direct 5 mm probehead.
Xenon-Cryptophane Binding Constant Determination.
The xenon-cryptophane binding constant for 219 and 64 had
been determined by using a 1:1 stoechiometry (only one xenon
atom able to enter in the cage). This assumption may be safely
extended to all other cryptophanes reported here, because their
cavity volume is smaller than that of 6. Although other
parameters, such as the nature of the solvent, influence
xenon-cryptophane binding, an apparent binding constant Ki
may be defined as Ki ) [Xe@Ci]/[Xe][Ci], where [Ci] and
[Xe@Ci] are the concentrations of the free and bound forms of
cryptophane i, respectively, enabling comparisons between
cryptophanes in the same solvent.
1
white crystals. Mp: 170 °C. H NMR (500 MHz, CDCl3, 20
°C): δ ) 7.3-7.2 (m, 15H; Bn), 6.77 (s, 6H; Ar), 6.70 (s, 6H;
2
Ar), 4.78 (m, 12H; OCH2), 4.52 (d, 6H, J(H,H) ) 13.5 Hz;
2
Ha), 4.3-4.2 (m, 12H; OCH2), 3.33 (d, 6H, J(H,H) ) 13.5
Hz; He). 13C NMR (126.7 MHz, CDCl3, 20 °C) δ ) 149.0 (6C),
147.5 (6C), 137.1 (6C), 134.2 (6C), 132.7 (6C), 128.4 (12C),
127.8 (6C), 127.4 (12C), 121.8 (6C), 117.8 (6C), 71.8 (6C;
OCH2), 69.6 (6C; OCH2), 36.2 (6C; Ca,e). HRMS (LSIMS),
calcd for C90H78O12Na [M°+] 1373.5391; found 1373.5315.
Cryptophane 10. Cryptophan-ol 15 (0.1 g, 0.12 mmol),
Cs2CO3 (0.48 g, 1.47 mmol), and DMF (6 mL) were placed in
a three-neck flask equipped with a reflux condenser. Then,
1-bromobutane (0.152 g, 0.12 mL, 1.1 mmol) was added in one
portion. The mixture was stirred overnight at 80 °C under an
argon atmosphere and then poured in water. The product was
extracted three times with CH2Cl2. The combined organic layers
were washed several times with brine and dried over Na2SO4.
Filtration followed by evaporation of the solvent under reduced
pressure leaves a residue that was purified by column chroma-
tography (CH2Cl2/acetone: 90/10) to provide cryptophane 10
as a white glassy product. Recrystallization in a mixture of
CHCl3 and EtOH gave cryptophane 10 as white crystals (0.11
1
g; 80%). Mp: 281 °C. H NMR (500 MHz, CDCl3, 20 °C): δ
2
) 6.72 (s, 6H; Ar), 6.66 (s, 6H; Ar), 4.53 (d, 6H, J(H,H) )
13.5 Hz; Ha), 4.15 (m, 6H), 4.05 (m, 6H), 3.94 (m, 6H), 3.80
2
(m, 6H), 3.36 (d, 6H, J(H,H) ) 13.5 Hz; He), 1.75 (p, 12H,
3J(H,H) ) 7.0 Hz; CH2), 1.54 (m, 12H, 3J(H,H) ) 7.0 Hz; CH2),
3
1.03 (t, 18H, J(H,H) ) 7.0 Hz; CH3). 13C NMR (126.7 MHz,
The Xe@1 binding constant at 278 K was evaluated by
following the same method as that previously used at 293 K,2
that is, by competition with 2. We took advantage of slow-
exchange conditions between certain proton signals of the
occupied and free cages of 1. A solution of 8.4 mM of 1 and
10.2 mM of 2 was degassed in the NMR tube. Then, xenon gas
was added while the tube was maintained at about 193 K in a
cold methanol bath, in order to drastically reduce the solvent
vapor pressure without freezing xenon. After a delay sufficient
for temperature equilibration, 1H and 129Xe spectra were
recorded under quantitative conditions. It can be demonstrated
that the ratio K1/K2 is independent of xenon solubility and is
given by equation 1
CDCl3, 20 °C) δ ) 149.6 (6C), 147.4 (6C), 134.5 (6C), 132.1
(6C), 122.3 (6C), 117.0 (6C), 69.7 (6C), 69.4 (6C), 36.2 (6C),
31.8 (6C), 19.4 (6C), 14.1 (6C). HRMS (LSIMS), calcd for
C72H90O12Na [M°+] 1169.6330; found 1169.6333.
Characterization of the Compounds. The synthesized
cryptophanes were characterized by mass spectrometry (HRMS
LSIMS) performed by the Centre de Spectrome´trie de Masse,
University of Lyon, on a Thermo-Finnigan MAT 95XL
1
spectrometer, and by H and 13C NMR spectra recorded on
Varian Unity 500, Bruker Avance II 500 spectrometers. Chemi-
cal shifts, δ(1H, 13C), are given relative to Me4Si. Column
chromatographic separations were carried out on Merck silica
gel 60 (0.040-0.063 mm). Analytical thin layer chromatography
(TLC) was performed on Merck silica gel TLC plates F-254.
Melting points were measured on a Perkin-Elmer DSC7
calorimeter. Solvents were distilled prior to use: DMF and
CH2Cl2 from CaH2, pyridine from KOH, and THF and Et2O
from Na/benzophenone ketyl. Lithium diphenylphosphide was
freshly prepared from a known procedure.15
KXe@1 ⁄ KXe@2 ) [γ(1 + ꢀ) ⁄ R] - ꢀ
(1)
where R is defined as the concentration ratio of cryptophanes 1
and 2, ꢀ is defined as the ratio of cryptophane 1 containing a
xenon atom to empty cryptophane 1, and γ is defined as the
concentration ratio of cages 1 and 2 containing xenon atoms.
