Ditopic ion transport systems: Anion-π interactions and halogen bonds at work

Article abstract of DOI:10.1002/anie.201104966

Single-atom exchange series are introduced to extract the individual contributions of halogen bonds and anion-π interactions to the transport of anions across lipid bilayer membranes (see picture). Known cation binding sites are used for counterion activa

Full text of DOI:10.1002/anie.201104966

DOI: 10.1002/anie.201104966
Halogen Bonds
Ditopic Ion Transport Systems: Anion–p Interactions and Halogen
Bonds at Work**
Andreas Vargas Jentzsch, Daniel Emery, Jiri Mareda, Pierangelo Metrangolo, Giuseppe Resnati,
and Stefan Matile*
Ion transport systems that operate in lipid bilayer mem-
branes[1–3] are emerging as attractive tools to probe the
functional relevance of weak interactions that are otherwise
difficult to observe.[1] This approach builds on the notion that
transport and catalysis operate best with weaker interactions
than the ones that are required for detection in binding
studies. Earlier important work in theory and in bulk
membranes has confirmed that transport efficiency follows
“Goldilocks principle”, with the stronger binders not being
the best transporter.[4] Herein, we introduce single-atom
mutation series with a new ditopic ion transport system to
demonstrate the general functional relevance of anion–p
interactions and to achieve, for the first time, anion transport
with halogen bonds.
The rarity of functional halogen bonds[5–10] compared to
the frequency of hydrogen bonds is reminiscent of the
situation with anion–p interactions[11] compared to the
ubiquitous cation–p interactions. Increasing in strength with
halogen atom polarizability, halogen bonds are most efficient
with iodine substituents.[5,6] They strongly depend on the
environment, including contributions from solvation/desol-
vation, and are strongest in noncompetitive hydrophobic
solvents.[7] Halogen bonds have been studied extensively in
solid-state crystal engineering.[5] Whereas pioneering exam-
ples for rational drug design[8] and anion binding with halogen
bonds in solution exist,[9] their application to catalysis[10] and
particularly to transport is essentially unexplored.
Figure 1. Ditopic ion transport systems made to study anion–p
interactions and halogen bonds at work, with effective concentrations
EC50 and Hill coefficients n to summarize the transport activities
found. Red colors indicate electron-rich, blue colors electron-poor
regions. Red balls indicate anions (e.g., chloride or hydroxide), blue
balls TMA cations, and dotted red lines possible anion–p (1), halogen-
bonding (2, 4, 5), or hydrogen-bonding (3, 6) interactions between
anions and transporters. These interactions are purely speculative and
are shown with the only intention to illustrate the design (compare
models, Figure 3b, S21, S22).
Calix[4]arenes 1–6 were designed to dissect the individual
contributions of halogen bonds, hydrogen bonds, and anion–p
[*] A. Vargas Jentzsch, D. Emery, Dr. J. Mareda, Prof. S. Matile
Department of Organic Chemistry, University of Geneva
Geneva (Switzerland)
interactions to anion transport (Figure 1). Calix[4]arene
cones[2,12–14] were selected for this study because they
1) offer convenient construction sites at the lower rim,
2) bind ammoniun cations at the upper rim,[15] and 3) are
well explored as transport systems in lipid bilayer mem-
branes.[2] These features indicated their suitability for the
construction of ditopic transporters with a tetramethylammo-
nium (TMA) binding site near a modular anion-binding site,
where nature and number of possible anion–p (1), halogen-
bonding (2, 4, 5), or hydrogen-bonding (3, 6) interactions with
anions can be varied systematically and without global
structural changes. Ditopic transporters are interesting
because they can operate by anion/cation symport and thus
avoid any temporary transmembrane separation of anions
and cations during transport. Their straightforward synthesis
is described in the Supporting Information (Scheme S1–S2,
Figure S10–S20).[16]
E-mail: stefan.matile@unige.ch
Homepage: http://www.unige.ch/sciences/chiorg/matile/
Prof. P. Metrangolo, Prof. G. Resnati
NFMLab and CNSC-IIT@POLIMI: Department of Chemistry,
Materials and Chemical Engineering “Giulio Natta”, Politecnico di
Milano
via Mancinelli 7, I-20131 Milan (Italy)
[**] We thank D. Jeannerat, A. Pinto, and S. Grass for NMR spectroscopy
measurements, the Sciences Mass Spectrometry (SMS) platform
for mass spectrometry services, F. Sansone (Parma) for samples,
and the University of Geneva, the European Research Council (ERC
Advanced Investigator), the National Centre of Competence in
Research (NCCR) Chemical Biology, and the Swiss NSF for financial
support. We also thank the Swiss Center for Scientific Computing
(CSCS Manno) for computing time (Cray-XT5).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201104966.
