Protein and small molecule recognition properties of deep cavitands in a supported lipid membrane determined by calcination-enhanced SPR spectroscopy

Article abstract of DOI:10.1021/ja102252d

This paper details the incorporation of a water-soluble deep cavitand into a membrane bilayer assembled onto a nanoglassified surface for study of molecular recognition in a membrane-mimicking setting. The cavitand retains its host properties, and real-time analysis of the host:guest properties of the membrane:cavitand complex via surface plasmon resonance and fluorescence microscopy is described. The host shows selectivity for choline-derived substrates, and no competitive incorporation of substrate is observed in the membrane bilayer. A variety of trimethylammonium-derived substrates are suitable guests, displaying varied binding affinities in a millimolar range. The membrane:cavitand:guest complexes can be subsequently used to capture NeutrAvidin protein at the membrane surface if a biotin-derived guest molecule is used. The surface coverage of NeutrAvidin is affected by the spacer used to derivatize the biotin. Increased distance from the bilayer allows a higher concentration of protein to be immobilized, suggesting a diminishing detrimental steric effect when the binding event is shifted away from the surface.

Full text of DOI:10.1021/ja102252d

Published on Web 07/09/2010
Protein and Small Molecule Recognition Properties of Deep
Cavitands in a Supported Lipid Membrane Determined by
Calcination-Enhanced SPR Spectroscopy
Ying Liu, Puhong Liao, Quan Cheng,* and Richard J. Hooley*
Department of Chemistry, UniVersity of California, RiVerside, California 92521
Received March 17, 2010; E-mail: richard.hooley@ucr.edu; quan.cheng@ucr.edu
Abstract: This paper details the incorporation of a water-soluble deep cavitand into a membrane bilayer
assembled onto a nanoglassified surface for study of molecular recognition in a membrane-mimicking setting.
The cavitand retains its host properties, and real-time analysis of the host:guest properties of the membrane:
cavitand complex via surface plasmon resonance and fluorescence microscopy is described. The host
shows selectivity for choline-derived substrates, and no competitive incorporation of substrate is observed
in the membrane bilayer. A variety of trimethylammonium-derived substrates are suitable guests, displaying
varied binding affinities in a millimolar range. The membrane:cavitand:guest complexes can be subsequently
used to capture NeutrAvidin protein at the membrane surface if a biotin-derived guest molecule is used.
The surface coverage of NeutrAvidin is affected by the spacer used to derivatize the biotin. Increased
distance from the bilayer allows a higher concentration of protein to be immobilized, suggesting a diminishing
detrimental steric effect when the binding event is shifted away from the surface.
Introduction
receptors typically consist of the covalent attachment of the
recognition motif to a lipid or steroid derivative that is
The surfaces of mammalian cells are decorated with a wide
variety of membrane-bound receptor molecules. These receptors
act as sensors for and transporters of small molecules in the
extracellular environment.1 For example, acetylcholinesterase
is a membrane-bound protein that binds the target acetylcholine
via noncovalent cation-π interactions with active site aromatic
residues.2 These natural receptors have inspired the discovery
of new methods to transport polar drug candidates across
hydrophobic membrane bilayers, a process of great importance
to medicinal chemistry.3 Small molecules have been exploited
to induce endocytosis4 or to shield the drug candidate hydro-
phobically as it is transported through the membrane.5 Many
examples of targeted drug delivery use natural receptors such
as glycolipids6 or polypeptides4b to recognize the target or study
the transport process alone by covalently linking the drug target
to a suitable vector.5 In these cases, the artificial membrane
incorporated in a synthetic membrane. The recognition motif
is displayed above the membrane surface in order to bind to
the target.
An alternate method of substrate recognition is the incorpora-
tion of a defined binding pocket inside the membrane bilayer.
This strategy is employed by membrane-penetrating proteins7
and transmembrane pore-forming peptides:8 the host is incor-
porated in the membrane itself and displays a cavity that allows
polar substrates to be shielded from the lipophilic membrane.9
This is mainly exploited for substrate transport through the
membrane, but a synthetic host that displayed this type of
binding motif would show a far greater range of application
than receptors created via the covalent derivatization of steroids
and lipids. To achieve this, we require a synthetic cavity that
can be incorporated into a membrane bilayer while still retaining
selective host properties. Most water-soluble synthetic host
molecules take advantage of the hydrophobic effect to recognize
their desired targets,10 which poses a problem for application
in natural systems. Hydrophobic substrates suffer from poor
water solubility and nonspecific localization in the lipophilic
membrane bilayers present in cells. In addition, competitive
binding of hydrophobic lipids themselves inside the host will
reduce the affinity for the target substrate. Because of these
(1) Conner, S. D.; Schmid, S. L. Nature 2003, 422, 37–44.