The R term was determined through integration of the aromatic
signals of the 1 + 2 solution mixture, and ꢀ was obtained from
integration of the aromatic protons at approximately 7.05 ppm
on the 1H spectrum of 1 (the doublet assigned to filled cages is
shifted downfield by approximately 0.03 ppm compared to that
In order to remove the remaining solvent traces before the
xenon binding experiments, the cryptophanes powders were
heated at 70 °C under vacuum during 1-2 h.
Xenon Polarization. Prior to measurement of 129Xe spectra,
xenon was polarized by optical pumping, by using the spin-

Cryptophane-Xenon Complexes in Organic Solvents
J. Phys. Chem. A, Vol. 112, No. 45, 2008 11367
assigned to empty cages, data not shown). The γ term was
determined from the thermally polarized xenon spectra of the
mixture. The binding constant value of 3900 M-1 for Xe@2
was taken from ref 19.
The in-out kinetics of xenon slightly influences its chemical
shift. The difference between the observed value and that
estimated in the case of infinitely slow exchange has been
assessed through a two-site model with unequal population.27
In a very unfavorable situation, reported in Figure S2 in the
Supporting Information for cryptophane 7, the chemical-shift
discrepancy of encapsulated and free xenon does not exceed
0.5 ppm. For 2 at the highest experimental temperature, the error
of the uncorrected chemical-shift value is on the order of 0.01
ppm. This is much lower than the uncertainties due to sample
temperature calibration. Therefore, experimental chemical shifts
were used without further correction. Moreover, the chemical
shift of xenon in solvent was negligibly modified by the presence
of a cryptophane in the solution. The linear dependence of this
chemical shift with temperature, -315 ( 3 ppb/K, already
reported by Bartik et al.,19 was used to calibrate this signal at
a chosen temperature for all cryptophanes. For some cryp-
tophanes, for example, 12 and 7, the calibration was performed
at the lowest temperature.
Free xenon chemical shifts in CDCl3 and CD2Cl2 at 293 K
were taken by reference equal to 216 and 193 ppm, respectively,
from their value at 296.5 K28 and from the variation of the
chemical shift of 1,1,2,2-tetrachloroethane-d2 with temperature.
Release Rate of Caged Xenon. Slow exchange conditions
on the xenon chemical-shift time scale were observed for all
cryptophanes under the experimental conditions of host con-
centration (2-11 mM), xenon pressure (3 × 103-1.6 × 105
Pa), and temperature (240-355 K). The apparent xenon release
rates were estimated by multiplying by π the full width at half
height of the signals assigned to bound xenon. The portion of
the cages that did not contain a xenon atom is not negligible
under certain experimental conditions. Therefore, the use of a
quasi exclusive second-order mechanism to explain the line
width of the bound-xenon signal19 did not necessarily apply.
However, the observed variation of the xenon line widths with
xenon pressure revealed a second-order mechanism (degenerate
state) for the xenon-cryptophane association that could be
described by the following equilibrium:
The Xe@7 binding constant was evaluated from the measure-
ment of the proton chemical shifts of 7 upon varying the amount
of bound xenon. In an NMR tube containing a reservoir, 190
µL of a 6.2 mM solution of 7 in 1,1,2,2-tetrachloroethane-d2
was placed. The solution was thoroughly degassed, and an initial
proton spectrum was acquired. Then, a series of known amounts
of xenon were added into the NMR tube by measuring the
decrease of pressure in a known volume through a membrane
gauge, while the solution is maintained frozen by immersion
in a liquid nitrogen bath. After xenon sublimation, the tube was
laid horizontally for 5 min with the solution in the reservoir in
order to increase the gas-liquid interface, at a temperature close
to that of the NMR study. Then, the NMR tube was inserted
into the spectrometer. The gas-liquid equilibrium of xenon was
reached after approximately 10 min, because the proton chemical
shifts did not evolve anymore, as observed through a series of
proton spectra. For each experimental point, the proportion of
xenon in the gas phase, free in solution and encapsulated in 7
depended on the xenon molar solubility and the Xe@7 binding
constant. Only the solubility value at 298 K was known.20
However, the variation of the molar fraction of dissolved xenon
with temperature at atmospheric pressure, from which molar
solubility could be deduced from the knowledge of the molar
volume of the solvent, is known for a wide set of solvents.21–25
For a given gas, this variation was fitted according to a currently
used parametric law25 (data not shown), and the ratio of molar
fractions at 278 and 298 K was extracted. Except for water,
this ratio was between 1.15 and 1.39. A value of 1.27 ( 0.13
was then used. As 1,1,2,2-tetrachloroethane density increases
by about 1.3% between 303.15 and 293.15 K,26 it has been
estimated to increase by 2.6% between 298 and 278 K. Finally,
solubility at 278K has been estimated as 1.29 ( 0.13 µM Pa-1
.
The binding constant was then evaluated by fitting the chemical
shift of an aromatic proton, which reflects the proportion of
cages filled by a xenon atom relative to the amount of added
xenon (see figure S2 in the Supporting Information). The
accuracy of the measurement depends strongly on the accuracy
of the xenon quantity added into the tube, especially for very
low pressures that do not imply a saturation of all cages. The
presence of the reservoir provided the possibility to reach an
equilibrium pressure of only 36 Pa after the addition of 0.14
micromoles of xenon. An error for the chemical-shift determi-
nation of 1.4 ppb has been estimated from the calculation of
reduced ꢁ2 from the model. By using relative errors from 2 to
10% on each experimental parameter, except 0.1% for temper-
ature and 1.4 ppb for chemical shifts, the median and standard
deviation of K7 were deduced through a Monte Carlo simulation
repeated 10 000 times.