Angew. Chem. Int. Ed. 2011, 50, 11675 –11678
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
11675

Communications
The transport activity of calix[4]arenes 1–6 was evaluated
with the 8-hydroxy-1,3,6-pyrenetrisulfonate (HPTS)
on interfering conformational constraints (not shown).[2]
Calix[4]arenes 1–6 were almost inactive as well in the
presence of all cations tested except for the TMA cation
(Figure S1, S5). This superb cation selectivity confirmed that
the known TMA complexes at the upper rim are essential for
function (Figure 1).[2,15] Similar counterion activation has
been observed with cesium-activated calix[4]pyrrole anion
transporters[18] and with many polyion transporters.[19]
Hill analysis of dose response curves of TMA-activated
calix[4]arene 1 gave an effective concentration EC50 = 25 mm
and a Hill coefficient n = 1.1 (Figure 1 and 2a&). The EC50
describes the effective concentration needed to reach 50%
activity and can be similar or proportional to the dissociation
constant KD of the transport-active complex.[17] The Hill
coefficient n can describe cooperativity and can indicate the
number of monomers needed for activity, although other
contributions can complicate the situation (e.g., the thermo-
dynamic stability of the active complex).[17,20]
assay.[1,16,17] In this assay, large unilamellar vesicles composed
of egg yolk phosphatidylcholine (EYPC LUVs) are loaded
with the pH-sensitive fluorophore HPTS, and a pH gradient is
applied.[17] The ability of transport systems to accelerate the
dissipation of this pH gradient is then reported ratiometrically
by the intravesicular pH probes. In this assay, ditopic
calix[4]arene transporters were expected to cause the collapse
of pH gradients by chloride/hydroxide antiport.[17] Results
described in the following support that this antiport occurs by
chloride/TMA and hydroxide/TMA symport, with TMA
participating as counterion activator of the transporter that
neither adds nor removes transmembrane gradients.
The HPTS assay is convenient because anion/cation
symport, anion/anion antiport, and cation/cation antiport
are detected.[1,17] In a typical experiment, HPTS-loaded LUVs
were first exposed to a base pulse. Then, the potential
transporter was added, and the change in HPTS emission was
recorded. At the end of each experiment, the pH was
equilibrated with excess gramicidin D to calibrate for 100%
Single-atom F!I mutation from fluorine to iodine in the
anion binding site of 2 caused nearly complete inactivation
~
(Figure 1 and 2 ). Single-atom substitution to hydrogen
&
fluorescence change (Figure 2c , S1).
atoms in the anion binding site of 3 partially restored activity
*
Under these conditions, several promising candidates for
chloride/hydroxide transport with halogen bonds were inac-
tive (not shown). Inactivity of analogues of 1–6 with tert-butyl
groups at the upper rim was consistent with previous reports
(Figure 1 and 2a ).
The obtained activity sequence demonstrated that the
anion–p binding site in calix[4]arene 1 generates the highest
activity (Figure 1 and 2a&). Perfluorophenyls are classical
p acids. Their quadrupole moment with Qzz =+ 9.5 B has
about half the strength of that of naphthalenediimide (NDI)
transporters with Qzz =+ 19 B.[1,11] In calix[4]arene 1, this
reduced p acidity should, however, be compensated by the
preorganized tetravalent array along the macrocycle.[12–14]
The dramatic drop in activity from calix[4]arene 1 to 2
demonstrated that F!I mutation destroys the operational
~
anion–p active site in calix[4]arene 1 (Figure 1 and 2 ). This
abrupt change suggested that chloride/hydroxide binding by
four proximal activated halogen-bond donors in calix[4]arene
2 firstly occurs but secondly does not lead to transport.
19F NMR titrations in dry [D6]acetone revealed detectable
chloride binding to the inactive 2 (KD = 18.0 ꢀ 1.0 mm for
tetrabutylammonium (TBA) chloride) but not to the active 1
and 3 (Figure 3, S7–S9, Table S1).[21,22] Similar transport
inhibition by excessive binding has been observed previous-
ly.[1a,4] To transport anions with halogen bonds, either the
thermodynamic or the kinetic stability of the chloride/
hydroxide complex of 2 would therefore have to be reduced.