(2) Sussman, J. L.; Harel, M.; Frolow, F.; Oefner, C.; Goldman, A.; Toker,
L.; Silman, I. Science 1991, 253, 872–879.
(3) (a) Langer, R. Science 1990, 249, 1527–33. (b) Langer, R. Acc. Chem.
Res. 1993, 26, 537–542. (c) Peterson, B. R. Org. Biomol. Chem. 2005,
3, 3607–3612.
(4) (a) Boonyarattanakalin, S.; Martin, S. E.; Dykstra, S. A.; Peterson,
B. R. J. Am. Chem. Soc. 2004, 126, 16379–16386. (b) Boonyarat-
tanakalin, S.; Hu, J.; Dykstra-Rummel, S. A.; August, A.; Peterson,
B. R. J. Am. Chem. Soc. 2007, 129, 268–269. (c) Boonyarattanakalin,
S.; Martin, S. E.; Sun, Q.; Peterson, B. R. J. Am. Chem. Soc. 2006,
128, 11463–11470.
(7) Mackenzie, K. R. Chem. ReV. 2006, 106, 1931–1977.
(8) (a) Ghadiri, M. R.; Granja, J. R.; Buehler, L. K. Nature 1994, 369,
301–304. (b) Granja, J. R.; Ghadiri, M. R. J. Am. Chem. Soc. 1994,
116, 10785–10786. (c) Thore´n, P. E. G.; Persson, D.; Esbjo¨rner, E. K.;
Gokso¨r, M.; Lincoln, P.; Norde´n, B. Biochemistry 2004, 43, 3471–
3489.
(5) (a) Janout, V.; Lanier, M.; Regen, S. L. J. Am. Chem. Soc. 1996, 118,
1573–1574. (b) Janout, V.; Lanier, M.; Regen, S. L. J. Am. Chem.
Soc. 1997, 119, 640–647. (c) Janout, V.; DiGiorgio, C.; Regen, S. L.
J. Am. Chem. Soc. 2000, 122, 2671–2672. (d) Kondo, M.; Mehiri,
M.; Regen, S. L. J. Am. Chem. Soc. 2008, 130, 13771–13777. (e)
Mehiri, M.; Chen, W.-H.; Janout, V.; Regen, S. L. J. Am. Chem. Soc.
2009, 131, 1338–1339.
(9) (a) Cho, W.; Stahelin, R. V. Annu. ReV. Biophys. Biomol. Struct. 2005,
34, 119–151.
(6) Bertozzi, C. R.; Kiessling, L. L. Science 2001, 291, 2357–2364.
(10) Biros, S. M.; Rebek, J., Jr. Chem. Soc. ReV. 2007, 36, 93–105.
9
10.1021/ja102252d  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 10383–10390 10383

A R T I C L E S
Liu et al.
Figure 1. (a) Cavitands 1-3, (b) minimized conformation of 1 (SPARTAN, AM1 force field) with one bound choline molecule (4) in the cavity, and (c)
the guests used in this study.
factors, the use of cavity-based synthetic receptors in natural
systems is underexplored.
to apply this proof of principle to more biorelevant settings,
analysis of the host properties in bilayers and vesicles is
required. NMR analysis of the micellar properties provides some
Deep cavitands11 are well-known as protein mimics, in that
they provide a cavity that can selectively recognize molecules
of the correct shape if they also contain a thin layer of positive
charge at their surface.12 Excellent selectivity for substituted
trimethylammonium salts is possible in both water13 and organic
solvents.14 Recognition of hydrocarbons15 and steroids16 via the
hydrophobic effect has also been reported. As well as binding
hydrophobic species, molecules such as Rebek’s tetracarboxylate
cavitand 1 (Figure 1) have shown the ability to be incorporated
in lipid micelles.17 The host:guest properties of these small
micellar aggregates were studied by 1D NMR and diffusion
NMR techniques, indicating that the molecules were able to
bind suitable guests in the presence of lipids above the critical
micelle concentration, albeit with reduced affinities with respect
to those in pure water. Analysis of the composition of the
micelles was not performed, nor were studies carried out on
larger aggregates such as vesicles or membrane bilayers. In order
1
information, but H NMR spectroscopy is poorly suited for
analysis of membrane bilayers due to the substantial signal
broadening observed. A different sensing technique is therefore
required.