Xe @ Ci + Xe * a Xe * @ Ci + Xe
(2)
A specific experimental protocol, in which samples contained
a mixture of cryptophanes, was therefore chosen, using either
hyperpolarized or thermally polarized xenon, in order to ensure
a kinetic comparison under the same conditions of temperature
and xenon pressure. The use of a solution containing several
cryptophanes is not expected to modify the results as it was
shown previously that the direct xenon exchange rate between
cryptophane cages was negligible compared to the release rate
into the solvent.10
T1 Measurements. The longitudinal relaxation times of xenon
encapsulated in cryptophanes were measured by one of the two
following methods. The first method consisted of monitoring
the evolution of the xenon peak area of the xenon signals, either
free in solvent or in the cage, obtained after very short excitation
pulses. Because fast in-out exchange rates on the xenon T1
time scale were present for all cryptophanes in our experimental
conditions, an average T1 was then deduced. As the longitudinal
relaxation time of xenon free in 1,1,2,2-tetrachloethane-d2 is
very long, on the order of hundreds of seconds,29 compared to
that in cryptophanes (see below), its exact value has a negligible
influence on the calculated T1 value of bound xenon, which is
therefore simply deduced from the ratio of the signal area of
xenon in both environments.
The Xe@5 binding constant was evaluated according to the
same method, by using the most downfield-shifted aromatic
proton signal. The error in chemical-shift determination has been
estimated as 1.9 ppb.
Chemical Shifts. The 129Xe chemical shifts were calibrated
through the resonance signal of xenon in tetrachloroethane-d2.
Segebarth et al.20 precisely measured the xenon chemical shifts
in this solvent as 223.60 ppm at 298 K, relative to the resonance
of xenon gas extrapolated to zero pressure. They also observed
that the chemical shift of dissolved xenon does not significantly
depend on pressure.
In the second method, the classical inversion-recovery
sequence was used with thermally polarized xenon. Both the

11368 J. Phys. Chem. A, Vol. 112, No. 45, 2008
Huber et al.
TABLE 2: Binding Constants Ki of Cryptophanes 1-7 with Xenon at 278 K in 1,1,2,2-Tetrachloroethane-d2 and Calculated
Internal Cavity Volume
cryptophane
1
2
3
4
5
6
7
Ki (M-1
volume (Å3)
)
28 000 ( 5000a
3900 ( 500b
2800b
102b
800b
117b
9.5 ( 1.0c
5-10a
1400 ( 150a,c
n. m.
81d
95b
110b
121b
a This work. b Reference 4. c See Figure S2 in the Supporting Information. d Reference 2; n. m., not measured.
noncomplete return to equilibrium and errors on the inversion
and lecture pulse angles were taken into account in the
calculation of T1.
affinity decreases as the cavity size increases (2, 3, 4, 6 series)4
is thus extended by one element: cryptophane 1. With host 1,
the apparent linear relationship between the binding constant
value and the cavity volume is broken (Table 2). For this
cryptophane, the ratio of the volume of xenon and the volume
of the cavity, r ) 0.52, is very close to the optimal value of
0.55 given by Rebek34 and significantly increases the affinity
for the noble gas. It is noteworthy that the water-soluble
congener of 1 would exhibit an even higher binding constant
in this solvent because of the hydrophobic character of the xenon
atom.
Clearly it can be argued that in contrast to 2-6, cryptophane
1 has no methoxy groups on the aromatic rings, thereby
questioning the validity of the comparison, because the elec-
tronic environment of the encapsulated xenon atom is signifi-
cantly modified by the removal of the six methoxy groups.
Therefore, the binding constant of xenon with 7 was also
measured. Cryptophane 7 differs from 1 by the length of the
linkers between the CTB bowls and differs from 2 by the
absence of methoxy substituents on the aromatic rings in
positions R1 and R2 (Table 1). The affinity of xenon for 7 is 20
times lower than that for 1 and 2.8 times lower than that for 2
(Table 2). It is unlikely that the change of cavity size resulting
from the presence of methoxy groups instead of hydrogen atoms
is solely responsible for the decrease of xenon affinity from 2
to 7 (otherwise, the affinity would be higher for 7 than for 3).
Rather, partial access of the solvent through the portals may
act as a competitor to xenon in the case of 7.
The estimated cavity volume of 5 (V ) 110 Å3) is smaller
than that of 4 (V ) 117 Å3), whereas the binding constant is
far lower (Table 2). This has been previously explained by a
higher flexibility of 5, which has a butanedioxy linker. Thus,
some conformations unfavorable for xenon binding may exist
for 5.
Xenon Chemical Shift. Our observations confirm previous
results,19 which showed that neither cryptophane concentration
nor xenon pressure significantly modify the chemical shift of
xenon free in solution, as long as the slow exchange regime
stands. The temperature dependence of the xenon chemical-
shift values for xenon encapsulated in cryptophanes 1-12 are
reported in Figure 1, at various temperatures. Let us first
consider the xenon chemical-shift values at 293 K (dashed
vertical line). The value for Xe@12 will be extrapolated, because
it has been measured at slightly lower temperature.
Results and Discussion
With the goal of optimizing the cryptophane cage moiety for
biological sensing applications, the influence of the structural
and electronic properties of cryptophane on key parameters such
as the xenon binding constant, chemical shift, in-out kinetics,
and longitudinal relaxation is examined. This analysis is based
on a set of 12 cryptophanes differing by their stereoisomerism,
their length of the linkers, and the number and type of
substituents on the aromatic rings. A comparative study within
this set renders possible the formulation of hypotheses on the
molecular origin of the differences observed for some xenon-
cryptophane binding properties. Obviously, none of these
cryptophanes is significantly soluble in water or blood and thus
will not be directly amenable for xenon biosensing. The presence
of a hydrophilic tether however will help solubilize the cage
molecule in biological fluids. Nevertheless, this study aims at a
better understanding of xenon-host interactions.