Thermodynamic destabilization of chloride/hydroxide
binding to 2 was achieved without global structural changes
by replacing the p-acidic with p-basic arenes (Figure 1). With
the weaker halogen-bond donors in calix[4]arene 5, chloride
binding became undetectable in 19F NMR titrations in dry
[D6]acetone,[16] whereas chloride/hydroxide transport activity
&
increased more than 30 times (Figure 1, 2 , S1). The EC50
=
Figure 2. Anion transport activity Y as a function of the concentration
32 mm of the halogen-bond transporter 5 was as good as that of
anion–p transporter 1 (Figure 1, 2). More than five times
decreasing activity in response to I!H mutation in calix[4]-
arene 6 provided corroborative evidence that transporter 5
indeed uses halogen bonds to transport chloride/hydroxide
anions (Figure 1, S1).
~
~
&
*
*
of a) 1 (&), 2 ( ), and 3 ( ), and b) 2 ( ), 4 ( ), and 5 ( ) with curve
&
fit to Hill equation (a, b). c) Original data for anion transport ( ) and
&
nonspecific leakage (ꢀ) mediated by 5 ( , 60 mm; ꢀ, 600 mm). These
&
original data show ( ) fractional emission IF during the addition of
NaOH (20 mL, 0.5m) and 5 to EYPC-LUVsꢁHPTS in TMACl (200 mm,
10 mm Hepes buffer, pH 7), with final calibration with gramicidin D
(200 s), and (ꢀ) the same for 5 added to EYPC-LUVsꢁCF.
11676 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 11675 –11678

Density functional theory (DFT) modelling at the
PBE1PBE/6-311G** level of theory showed that out of four
aryl groups, only three participate in the anion–p interactions
of the TMACl complex of calix[4]arene 1 (Figure 3b, S21),
resulting in a binding energy of ꢃ58.3 kcalmolꢃ1.[23] For the
meta-isomer 2, DFT models confirmed the reinforcement of
the binding energy (ꢃ70.9 kcalmolꢃ1) with the formation of
two halogen bonds, whereas the other two iodoarene donors
rest nearby to, we speculated, possibly increase probability
effects and inertness of the complex (Figure S21). In the
absence of perfluorinated iodoarenes, the two halogen bonds
of the TMACl complex of calix[4]arene 5 were maintained
(Figure S22). However, the binding energy clearly decreased
to ꢃ59.7 kcalmolꢃ1, which correlates with the observed regain
in the transport activity with 5.
All pertinent control experiments confirmed that calix[4]-
arenes 1–6 act as counterion-activated anion transporters with
operational halogen bonds and anion–p interactions. Namely,
external anion exchange produced changes in transport
activity, and revealed a general selectivity for chloride ions
(Figure S4). This responsiveness to external anion exchange
indicated that weak anion binding occurs and matters for
transport in all cases.[1,17] The nonconformity to the Hofmeis-
ter topology suggested that the cost of at least partial anion
desolvation is compensated by binding.[17] Inability to trans-
port 5(6)-carboxyfluorescein (CF) excluded the occurrence of
nonspecific leakage through larger defects in the membrane
(Figure 2c ꢀ , S6).
In summary, the absolute transport activity of calix[4]ar-
ene transporters was about as modest as expected from
literature.[2,3] However, the calix[4]arene scaffold was very
useful to orchestrate transport with halogen bonds and anion–
p interactions. Best transport activity was obtained for anion
binding at the focal point of four pentafluorobenzyl p acids,
thus demonstrating that the functional relevance of anion–p
interactions is general, independent of the structural motif
involved. Direct substitution of these p acids by halogen-bond
donors did not afford active anion transporters, whereas
anion binding was clearly detectable. Thermodynamic and
kinetic destabilization of this too efficient binder by weaken-
ing of the strength and the proximity of the halogen-bond
donors, respectively, provided rational approaches to active
anion transporters. These remarkably consistent results con-
firm synthetic transporters as unique systems to explore the
functional relevance of weak interactions that are otherwise
difficult to detect. This is the first time anion transport in lipid
bilayers has been achieved with halogen bonds.