Supported lipid bilayers are an excellent mimic of natural
membranes, as they maintain similar fluidity properties18 and
substrate mobility. These membrane mimics have been suc-
cessfully used for the study of a variety of biological phenom-
ena.19 The predominant technique for characterizing activities
on the membrane has been fluorescence microscopy, while in
recent years label-free methods such as surface plasmon
resonance have gained considerable application due to simple
experimental procedure and real-time measurement capability.20
Proper functionalization of the gold substrates by hydrophilic
polymers in SPR is required in order to fabricate quality
membrane mimics. Recently an improved approach using a
calcinated nanoglassified gold surface has come forward and
allowed direct assembly of the supported membranes at the
sensing interface,21 allowing the real-time detection of binding
events in a supported membrane bilayer. The binding properties
(11) Rudkevich, D. M.; Hilmersson, G.; Rebek, J., Jr. J. Am. Chem. Soc.
1998, 120, 12216–12225.
(12) Hooley, R. J.; Rebek, J., Jr. Chem. Biol. 2009, 16, 255–264.
(13) (a) Hof, F.; Trembleau, L.; Ullrich, E. C.; Rebek, J., Jr. Angew. Chem.,
Int. Ed. 2003, 42, 3150–3153. (b) Biros, S. M.; Ullrich, E. C.; Hof,
F.; Trembleau, L.; Rebek, J., Jr. J. Am. Chem. Soc. 2004, 126, 2870–
2876.
(14) Menozzi, E.; Onagi, H.; Rheingold, A. L.; Rebek, J., Jr. Eur. J. Org.
Chem. 2005, 3633–3636.
(18) Cremer, P. S.; Boxer, S. G. J. Phys. Chem. B 1999, 103, 2554–2559.
(19) (a) Castellana, E. T.; Cremer, P. S. Surf. Sci. Rep. 2006, 61, 429–444.
(b) Tanaka, M.; Sackmann, E. Nature 2005, 437, 656–663. (c) Groves,
J. T.; Boxer, S. G. Acc. Chem. Res. 2002, 35, 149–157. (d) Schiller,
S. M.; Naumann, R.; Lovejoy, K.; Kunz, H.; Knoll, W. Angew. Chem.,
Int. Ed. 2003, 42, 208–211. (e) Perez, J. B.; Martinez, K. L.; Segura,
J. M.; Vogel, H. AdV. Funct. Mater. 2006, 16, 306–312.
(15) (a) Hooley, R. J.; van Anda, H. J.; Rebek, J., Jr. J. Am. Chem. Soc.
2007, 129, 13464–13473. (b) Hooley, R. J.; Biros, S. M.; Rebek, J.,
Jr. Chem. Commun. 2006, 50, 9–510. (c) Gibb, C. L. D.; Gibb, B. C.
J. Am. Chem. Soc. 2006, 128, 16498–16499.
(16) Gibb, C. L. D.; Gibb, B. C. J. Am. Chem. Soc. 2004, 126, 11408–
11409.
(20) Bieri, C.; Ernst, O. P.; Heyse, S.; Hofmann, K. P.; Vogel, H. Nat.
Biotechnol. 1999, 11, 1105–1108. (b) Reimhult, E.; Larsson, C.;
Kasemo, B.; Hook, F. Anal. Chem. 2004, 76, 7211–7220. (c) Wang,
Z. Z.; Wilkop, T.; Cheng, Q. Langmuir 2005, 21, 10292–10296.
(17) (a) Trembleau, L.; Rebek, J., Jr. Chem. Commun. 2004, 5, 8–59. (b)
Schramm, M. P.; Hooley, R. J.; Rebek, J., Jr. J. Am. Chem. Soc. 2007,
129, 9773–9779.
9
10384 J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010

Recognition Properties of Deep Cavitands
A R T I C L E S
of proteins such as lectins22 and bacterial toxins23 have been
extensively assessed with the calcinated gold chips.
Inc., Andover, MA, and used without further purification. 1-Palmi-
toyl-2-oleoyl-sn-glycero-3-phosphocholine was purchased from
Avanti Polar Lipids. All other materials were obtained from Aldrich
Chemical Co., St. Louis, MO and were used as received. Solvents
were dried through a commercial solvent purification system (SG
Water, Inc.). Cavitands 1-3 were synthesized according to literature
procedures.13b,14
Synthetic host molecules, such as calixarenes, shallow
cavitands, and cyclodextrins, have been directly attached to
surfaces for sensing applications.24 They have been tested for
detection of a range of small molecules, including gas vapors,
polycyclic aromatic hydrocarbons via surface-enhanced Raman
spectroscopy (SERS),25 and adrenaline/catecholamine via elec-
trochemistry.26 The results show satisfactory sensitivity, however
detection of small molecules by SPR can be challenging to
implement, as low molecular weight compounds are insuf-
ficiently large to generate a measurable refractive index change.