Binding Constants. The binding constants, Ki, of xenon with
cryptophanes 1-7 in 1,1,2,2-tetrachloroethane-d2 at 278 K are
reported in Table 2. As already noticed, these average values
compete favorably with those encountered for xenon in other
cage molecules: for instance, values of 20 M-1 for R-cyclo-
dextrin in water at room temperature30 and of 200 M-1 for
cucurbit[6]uril in acidic conditions31 are reported. The exception
is the very recent measurement of a binding constant of 3000
M-1 for xenon in a cucurbit[6]uril derivative.32 Anyway,
cryptophanes seem really to be among the best candidates to
constitute the core of xenon sensors.
Cryptophanes 2-6 differ only by the length of the alkyl
chains linking the two CTB units. The decrease of the xenon
binding constant value in the series 2, 3, 4, and 6 may be
explained by a decrease of stabilizing cryptophane-xenon
interactions through London forces between the noble gas and
the accessible atoms of the host. It has indeed been shown that
xenon is stabilized by close proximity to ethanedioxy linkers
in preference to longer alkyldioxy chains.33 Also, cryptophane
sampling of conformational space when binding xenon, or not,
may vary in the series and affect the entropic part of Ki. The
role of the solvent, which for large cavities can act as a
competitor, could explain the low binding constant value for
Xe@5, because this host has one butanedioxy and two
ethanedioxy linkers. In this case, more flexible portals may allow
increased interaction between solvent molecules and the en-
capsulated noble gas.
The binding constant of xenon with 1, K1, had previously
been estimated as 10 000 M-1 at 293 K, on the basis of
competition experiments with 2.2 A K1 value of 28 000 ( 5000
M-1 at 278 K is now determined by using the same method.
This is by far the highest binding constant between xenon and
a cage molecule ever observed in an organic solvent. The already
mentioned trend which indicates that the xenon-cryptophane
The arrangement of the O-CH2-CH2-O linkers has a strong
influence on the chemical shift of encapsulated xenon. This is
exemplified by a larger downfield xenon chemical shift for 12
than for 11. It may be correlated with a different shape of inner
space available for xenon, less spherical in 12 than in 11, as
observed by X-ray crystallography when a dichloromethane
molecule is encapsulated.6
The decrease of xenon chemical shift observed with the
increase of portal size in the 2, 3, 4, 6 series had been
rationalized by ab initio calculations.35 It was found that xenon
is more deshielded upon closer approach to a host atom;

Cryptophane-Xenon Complexes in Organic Solvents
J. Phys. Chem. A, Vol. 112, No. 45, 2008 11369
Figure 1. Chemical shift of xenon encapsulated in cryptophanes in
(CDCl2)2 as a function of temperature for 1 (blue squares), 2 (red
circles), 3 (orange triangles), 4 (orange diamonds), 5 (orange circles),
6 (red squares), 7 (blue circles), 8 (green squares), 9 (violet triangles),
10 (pink circles), 11 (black triangles), and 12 (black squares). Symbols:
experimental data. Lines: linear fit of experimental points. The dashed
vertical line corresponds to T ) 293 K.
Figure 3. Temperature dependence of the chemical shift of xenon
encapsulated in cryptophanes 1-12 in 1,1,2,2-tetrachloroethane-d2 in
ppb K-1
.
characteristic of bound xenon is downfield shifted by ap-
proximately 13 ppm in 1,2-dichloroethane and by 5 ppm in 1,2-
dibromoethane compared to that in 1,1,2,2-tetrachloroethane,
whereas it is only slightly upfield shifted by 2 ppm in
chloroform. Conversely, the chemical shift of xenon in 1 is quasi
independent of the nature of the solvent. The explanation we
propose is that, in the case of 2 in (CD2Cl)2 and to a lesser
extent in (CD2Br)2, a solvent molecule partially enters the cavity,
leading to a downfield shift of the bound xenon signal. This
does not occur in 1,1,2,2-tetrachloroethane where steric hin-
drance of each end of the molecule is increased by the presence
of a second halogen atom, nor in chloroform which contains
three chlorine atoms. For 1, the halogenated solvent molecules
are too big to enterseven partiallysinto the cavity, therefore
giving rise to a quasi invariant chemical shift of bound xenon
whatever the solvent. Moreover, xenon chemical shift in 1 is
modified by less than 0.03 ppm when alkoxy linkers are
deuterated (data not shown). This confirms both the conforma-
tion of linkers in 1, that point outward the cavity, and confirms
the low flexibility of this cryptophane.
Figure 2. Hyperpolarized 129Xe spectrum of a solution of 2 in 1,1,2,2-
tetrachloroethane-d2 (a), chloroform-d (b), 1,2-dichloroethane-d4 (c),
and 1,2-dibromoethane (d), obtained in one scan at 293 K. Chemical
shifts are referenced (a-c) to the signal of xenon free in solution19,20,28
or (d) to a signal at 0 ppm assigned to gas bubbles (data not shown).
The 129Xe spectra for 2 in dichloromethane and carbon
tetrachloride reveal only one signal (data not shown). This
clearly confirms that CD2Cl2 has a good affinity for the cavity
of 2,38 inducing a far lower apparent binding constant for xenon.
The unique signal observed in carbon tetrachloride may come
from a fast exchange between bound and free xenon, because
a significantly wider signal is observed at 258 K compared to
293 K (data not shown). Because all other tested solvents contain
at least one proton, they may reduce the xenon exchange rate
by the partial entry of that proton. This does not occur for carbon
tetrachloride.
Cryptophanes 2, 7, 8, 9, 10, and 11 differ only by the nature
of the substituents R1 and R2 on the aromatic rings. The chemical
shifts of xenon in 2, 8, 9, and 10 are very similar at room
temperature. In this series, Xe@7 is the most upfield shifted.