Figure 3. Too strong binding hinders transport: a) Part of the 19F NMR
spectra in dry [D6]acetone of the least active transporter 2 and a,a,a-
trifluorotoluene (ꢃ63.72 ppm) in the presence of increasing concen-
trations of TBACl (0–500 mm, bottom to top). b) DFT-optimized
(PBE1PBE/6-311G**) TMACl complex of the most active transporter 1;
TBACl binding by 1 was not detectable by 19F NMR spectroscopy in dry
[D6]acetone (ball-and-stick representation: C: cyan; H: white; O: red;
N: blue; Cl ion: green; F: magenta).
Kinetic destabilization was envisioned by moving the
halogen-bond donors from meta-position in 2 to para-position
in 4.[22] This constitutional isomerization should increase the
distance between halogen-bond donors and move the active
site to the periphery of the complex (Figure 1). Consistent
with mainly kinetic complex destabilization, TBACl binding
by para-isomer 4 remained detectable in 19F NMR titrations
(KD = 13.3 ꢀ 0.6 mm for TBACl in dry [D6]acetone, Fig-
ure S7–S9), whereas transport activity increased 15 times to
[21]
*
an EC50 = 68 mm (Figure 1, 2 , S1).
Reduced proximity of the halogen-bond donors in para-
isomer 4 further caused an increase from the usual n ꢂ 1 to a
*
Hill coefficient n ꢂ 2 (Figure 1, 2 ). This change can suggest
that, with peripheral active sites, two calix[4]arenes are
required to sufficiently surround and transport one chloride/
hydroxide ion, whereas the number of proximal halogen-bond
donors in meta-isomers is already sufficient in 1:1 com-
plexes.[20] Job plots of 19F NMR titrations in dry [D6]acetone
supported the 1:1 stoichiometry of complexes with meta-
isomers such as 2, whereas in acetone, para-isomers 4 could
bind around two chloride ions (Figure S9b). This inversion of
stoichiometry with para-isomers 4 provided compelling
evidence that binding of TBACl in acetone and transport of
TMACl across lipid bilayers must be compared with highest
caution.[7,21,23]
Replacement of one or two haloaryl substituents in the
too efficient binder 2 with one or two methoxy groups to
reduce the number of halogen-bond donors did not afford
active transporters.[16,S5] This persistent inactivity was mean-
ingful because neither power nor proximity of the individual
halogen-bond donors are significantly changed by this
modification. Moreover, the essential cation-binding site
could possibly be perturbed by the appearance of stable
calix[4]arene isomers.[12,13] Detectability of TMA-independ-
ent chloride ion binding by homologues of 2 without one or
two haloaryl substituents by 19F NMR spectroscopy in dry
[D6]acetone confirmed that reduction in multivalency is
insufficient to significantly reduce the thermodynamic stabil-
ity of the halogen-bond chloride complexes.[S5]
Received: July 15, 2011
Revised: September 20, 2011
Published online: October 13, 2011
Keywords: anions · anion transport · anion–p interactions ·
.
halogen bonds
[1] a) R. E. Dawson, A. Hennig, D. P. Weimann, D. Emery, S.
Gabutti, J. Montenegro, V. Ravikumar, M. Mayor, J. Mareda,
C. A. Schalley, S. Matile, Nat. Chem. 2010, 2, 533 – 538; b) J.
ˇ
Mꢁsek, A. Vargas Jentzsch, S. Sakurai, D. Emery, J. Mareda, S.
Angew. Chem. Int. Ed. 2011, 50, 11675 –11678
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 11677

Communications
Matile, Angew. Chem. 2010, 122, 7846 – 7849; Angew. Chem. Int.
Ed. 2010, 49, 7680 – 7683.
Soc. 2007, 129, 48 – 58; e) D. QuiÇonero, C. Garau, C. Rotger, A.
Frontera, P. Ballester, A. Costa, P. M. Deya, Angew. Chem. 2002,
114, 3539 – 3542; Angew. Chem. Int. Ed. 2002, 41, 3389 – 3392;
f) M. Mascal, A. Armstrong, M. D. Bartberger, J. Am. Chem.
Soc. 2002, 124, 6274 – 6276; g) I. Alkorta, I. Rozas, J. Elguero, J.
Am. Chem. Soc. 2002, 124, 8593 – 8598.