SPR sensing of small molecule interactions in a supported
bilayer poses an even greater challenge, as the recognition event
is displaced from the surface. In addition, the binding constants
obtainable by synthetic receptors are generally on the order of
millimolar (with some notable exceptions27), rather than the
micro- and nanomolar binding affinities displayed by proteins.
Binding properties of lectins that show millimolar binding
affinities are detectable,28 but any receptor incorporated in the
membrane must display at least millimolar affinity for substrate
in order for accurate analysis to be possible.
In this work, small molecule interactions with cavitand
incorporated supported bilayer lipid membrane are investigated
by SPR spectroscopy. The use of calcinated surface offers an
enhanced mode of SPR detection, allowing study of the
interactions of cavitands such as 1 with a series of choline
derivatives in a near-native cell membrane mimic. The binding
events are exploited to understand the interactions between polar
guests and the water-soluble host in biomimetic media, while
the employment of biotin-tagged guests aims to further explore
the protein binding properties at the membrane surface via a
host:guest handle.
2-(Thioureidonaphthalene)ethyltrimethylammonium Chloride
5. 2-Naphthyl isothiocyanate (19 mg, 0.10 mmol) was added to a
solution of 2-aminoethyltrimethylammonium chloride (18 mg, 0.10
mmol) and Et3N (10 µL) in DMSO (0.6 mL). After stirring for 1 h
at 70 °C, the solution was added to acetone (5 mL) and hexane (5
mL). The oily residue was separated by centrifugation and then
added to 10% aqueous NaOH solution (1.5 mL) and washed with
ethyl acetate (5 mL × 2). After neutralizing with 10% HCl solution,
the solvent was removed in vacuo to furnish product 5 (15 mg,
1
46%): H NMR (400 MHz, DMSO-d6) δ 11.93 (s, 1H), 10.40 (s,
1H), 8.20 (s, 2H), 7.80 (dd, J ) 8.6 Hz, 17.7, 7H), 7.69 (dd, J )
2.1 Hz, 8.8, 2H), 7.42 (dt, J ) 7.5 Hz, 15.9 Hz, 4H), 3.56 (t, J )
6.3 Hz, 4H), 3.31 (t, J ) 6.9 Hz, 4H), 3.17 (s, 9H); 13C NMR (100
MHz, DMSO-d6) δ 182.27, 139.00, 134.09, 131.32, 128.47, 128.17,
127.11, 125.66, 124.32, 119.38, 105.16, 64.80, 53.68, 38.52. HRMS
(ESI) m/z calcd for C16H22N3S (M+) 288.1529, found 288.1537.
2-(Thioureidofluorescein)ethyltrimethylammonium Chloride 6.
Fluorescein isothiocyanate (39 mg, 0.10 mmol) was added to a
solution of 2-aminoethyltrimethylammonium chloride (17 mg, 0.10
mmol) and K2CO3 (28 mg, 0.20 mmol) in water (1.5 mL). After
stirring for 18 h at room temperature, the solution was filtered and
the filtrand was added to acetone (5 mL). The precipitate was
separated by centrifugation and washed with acetone (3 mL) twice,
providing the desired compound 6 as a red solid (18 mg, 34%): 1H
NMR (400 MHz, D2O) δ 7.70 (d, J ) 2.1 Hz, 1H), 7.50 (dd, J )
2.2 Hz, 8.2 Hz, 1H), 7.21 (m, 3H), 6.68-6.60 (m, 4H), 4.13 (t, J
) 6.6 Hz, 2H), 3.63 (t, J ) 6.8 Hz, 2H), 3.25 (s, 10H); 13C NMR
(100 MHz, D2O) δ 180.79, 174.09, 157.75, 148.99, 143.98, 141.19,
131.49, 130.03, 126.16, 124.86, 123.02, 112.52, 103.77, 63.91,
53.54, 38.73; HRMS (ESI) m/z calcd for C26H26N3O5S (M+)
492.1588, found 492.1587.