Because calculations indicate that xenon is deshielded when it
approaches closely a host atom,35 it seems that the deshielding
effect is more important when an alkoxy group is in position
R1 or R2, as in 2, 8, 9, and 10, than when it is replaced by a
proton, as in 7. Cryptophane 11 has an intermediate situation
in which only half the aromatic rings carry an alkoxy group,
inducing an average value of the xenon chemical shift between
those of 7 and 2. However, the exact formula of the alkoxy
group seems to only slightly influence xenon shielding, because
the xenon frequency bandwidth is very narrow (between 66.0
and 67.2 ppm) for 2, 8, 9, and 10. These results can be compared
to those obtained by Pines et al. on derivatives of cryptophane
2 bearing one bulky biotinylated substituent.9,39,40 In these cases,
therefore, xenon in the largest cage (6) is the most upfield
shifted. This does not apply to cryptophane 1. However, it was
recently reported that relativistic effects, that were neglected
by Sears and Jameson,35 may significantly influence xenon
chemical shift in other cage molecules.36,37
Obviously, the nature of the substituents on the aromatic rings
has its own influence on the chemical shift of encapsulated
xenon. For instance, the difference in shift of approximately
20 ppm between Xe@2 and Xe@7 is easily explained by the
absence of the methoxy groups in the latter complex, which
leads to a decrease of the electronic density in the aromatic ring
of 7 and therefore a larger shielding effect.
Comparison of two cryptophanes having the same substituents
but different sizes, 1 and 7, shows an unexpected shielding effect
for the former, which could be explained by a solvent effect.
In the present study, a detailed comparison of the 129Xe NMR
parameters of the complexes is made in 1,1,2,2-tetrachloroet-
hane-d2 (C2D2Cl4). However, it should be kept in mind that the
solvent may also have its own influence on parameters such as
the chemical shift of encapsulated xenon. As an example, 129Xe
spectra of 1 and 2 were also recorded at 293 K in chloroform-
d, 1,2-dichloroethane-d4, and carbon tetrachloride (see Figures
2 and S3 in Supporting Information). The 129Xe spectra of 2 in
dichloromethane-d2 and 1,2-dibromoethane were also recorded.
Two signals are observed in 1,2-dichloroethane, 1,1,2,2-tetra-
chloroethane, chloroform, and 1,2-dibromoethane, indicating a
slow exchange on the chemical-shift time scale. For 2, the signal

11370 J. Phys. Chem. A, Vol. 112, No. 45, 2008
Huber et al.
and especially when the substituent is bound to avidin, the signal
of caged xenon is shifted downfield by several ppm. This seems
to come from a cage distortion.35 This is not observed when
the cryptophane moiety bears smaller substituents such as those
of 8, 9, and 10, perhaps because O-benzyl and butoxy substit-
uents are not bulky enough to significantly modify the cage
conformation.
Release Rate of Bound Xenon. The in-out exchange rate
of xenon is an important parameter for biosensing applications.
It has to be slow on the xenon chemical-shift time scale in order
to enable observation of a specific signal for caged xenon and
furthermore to give rise to a sharp signal. But the release rate
must also be high enough to allow the replenishment of
cryptophane cages with polarized xenon between acqui-
sitions.39,41,42 Previous work evidenced the importance of the
nature of the tether linking the cryptophane moiety to the
recognition antenna on the caged xenon line width.40 One of
the aims of the present study is to examine the effect of various
cryptophane cores and their substituents on the xenon in-out
exchange rates.
In the present study, the xenon release rate is estimated from
the line width of the encapsulated noble gas signal (see Materials
and Methods). Although this measurement is not precise, most
of the spurious effects are canceled by recording spectra at
different temperatures. For instance, the magnetic-field inho-
mogeneity contribution to the line width has been estimated
from experiments at the lowest temperature examined, where
exchange is the slowest. Furthermore, the xenon release rates
are used only in a comparative way by using mixtures of
cryptophanes. This method ensures that both temperature and
xenon concentration, parameters that greatly affect the kinetics,
are the same along a series.
The influence of the cage stereoisomerism on xenon in-out
kinetics has been assessed through the study of a mixture of
cryptophanes 11 and 12. 129Xe NMR spectra under conditions
of a large excess of thermally polarized xenon indicate that the
release rate is much faster for 12 than for 11 (data not shown).
Portal sizes were found to be similar in the crystallographic
structures of 11 and 12.6 However, when the cages contain a
xenon atom, the aromatic protons in R1 positions, according to
Scheme 1, are deshielded by 0.6 ppm in 12 compared to those
of 11.43 This suggests that both cryptophanes explore signifi-
cantly different conformational spaces, inducing different global
xenon release rates.
It appears that the presence of the methoxy groups sterically
hinders the release of xenon. For example, the rate is about
one order of magnitude higher for 7 than for 2, as observed in
the 129Xe spectrum of a mixture of both cryptophanes (Figure
S4 in the Supporting Information). This observation is supported
by a slightly higher release rate of xenon for 11 than for 2, at
least for temperatures above 265 K (Figure S5 in the Supporting
Information). At these low temperatures, the small kinetic
difference can be interpreted as one methoxy group per portal
being sufficient to hinder the release of xenon. At higher
temperature, the cryptophane can attain conformations where a
single methoxy group is less able to hinder the passage of xenon.
The presence of a second methoxy group close to each portal,
as in 2, may further slow down the release of the noble gas.
Evolution of Xenon Chemical Shift with Temperature. The
chemical shift of xenon encapsulated by each of the cryp-
tophanes 1-12 is downfield shifted with temperature. The slopes
of the caged xenon chemical shift as a function of temperature
are reported in Figure 3 and can be sorted into four groups:
- Xe@10: 84 ppb K-1
- Xe@1, Xe@8, Xe@9: approximately 150 ppb K-1
- Xe@2, Xe@3, Xe@4, Xe@6, Xe@11, Xe@12: approximately
225 ppb K-1
- Xe@5, Xe@7: approximately 300 ppb K-1
The temperature-dependent behavior of the xenon chemical
shift can be interpreted as xenon exploring regions not easily
accessible at lower temperature, as a consequence of the
deformation of the xenon electron cloud and increased confor-
mational flexibility of the cage as temperature increases.35 The
importance of these effects varies among the different cryp-
tophanes. Cryptophane 5 bears the same substituents on the
aromatic rings as cryptophanes 2, 3, 4, and 6 but exhibits a
significantly higher chemical-shift variation. Because the cal-
culated cavity volume of 5 is slightly smaller than those of 4
and 6 (Table 2), the xenon mobility in the cavity of 5 may be
close to that in the latter two and therefore cannot explain the
observed effect. The temperature dependence of the xenon
chemical shift in 5 may be due instead to the relatively low
energy barriers that the butanedioxy linker has to cross in order
to adopt twisted conformations that tend to minimize the size
of the cavity or extended conformations that maximize it.