[2] a) O. Lawal, K. S. J. Iqbal, A. Mohamadi, P. Razavi, H. T. Dodd,
M. C. Allen, S. Siddiqui, F. Fucassi, P. J. Cragg, Supramol. Chem.
2009, 21, 55 – 60; b) J. C. Iglesias-Sꢂnchez, W. Wang, R. Ferdani,
P. Prados, J. de Mendoza, G. W. Gokel, New J. Chem. 2008, 32,
878 – 890; c) O. A. Okunola, J. L. Seganish, K. J. Salimian, P. Y.
Zavalij, J. T. Davis, Tetrahedron 2007, 63, 10743 – 10750; d) Y.
Tanaka, Y. Kobuke, M. Sokabe, Angew. Chem. 1995, 107, 717 –
719; Angew. Chem. Int. Ed. Engl. 1995, 34, 693 – 694.
[3] a) J. K. Chui, T. M. Fyles, Chem. Soc. Rev. 2011, DOI: 10.1039/
C1CS15099E; b) S. Matile, A. Vargas Jentzsch, A. Fin, J.
Montenegro, Chem. Soc. Rev. 2011, 40, 2453 – 2474; c) U. Devi,
J. R. Brown, A. Almond, S. J. Webb, Langmuir 2011, 27, 1448 –
1456; d) F. Otis, C. Racine-Berthiaume, N. Voyer, J. Am. Chem.
Soc. 2011, 133, 6481 – 6483; e) H. Cho, L. Widanapathirana, Y.
Zhao, J. Am. Chem. Soc. 2011, 133, 141 – 147; f) J. T. Davis, P. A.
Gale, O. A. Okunola, P. Prados, J. C. Iglesias-Sanchez, T.
Torroba, R. Quesada, Nat. Chem. 2009, 1, 138 – 144; g) A. P.
Davis, D. N. Sheppard, B. D. Smith, Chem. Soc. Rev. 2007, 36,
348 – 357.
[12] Stable calix[4]arene isomers appeared in the 1H, 13C and
19F NMR spectra only upon removal of at least one bulky
substituent at the lower rim.[13] The energetics of conformational
changes of the flexible side chains has been determined in
rigidified cavitand analogues.[14]
[13] P. F. Hudrlik, A. M. Hudrlik, L. Zhang, W. Arasho, J. Cho, J. Org.
Chem. 2007, 72, 7858 – 7862.
[14] V. A. Azov, A. Beeby, M. Cacciarini, A. G. Cheetham, F.
Diederich, M. Frei, J. K. Gimzewski, V. Gramlich, B. Hecht, B.
Jaun, T. Latychevskaia, A. Lieb, Y. Lill, F. Marotti, A. Schlegel,
R. R. Schlittler, P. J. Skinner, P. Seiler, Y. Yamakoshi, Adv.
Funct. Mater. 2006, 16, 147 – 156.
[15] a) H.-J. Schneider, D. Guettes, U. Schneider, J. Am. Chem. Soc.
1988, 110, 6449 – 6454; b) J. M. Harrowfield, W. R. Richmond,
A. N. Sobolev, A. H. White, J. Chem. Soc. Perkin Trans. 2 1994,
5 – 9; c) W. Abraham, J. Inclusion Phenom. Macrocyclic Chem.
2002, 43, 159 – 174; d) L. M. Salonen, C. Bucher, D. W. Banner,
W. Haap, J.-L. Mary, J. Benz, O. Kuster, P. Seiler, W. B.
Schweizer, F. Diederich, Angew. Chem. 2009, 121, 825 – 828;
Angew. Chem. Int. Ed. 2009, 48, 811 – 814.
[4] J.-P. Behr, M. Kirch, J.-M. Lehn, J. Am. Chem. Soc. 1985, 107,
241 – 246.
[5] P. Metrangolo, F. Meyer, T. Pilati, G. Resnati, G. Terraneo,
Angew. Chem. 2008, 120, 6206 – 6220; Angew. Chem. Int. Ed.
2008, 47, 6114 – 6127.
[6] P. Politzer, J. S. Murray, T. Clark, Phys. Chem. Chem. Phys. 2010,
12, 7748 – 7757.
[7] Y. Lu, H. Li, X. Zhu, W. Zhu, H. Liu, J. Phys. Chem. A 2011, 115,
4467 – 4475.
[16] See the Supporting Information.
[17] S. Matile, N. Sakai, The Characterization of Synthetic Ion
Channels and Pores. Analytical Methods in Supramolecular
Chemistry (Ed.: C. A. Schalley), Wiley, Weinheim, 2007,
pp. 391 – 418.