Experimental Section
2-(Biotinamidyl)ethyltrimethylammonium Chloride 7. Isobu-
tyl chloroformate (16 µL) was added to a solution of biotin (25
mg) in DMF (0.6 mL) containing tri-N-butylamine (32 µL). After
10 min at room temperature, the mixture was added to a solution
of 2-aminoethyltrimethylammonium chloride (18 mg, 0.10 mmol)
in DMF/water (1:1). After stirring at room temperature overnight,
the mixture was added to 10% NaOH solution (1.5 mL) and washed
with ethyl acetate (5 mL × 2). After neutralizing with 10% HCl
solution, the solvent was removed in vacuo to furnish product 7
(18 mg, 49%): 1H NMR (400 MHz, D2O) δ 4.66 (dd, J ) 4.7 Hz,
7.8, 1H), 4.47 (dd, J ) 4.5 Hz, 7.9, 1H), 3.74 (t, J ) 6.3 Hz, 2H),
3.54 (t, J ) 6.7 Hz, 2H), 3.44-3.34 (m, 1H), 3.23 (s, 9H), 3.05 (q,
J ) 5.1 Hz 1H), 2.83 (d, J ) 13.1 Hz, 1H), 2.35 (t, J ) 7.3 Hz,
2H), 1.84-1.56 (m, 4H), 1.52-1.38 (m, 2H); 13C NMR (100 MHz,
D2O) δ 177.54, 165.52, 64.34, 62.26, 60.42, 55.52, 53.50, 39.84,
35.41, 33.59, 28.07, 27.81, 24.97; HRMS (ESI) m/z calcd for
C15H29N4O2S (M+) 329.2006, found 329.2007.
1H spectra were recorded on a Varian Inova 400 spectrometer.
Proton (1H) chemical shifts are reported in parts per million (δ)
with respect to tetramethylsilane (TMS, δ ) 0) and referenced
internally with respect to the protio solvent impurity. Deuterated
NMR solvents were obtained from Cambridge Isotope Laboratories,
(21) (a) Phillips, K. S.; Han, J.; Martinez, M.; Wang, Z.; Carter, D.; Cheng,
Q. Anal. Chem. 2006, 78, 596–603. (b) Han, J.; Taylor, J. D.; Phillips,
K. S.; Wang, X.; Feng, P.; Cheng, Q. Langmuir 2008, 24, 8127–8133.
(c) Linman, M. J.; Culver, S. P.; Cheng, Q. Langmuir 2009, 25, 3075–
3082.
(22) Linman, M. J.; Taylor, J. D.; Yu, H.; Chen, X.; Cheng, Q. Anal. Chem.
2008, 80, 4007–4013.
(23) Taylor, J. D.; Linman, M. J.; Wilkop, T.; Cheng, Q. Anal. Chem. 2009,
81, 1146–1153.
(24) (a) Daly, S. M.; Grassi, M.; Shenoy, S. K.; Ugozzoli, F.; Dalcanale,
E. J. Mater. Chem. 2007, 17, 1809–1818. (b) Schierbaum, K. D.;
Weiss, T.; Thoden van Velzen, E. U.; Engbersen, J. F. J.; Reinhoudt,
D. N.; Go¨pel, W. Science 1994, 265, 1413–1415. (c) Thoden van
Velzen, E. U.; Engbersen, J. F. J.; de Lange, P. J.; Mahy, J. W. G.;
Reinhoudt, D. N. J. Am. Chem. Soc. 1995, 117, 6853–6862. (d) Thoden
van Velzen, E. U.; Engbersen, J. F. J.; Reinhoudt, D. N. J. Am. Chem.
Soc. 1994, 116, 3597–3598. (e) Ludden, M. J. W.; Reinhoudt, D. N.;
Huskens, J. Chem. Soc. ReV. 2006, 35, 1122–1134.
2-(Biotinamidylcaproyl)ethyltrimethylammonium Chloride
8. Isobutyl chloroformate (9 µL) was added to a solution of
biotinylaminohexanoic acid29 (18 mg) in DMF (0.6 mL) containing
tri-N-butylamine (30 µL). After 10 min at room temperature, the
mixture was added to a solution of 2-aminoethyltrimethylammo-
nium chloride (9 mg, 0.05 mmol) in DMF/water (1:1). After stirring
at room temperature overnight, the mixture was added to 10%
NaOH solution (1.5 mL) and washed with ethyl acetate (5 mL ×
2). After neutralizing with 10% HCl solution, the solvent was
(25) Guerrini, L.; Garcia-Ramos, J. V.; Domingo, C.; Sanchez-Cortes, S.
Anal. Chem. 2009, 81, 1418–1425.
(26) Nikolelis, D. P.; Petropoulou, S. S. E.; Theoharis, G. Electroanalysis
2003, 15, 1616–1624.