It appears that, for cryptophanes with ethanedioxy linkers,
the slope decreases if bulky moieties are attached to the aromatic
rings, as can be observed by comparing cryptophane 7 (no
aromatic substituents), 11 or 2 (methoxy groups), and 8 or 9
(benzyloxy groups). This may be interpreted as an increase in
the energy barrier to cage distortion when bulkier groups are
present, thus making high-energy conformations less accessible.
Moreover, Figure 3 shows that the number of aromatic ring
substituents does not significantly affect the slope as shown by
comparison of cryptophanes 8 and 9 or cryptophanes 2 and 11.
Cryptophane 10 exhibits by far the lowest temperature-
dependent variation of the series, which can be interpreted as
due to the presence of flexible butyl groups that hinder the access
of solvent molecules to bound xenon. Moreover, van der Waals
interaction between butoxy groups may reduce the cage’s con-
formational freedom.
The difference in the xenon temperature coefficient slopes
observed for 11 and 12, which differ by the connectivity of
their linkers (syn- and anticryptophanes), cannot easily be
explained without the help of the calculation of the potential
energy surface, particularly with regard to the relative position
of the two cyclotriveratrylene units.
The 129Xe spectra of a mixture of cryptophanes 2, 3, 4, 6
enable a comparison of the xenon release rates (Figure 4). At
higher pressure, similar line widths are observed for Xe@2,
Xe@3, Xe@4, and Xe@6. This may be explained in part by
lower magnetic-field homogeneity due to a macroscopic effect
of susceptibility variation caused by xenon bubbles.
At lower noble-gas pressure, the line widths reveal the
intrinsic cryptophane-xenon dissociation rates. The line widths,
decreasing from Xe@2 to Xe@6, indicate that xenon exchange
is faster for the smaller cavities than for the larger ones. This
behavior, opposite to that observed for the series of water-soluble
cryptophane congeners, where the rates increased along the
series, appears surprising.41 By considering that the van der
The experimental slope for Xe@1 is significantly lower than
that of Xe@7. Because of a much smaller volume available for
encapsulated xenon in 1 than in 7, the noble gas may interact
with all atoms that constitute the cavity of 1 simultaneously, a
situation that does not occur for 7. Also, 1 is more rigid than 7,
particularly when filled with xenon. Finally, because of the large
portals of 7, the influence of the solvent could accentuate the
Xe@7 chemical-shift variation.

Cryptophane-Xenon Complexes in Organic Solvents
J. Phys. Chem. A, Vol. 112, No. 45, 2008 11371
TABLE 3: Longitudinal Relaxation Times of Encapsulated
Xenon at 293 K
Cryptophane
1
2
7
8
11
15.1 ( 2.4a
3.9 ( 0.5a 9.3 ( 1.1a 5.7 ( 0.3a
caged
T1
(s)
12.1 ( 0.7b 19.1 ( 4.7b
a Method using peak area monitoring. b Inversion-recovery
method.
The T1 decrease in the 2, 11, 7 series, along which the number
of methoxy substituents decreases, can be interpreted as due to
a more efficient relaxation of the noble gas with the aromatic
ring protons than with the more distant and mobile methoxy
groups. For 1 and 7, which differ by the length of the three
linkers, it appears that the T1 value of xenon bound to the former
is significantly higher than in the latter, partly because each
linker bears only two protons instead of four. Moreover, the
two protons of the O-CH2-O linker point away from the cavity
and are therefore more distant from the encapsulated xenon
atom. The T1 value of xenon in 1 where alkoxy linkers are
deuterated has been measured to be 12.8 ( 1.5 s. This value is
only slightly higher than that of the fully protonated form of 1.
As for the chemical shifts, this is in agreement with a
conformation of O-CH2-O linkers where hydrogen atoms point
outward the cavity. These experimental results may also be
Figure 4. High-field region of the laser-polarized 129Xe NMR spectrum
of a mixture of 2, 3, 4, 6 at concentrations of 1.8, 3.0, 6.4, and 10.7
mM, respectively, under a xenon pressure of (a) 1.6 ×105 Pa and (b)
3.104 Pa, at T ) 295 K (one scan). A 10 Hz Lorentzian apodization
has been applied on the FID before Fourier transformation.
Waals diameter of chlorine is comparable to the portal size in
6, a possible explanation is that solvent molecules partially
penetrate the cavity, thereby slowing down the xenon in-out
exchange rate. As the size of the portals progressively decreases
in the order 6, 4, 3, 2, the blocking interaction of the solvent
with the portal is limited.44 At higher pressure however, the
replacement of a xenon atom in the cavity by another xenon
atom appears more efficient for the largest cryptophane, 6.
Another explanation could be a folding of the structure of 6 at
low xenon pressure, giving rise to a slow xenon exchange
through a concerted process involving deformation of the cage
and encapsulation of xenon.
Finally, the recent synthesis of 1 provides a cage molecule
with much lower xenon release rate: 2.4 Hz at 293 K,2 about 2
orders of magnitude lower than for 2. The in-out exchange of
xenon is drastically reduced in 1 because the average size of
the portals is smaller than the diameter of xenon and because
the host appears rather rigid. That exchange is still possible,
maybe because of the distortion of the xenon electronic cloud.