[18] C. C. Tong, R. Quesada, J. L. Sessler, P. A. Gale, Chem.
Commun. 2008, 6321 – 6324.
[8] a) L. A. Hardegger, B. Kuhn, B. Spinnler, L. Anselm, R.
Ecabert, M. Stihle, B. Gsell, R. Thoma, J. Diez, J. Benz, J. M.
Plancher, G. Hartmann, D. W. Banner, W. Haap, F. Diederich,
Angew. Chem. 2011, 123, 329 – 334; Angew. Chem. Int. Ed. 2011,
50, 314 – 318; b) P. Auffinger, F. A. Hays, E. Westhof, P. S. Ho,
Proc. Natl. Acad. Sci. USA 2004, 101, 16789 – 16794.
[9] a) M. G. Chudzinski, C. A. McClary, M. S. Taylor, J. Am. Chem.
Soc. 2011, 133, 10559 – 10567; b) A. Caballero, N. G. White, P. D.
Beer, Angew. Chem. 2011, 123, 1885 – 1888; Angew. Chem. Int.
Ed. 2011, 50, 1845 – 1848; c) M. G. Sarwar, B. Dragisic, L. J.
Salsberg, C. Gouliaras, M. S. Taylor, J. Am. Chem. Soc. 2010, 132,
1646 – 1653; d) A. Mele, P. Metrangolo, H. Neukirch, T. Pilati, G.
Resnati, J. Am. Chem. Soc. 2005, 127, 14972 – 14973.
[10] a) S. Walter, F. Kniep, E. Herdtweck, S. Huber, Angew. Chem.
2011, 123, 7325 – 7329; Angew. Chem. Int. Ed. 2011, 50, 7187 –
7191; b) S. Dordonne, B. Crousse, D. Bonnet-Delpon, J. Legros,
Chem. Commun. 2011, 47, 5855 – 5857; c) D. A. Kraut, M. J.
Churchill, P. E. Dawson, D. Herschlag, ACS Chem. Biol. 2009, 4,
269 – 273; d) A. Bruckmann, M. A. Pena, C. Bolm, Synlett 2008,
900 – 902.
[19] T. Takeuchi, J. Montenegro, A. Hennig, S. Matile, Chem. Sci.
2011, 2, 303 – 307.
[20] S. Bhosale, S. Matile, Chirality 2006, 18, 849 – 856.
[21] Chloride ion binding studies by 19F NMR spectroscopy were not
possible with TMACl because of mutually incompatible solu-
bilities at higher concentrations. Successful binding studies with
tetrabutylammonium chloride (TBACl) in dry [D6]acetone have
to be interpreted carefully because TBA is too large for central
binding to the macrocycle,[15] a feature that is essential for
transport. Moreover, binding studies were highly solvent-
dependent, with binding constants decreasing rapidly with
traces of water in acetone.[7,23] KD values were determined by
nonlinear fitting of Dd vs. TBACl concentration as described in
literature.[9a,16]
[22] The synthesis of the ortho-isomer was not considered because
the interior is too overcrowded for hosting and external chloride
ion binding was already covered with 4.
[11] a) A. Frontera, P. Gamez, M. Mascal, T. J. Mooibroek, J.
Reedijk, Angew. Chem. 2011, 123, 9736 – 9756; Angew. Chem.
Int. Ed. 2011, 50, 9564 – 9583; b) L. M. Salonen, M. Ellermann, F.
Diederich, Angew. Chem. 2011, 123, 4908 – 4944; Angew. Chem.
Int. Ed. 2011, 50, 4808 – 4842; c) H. T. Chifotides, B. L. Schottel,
K. R. Dunbar, Angew. Chem. 2010, 122, 7360 – 7365; Angew.
Chem. Int. Ed. 2010, 49, 7202 – 7205; d) O. B. Berryman, V. S.
Bryantsev, D. P. Stay, D. W. Johnson, B. P. Hay, J. Am. Chem.
[23] Gas phase calculations were preferred because they presumably
mimic the situation in the apolar core of the lipid bilayer more
closely than results from competing polar solvents.[1,7,21] In
general, computational results that fail to include the changing
complex environments experienced during transport in lipid
bilayers have to be considered with appropriate reservation. For
the same reasons, efforts to characterize complexes in solid
state[5,8–10] appeared less meaningful in the context of this study.
11678 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 11675 –11678