1
(27) Rekharsky, M. V.; Mori, T.; Yang, C.; Ko, Y. H.; Selvapalam, N.;
Kim, H.; Sobransingh, D.; Kaifer, A. E.; Liu, S.; Isaacs, L.; Chen,
W.; Moghaddam, S.; Gilson, M. K.; Kim, K.; Inoue, Y. Proc. Natl.
Acad. Sci., U.S.A. 2007, 104, 20737–20742.
removed in vacuo to furnish product 8 (20 mg, 83%): H NMR
(29) Skander, M.; Humbert, N.; Collot, J.; Gradinaru, J.; Klein, G.; Loosli,
A.; Sauser, J.; Zocchi, A.; Gilardoni, F.; Ward, T. R. J. Am. Chem.
Soc. 2004, 126, 14411–14418.
(28) Mann, D. A.; Kanai, M.; Maly, D. J.; Kiessling, L. L. J. Am. Chem.
Soc. 1998, 120, 10575–10582.
9
J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010 10385

A R T I C L E S
Liu et al.
(400 MHz, D2O) δ 4.64 (dd, J ) 5.0, 8.0, 1H), 4.46 (dd, J ) 4.5,
7.9, 1H), 3.72 (t, J ) 6.7, 2H), 3.51 (t, J ) 6.7, 2H), 3.44-3.33
(m, 1H), 3.21 (s, 9H), 3.03 (q, J ) 5.0 Hz, 1H), 2.82 (d, J ) 13.1
Hz, 1H), 2.35-2.24 (m, 4H), 1.82-1.50 (m, 8H), 1.48-1.40 (m,
2H), 1.39-1.30 (m, 2H); 13C NMR (100 MHz, D2O) δ 177.52,
176.79, 165.50, 64.32, 62.24, 60.41, 55.52, 53.50, 39.87, 39.19,
35.63, 33.60, 28.16, 27.98, 27.83, 25.76, 25.35, 24.88, 23.37; HRMS
(ESI) m/z calcd for C21H40N5O3S (M+) 442.2846, found 442.2845.
Calcinated Chip Preparation. Gold substrates were fabricated
with a 2 nm thick chromium adhesion layer, followed by deposition
of a 46 nm thick gold layer via e-beam evaporation onto cleaned
glass slides. The nanoglassified layers were constructed on the
surface based on a previous layer-by-layer protocol.30 Clean gold
substrates were immersed in 10 mM 3-mercaptopropionic acid
(MPA) ethanol solution overnight to form a self-assembled mono-
layer. After extensive rinsing with ethanol and nanopure water and
drying with nitrogen gas, modified gold substrates were alternately
dipped into sodium silicate solution (22 mg/mL, adjusted to pH
9.5) and poly(allylamine hydrochloride) solution (1 mg/mL, adjusted
to pH 8.0) for 1 min to form a layer by layer assembly structure,
with sufficient nanopure water rinsing between layers. This dipping
process was repeated eight times to build up a multilayered chip,
followed by calcination in a furnace by heating to 450 °C at a rate
of 17 °C per min and subsequent cooling to room temperature 4 h
later.
Vesicle Preparation. PC lipid stock solution was transferred to
a small vial and the organic solvent was purged from the vial with
N2 to form a dry lipid film on the vial wall, which was then
rehydrated with 20 mM phosphate buffered saline (150 mM NaCl,
pH 7.40) to a lipid concentration of 1 mg/mL. The resuspended
lipids were probe sonicated for 20 min, followed by centrifugation
at 8000 rpm for 6 min to remove any titanium particles released
from the probe tip. The supernatant was then extruded with 11
passes through a polycarbonate membrane of pore size 100 nm to
ensure formation of small unilamellar vesicles. The solution was
then incubated at 4 °C for at least 1 h before use.
formula (eq 1) when the membrane between adsorbate and metal
surface is very thin compared to plasmon decay length.31 Cavitand
1 surface coverage (in units of molecules/cm2) can be estimated
by the SPR signal increase
θ ) NId/2[R/m(ηa - ηs)]
(1)
where N is the bulk number density of the adsorbate, Id is roughly
estimated as 0.37 of the light wavelength (670 nm here), R is SPR
response (in the unit of degrees) via binding, m is a instrument
constant (determined experimentally by calibrating the measured
sensor response to changes in refractive index), ηa is the refractive
index of adsorbed molecule, and ηs is the refractive index of bulk
solution (1.33 for water). Id is much larger than the thickness of
bilayer lipid membrane (∼5 nm), allowing correct application of
eq 1.31
To calculate the surface coverage of cavitand 1, the bulk
refractive index of cavitand 1 was estimated as 1.6013 and bulk
density was estimated as 1.1254 g/cm3.32 From the SPR signal
increase upon cavitand 1 binding (0.20 ( 0.02 degree), θ was
determined to be 8.02 ( 0.93 × 1013 molecules/cm2. NeutrAvidin
binding to the 1•7 complex caused a signal increase of 0.25° (
0.05°. Crystalline protein refractive index (1.60) was used as bulk
refractive index,31 and 1.43 g/cm3 as bulk density;33 thus, the θ of
NeutrAvidin was determined to be 2.94 ( 0.62 × 1012 molecules/
cm2.