Indeed, recent work supports this assumption, because a
chloromethane molecule possessing a similar van der Waals
volume as xenon but a lower polarizability does not enter the
cavity of 1 under the same experimental conditions. This result
will be reported in due course.
compared with those obtain with a water-soluble cryptophane
caged
derivative of 2, where T1
significantly depends on the
deuteration level of the alkoxy linkers and acetate substituents
and where the encapsulated xenon chemical shift differs by ca.
1 ppm between both isotopomers.41
The two trends observed for 1, 2, and 7 suggest that a
cryptophane molecule having short linkers such as those in 1
and bearing methoxy groups such as in 2 would result in an
even longer T1 and thus lead to better conservation of the out-
of-equilibrium xenon polarization. Obviously, this analysis only
considers structural aspects of the Xe@cryptophane complexes.
Undoubtedly, the dynamics of the complexes also have a large
influence through the rotational correlation time of the encap-
sulated noble gas. The difference of longitudinal relaxation time
for xenon encapsulated in 7 and in 8 evidences that the aromatic
protons contribute greatly to the relaxation of the encapsulated
xenon. Their replacement by benzyloxy groups on three of the
six aromatic rings is sufficient to lengthen the relaxation time
by a factor 2.4.
Xenon Longitudinal Relaxation. The longitudinal relaxation
time values, T1, measured on signals characteristic of xenon
free in solution and encapsulated in cryptophanes are always
similar, within the experimental error. This comes from a fast
xenon exchange (faster than 10 Hz) on the relaxation time scale.
The averaged xenon T1 is expressed as:
Conclusion
The analysis of a family of 12 cryptophanes shows that their
interaction with xenon involves a high variability of NMR
parameters. In order to rationally design biosensors for specific
applications based on 129Xe NMR, several trends related to their
physical behavior can be drawn from this work. As expected,
the affinity of the host for xenon strongly depends on the quality
of the match between the cavity size and the diameter of the
noble gas atom. The smallest cryptophane exhibits the highest
binding constant with xenon, at the price however of slower
in-out xenon exchange less adequate for the 129Xe-NMR-based
biosensing approach. The exchange rate is an important
parameter for this application, because it will govern the
capability to reload the cage with laser-polarized xenon after
each radio frequency excitation.45 The cavities of cryptophanes
delimited by portals of a size similar to that of solvent molecules
enable a wide range of in-out exchange rates to be spanned.
We have emphasized the fact that size is not the sole parameter
that acts on the exchange rate; flexible hydrophobic substituents
1
T1
1
1
T1free
)
xcaged
+
(1 - xcaged
)
(3)
T1caged
where xcaged is the fraction of caged xenon.
Because the relaxation time of the xenon free in solution is
caged
very long, on the order of several minutes, the T1
value of
encapsulated xenon may be extracted directly from the experi-
mental T1 and the relative xenon populations in both environ-
ments, measured by the area of the corresponding signals. The
T1caged values of xenon in five cryptophanes, 1, 2, 7, 8, and 11,
at 293 K, are given in Table 3. The choice of the method only
slightly influences the extracted T1 value. The value obtained
for 2 is slightly higher than the value of 16.4 s estimated by
Luhmer et al.29 but remains within the experimental error.

11372 J. Phys. Chem. A, Vol. 112, No. 45, 2008
Huber et al.
(7) 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.
(8) Hilty, C.; Lowery, T. J.; Wemmer, D. E.; Pines, A. Angew. Chem.
Int. Ed. 2006, 45, 70.
(9) 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.
(10) Brotin, T.; Devic, T.; Lesage, A.; Emsley, L.; Collet, A. Chem.
Eur. J. 2001, 7, 1561.
(11) Brotin, T.; Roy, V.; Dutasta, J.-P. J. Org. Chem. 2005, 70, 6187.
(12) Cavagnat, D.; Buffeteau, T.; Brotin, T. J. Org. Chem. 2008, 73,
66.
(13) Canceill, J.; Collet, A. J. Chem. Soc., Chem. Commun. 1983, 1145.
(14) Canceill, J.; Collet, A.; Gottarelli, G. J. Am. Chem. Soc. 1984, 106,
5997.
(15) Vedejs, E.; Fuchs, P. L. J. Am. Chem. Soc. 1973, 95, 822.
(16) Herman, R. M. Phys. ReV. 1965, 137A, 1062.
(17) Walker, T. G.; Happer, W. ReV. Mod. Phys. 1997, 69, 629.
(18) Desvaux, H.; Gautier, T.; Le Goff, G.; Pe´tro, M.; Berthault, P. Eur.
Phys. J. D 2000, 12, 289.
(19) Bartik, K.; Luhmer, M.; Dutasta, J.-P.; Collet, A.; Reisse, J. J. Am.
Chem. Soc. 1998, 120, 784.
(20) Segebarth, N.; A¨ıtjeddig, L.; Locci, E.; Bartik, K.; Luhmer, M. J.
Phys. Chem. A 2006, 110, 10770.
(21) Pollack, G. L.; Himm, J. F. J. Chem. Phys. 1982, 77, 3221.
(22) Pollack, G. L.; Himm, J. F.; Enyeart, J. J. Chem. Phys. 1984, 81,
3239.
(23) Pollack, G. L.; Kennan, R. P.; Himm, J. F. J. Chem. Phys. 1989,
90, 6569.
(24) Kennan, R. P.; Pollack, G. L. J. Chem. Phys. 1988, 89, 517.
(25) Clever, H. L. IUPAC solubility data series; Pergamon Press: Oxford,
1979.
can seriously slow down the exchange, as in the case of 10.
Moreover, the exchange rate is significantly increased in the
syn-isomer cryptophane compared to the anti-isomer, as exem-
plified by 11 and 12.