Kinetic Analysis. Saturation binding mode (eq 2)23 was applied
here to determine the equilibrium dissociation constant (KD) value
for the interaction between cavitand 1 and guest 6. Increasing
concentrations of guest 6 (0.5-5 mM) were injected over the
cavitand 1:membrane complex, and the minimum angle shift was
recorded
ABeq ) ABmax{1/(1 + KD/[A])}
(2)
where ABeq is the average of the response signal at equilibrium
and ABmax is the maximum response that can be obtained for guest
Fabrication of Cavitand 1 Receptor Layer and Guest
Binding Measurement. The fabrication of cavitand 1-membrane
complex and subsequent guest binding was monitored through
surface plasmon resonance (SPR) spectrometry and fluorescence
microscopy. The shift of SPR minimum angle characterized surface
thickness and surface refractive index change, demonstrating
adsorption or binding on the surface. The calcinated gold substrate
was first rinsed with ethanol and nanopure water and after drying
under a gentle stream of N2 gas was then clamped down by a flow
cell on a high-refractive index prism for SPR measurement. PC
vesicles (1 mg/mL) in 20 mM phosphate buffered saline (150 mM
NaCl, pH 7.40) were injected through a flow-injection system and
incubated for 1 h to allow vesicle fusion on the hydrophilic
calcinated gold surface, forming a smooth bilayer membrane. After
10 min of rinsing to remove excess vesicle from the surface, 2
mg/mL cavitand 1 in 10% DMSO solution was subsequently
injected and incubated for 20 min. The surface was extensively
rinsed with nanopure water, followed by incubation with 2 mM
aqueous solution of guest 4-8 for varied times. Control experiments
were performed under identical conditions in the absence of cavitand
1.
6 binding and [A] is the concentration of guest 6 injection. ABmax
/
ABeq was plotted against 1/[A], and the slope is equal to the KD
value (3.54 ( 0.92 mM). KA, the equilibrium association constant
(282 ( 73 M-1), can be determined as the reciprocal value of KD.
This process was used to determine KD/KA for guests 6-8 and
NeutrAvidin.
In transient SPR response mode,34 the initial rate kinetic method
was employed to extract association constant (kassoc), dissociation
constant (kdiss), and KD value for the interaction between cavitand
1 and guest 6. At t ) 0, the equation for initial rate analysis is
dAB/dt ) ABmax[A]kassoc
(3)
By plotting the initial rate against [A], kassoc was calculated as
289.76 M-1 min-1. kdiss can be determined from the dissociation
curve of SPR sensorgram
ABt ) (AB0 - AB∞)[exp(-kdisst)] + AB∞
(4)
where AB0 is the initial response at the beginning of the dissociation
curve and AB∞ is the final response once completely dissociated.
Given that the kdiss value is 1.304 min-1, KD is thus determined to
For fluorescein guest 6, fluorescence microscopy experiments
were also performed on the same chip setup. After sufficient rinsing
with nanopure water, fluorescence microscopy was carried out on
a Zeiss LSM 510 confocal laser scanning microscope with 488 nm
argon laser excitation and a CCD camera. For comparison, an
aqueous solution of guest 6 was injected onto the PC membrane
surface as before, in the absence of cavitand 1.
be 3.54 mM via KD ) kdiss/kassoc
.
Results and Discussion
Three cavitands were initially tested for their associative
properties with the membrane bilayer: water-soluble tetracar-
Surface Coverage Calculations. The surface coverage of
adsorbates on a membrane bilayer can be estimated via Jung’s
(31) Jung, L. S.; Campbell, C. T.; Chinowsky, T. M.; Mar, M. N.; Yee,
S. S. Langmuir 1998, 14, 5636–5648.
(32) Lide, D. R., Ed. Handbook of Chemistry Physics, 71st ed.; CRC Press:
Boston, 1990.
(30) Linman, M. J.; Culver, S. P.; Cheng, Q. Langmuir 2009, 25, 3075–
(33) Connolly, M. L. J. Mol. Graph. 1993, 11, 139–141.