The longitudinal self-relaxation time of encapsulated xenon
is also a key parameter, because it influences the sensitivity of
the NMR experiments. It is observed that cryptophanes bearing
the shortest alkoxy linkers and the most substituted aromatic
rings induce the slowest xenon relaxation because of the smaller
number of protons in close proximity to the encapsulated noble
gas. Indeed, it was shown previously that for some water soluble
cryptophanes, deuteration of the aromatic ring proton sites also
acts significantly on xenon T1.41 Obviously, our study, performed
in halogenated solvents, cannot fully prefigure the xenon
properties experienced in water, but it gives trends to aid in the
conception of new cage molecules adapted to the environment
of in vivo experiments around 310 K.
The effect of the solvent has been suggested not only for the
in-out exchange rate but also for the resonance frequency of
caged xenon. In aqueous solvents, it is nevertheless expected
that the water molecules will have a smaller effect on the
chemical shifts. However, when modifying the size of the cage,
its stereochemistry, or to a lesser extent the substituents on the
aromatic rings, will give rise to specific spectral signatures of
the new hosts, enabling multiplexing applications.
From our study, it is interesting to note that cryptophanes 1
and 6 (i.e., the smallest and the biggest host molecules) show
some resemblance when dealing with the interaction with xenon.
In 1,1,2,2-tetrachloroethane, the chemical shifts of caged xenon
are the most upfield shifted among those of the 12 cryptophanes.
Also, the line widths of these signals are the sharpest at low
xenon concentration, indicating a very slow exchange. This
could indicate that thanks to the flexibility of the alkyldioxy
chains of 6, this host exists under a folded form, with a small
cavity offered to xenon. Such an adaptation effect may occur
as soon as the linker length has three carbon atoms, as evidenced
by the line widths observed on the 129Xe spectrum of 2, 3, 4, 6
at low xenon concentration.
(26) Al-Hayan, M. N. M. J. Chem. Thermodyn. 2006, 38, 427.
(27) Rogers, M.; Woodbrey, J. J. Phys. Chem. 1962, 66, 540.
(28) Miller, K. W.; Reo, N. V.; Schoot Uiterkamp, J. M.; Stengle, D. P.;
Stengle, T. R.; Williamson, K. L. Proc. Natl. Acad. Sci. U.S.A. 1981, 78,
4946.
(29) Luhmer, M.; Goodson, B. M.; Song, Y.-Q.; Laws, D. D.; Kaiser,
L.; Cyrier, M. C.; Pines, A. J. Am. Chem. Soc. 1999, 121, 3502.
(30) Bartik, K.; Luhmer, M.; Heyes, S. J.; Ottinger, R.; Reisse, J. J.
Magn. Reson. B 1995, 109, 164.
(31) El Haouaj, M.; Luhmer, M.; Ho Ko, Y.; Kim, K.; Bartik, K.
J. Chem. Soc., Perkin Trans. 2 2001, 804.
(32) Kim, B. S.; Ko, Y. H.; Kim, Y.; Lee, H. J.; Selvapalam, N.; Lee,
H. C.; Kim, K. Chem. Commun. 2008, 2756.
(33) Desvaux, H.; Huber, J. G.; Brotin, T.; Dutasta, J.-P.; Berthault, P.
ChemPhysChem 2003, 4, 384.
(34) Mecozzi, S.; Rebek, J., Jr. Chem. Eur. J. 1998, 4, 1016.
(35) Sears, D. N.; Jameson, C. J. J. Chem. Phys. 2003, 119, 12231.
(36) Straka, M.; Lantto, P.; Vaara, J. J. Phys. Chem. A 2008, 112, 2658.
(37) Xe@5, the signal of which is strongly low-field shifted, cannot be
incorporated in the previous series. We are unable to explain such an effect.
(38) Canceill, J.; Lacombe, L.; Collet, A. J. Chem. Soc., Chem. Commun.
1987, 219.
(39) 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.
Acknowledgment. Financial support from the French Min-
istry of Research (ANR Program Physico-Chimie du Vivant
2006) is greatly acknowledged. H.F. is grateful for financial
support from a U.S. National Science Foundation International
Research Fellowship grant (no. 0502393).
(40) 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.
Supporting Information Available: This material is avail-
able free of charge via the Internet at http://pubs.acs.org.
(41) Huber, G.; Brotin, T.; Dubois, L.; Desvaux, H.; Dutasta, J.-P.;
Berthault, P. J. Am. Chem. Soc. 2006, 128, 6239.
(42) Schro¨der, L.; Lowery, T. J.; Hilty, C.; Wemmer, D. E.; Pines, A.
Science 2006, 314, 446.
(43) Canceill, J.; Lacombe, A.; Collet, A. C. R. Acad. Sci. Paris Ser. II
1984, 298, 39.
(44) For water-soluble cryptophanes, the water molecule, contrary to
1,1,2,2-tetrachloroethane-d2, is small enough to enter the cavity of all
members of the series.
(45) Schro¨der, L.; Chavez, L.; Meldrum, T.; Smith, M.; Lowery, T. J.;
Wemmer, D. E.; Pines, A. Angew. Chem. Int. Ed. 2008, 47, 4316.
References and Notes
(1) Canceill, J.; Lacombe, L.; Collet, A. J. Am. Chem. Soc. 1985, 107,
6993.
(2) Fogarty, H. A.; Berthault, P.; Brotin, T.; Huber, G.; Desvaux, H.;
Dutasta, J.-P. J. Am. Chem. Soc. 2007, 129, 10332.
(3) Gabard, J.; Collet, A. J. Chem. Soc., Chem. Commun. 1981, 1137.
(4) Brotin, T.; Dutasta, J.-P. Eur. J. Org. Chem. 2003, 973.
(5) Canceill, J.; Lacombe, L.; Collet, A. J. Am. Chem. Soc. 1986, 108,
4230.
(6) Canceill, J.; Cesario, M.; Collet, A.; Guilhem, J.; Riche, C.; Pascard,
C. J. Chem. Soc., Chem. Commun. 1986, 339.
JP807425T