(34) Edwards, P. R.; Leatherbarrow, R. J. Anal. Biochem. 1997, 246, 1–6.
3082.
9
10386 J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010

Recognition Properties of Deep Cavitands
A R T I C L E S
Scheme 1. A Cartoon Representation of Guests Binding to
Cavitand 1 Incorporated in a PC Bilayer Membrane
Figure 2. SPR sensorgrams for guest 6 interaction with PC bilayer
membrane in the presence (left) and absence (right) of cavitand 1.
The sensorgram does not, however, give any indication of
the orientation of cavitand 1 in the membrane. To determine
whether its binding properties remained intact, the cavitand:
membrane system was exposed to the guest molecules shown
in Figure 1c. To minimize nonspecific interaction between the
guest and the membrane bilayer, commercially available long
chain alkyltrimethylammonium salts were avoided. The guests
used in this study were synthesized by the coupling of
2-(trimethylammonium)ethylamine chloride with either the
isothiocyanate (5, 6) or acid chloride (7, 8) of the corresponding
headgroup (Figure 1c). All showed good water solubility (g20
mM) and were analyzed by 1H NMR spectroscopy before
injection.
Upon addition of choline 4, no change in resonance angle
was observed. This is not entirely unexpected, as molecular
modeling (Figure 1b) indicates that choline is completely
surrounded by the cavity upon binding and does not protrude
into the solvent above the membrane. The small change in the
membrane composition upon small guest binding would be
difficult to detect, however, so host:guest affinity cannot be ruled
out. Upon addition of larger guests 5-8 to the system, a small,
reproducible change in resonance angle was observed, indicating
guest binding and concomitant change in refractive index of
the membrane. The SPR response to guest binding was
dependent on the molecular weight of the substrate. SPR signal
response increased with increasing guest molecular weight and
a corresponding increase in perturbation of the environment at
the membrane-water interface. The change in resonance angle
for the three largest guests 6-8 was large enough to allow
reasonably accurate calculation of KD/KA values, although this
was not possible for smaller guests 4 and 5.
The values of equilibrium dissociation constant (KD) and
binding affinity (KA) for the host:guest interactions between
guests 6-8 and cavitand 1 were determined from SPR sensor-
grams via saturation binding mode (see Experimental Section).
Guest 6 showed millimolar binding affinity (KD ) 3.54 ( 0.92
mM, KA ) 282 ( 73 M-1). This value is 2 orders of magnitude
smaller than that observed for substituted trimethylammonium
salts in pure water but is consistent with the binding affinity
observed under physiological conditions (phosphate buffer, as
used in these experiments).13b For comparison, the transient SPR
response method was also used to determine the binding constant
to the cavitand:membrane complex. The KD value (4.55 mM)
agrees well with the number derived from equilibrium binding
analysis. A stronger affinity was determined for biotin-derived
boxylate cavitand 1 and its more lipophilic counterparts 2 and
3. The synthesis of 1-3 and their binding properties in both
water and organic solvents have been previously reported.13,14,35
Cavitand 1 is soluble in water at millimolar concentrations,
whereas lipophilic cavitands 2 and 3 are only sparingly soluble
in pure water but can be incorporated into micelles in the
presence of added lipid.17b All three hosts show millimolar
affinity for choline and related trimethylammonium salts due
to favorable cation-π interactions between the aromatic cav-
itand walls and the guest. In water, choline has a binding affinity
for 1 of 2.6 × 104 M-1, although the KA was lessened in an
SDS micelle.17a A significant advantage of cavitands as hosts
is their open-ended character; long guests can extend out of
the cavity, presenting large functional groups into the bulk
solvent. This allows a variety of guest derivatives to be tested.
The experiment is illustrated in Scheme 1. The membrane
was deposited on a nanoglassified gold chip in a flow-cell (see
the Experimental Section).22 The membrane was fabricated by
the injection of preformed L-R phosphatidylcholine (PC) vesicles
that fuse readily on this chip. Cavitands 1-3 were subsequently
injected into the system in a 10% DMSO:H2O solution (to
maximize solubility) followed by a rinsing stage to remove all
DMSO and unincorporated cavitand from the flow cell. Finally,
a suitable guest (Figure 1c) was introduced to the system by
injection, followed by copious rinsing to remove the unincor-
porated excess. PC was chosen as the constituent lipid to
minimize the host:guest binding of the lipid inside the cavitand.
The two alkyl chains in PC (oleoyl and palmitoyl) cannot both
fit inside the cavity (as has been observed with SDS17a), and
the phosphate anion in the phosphocholine group minimizes
binding at that terminus via a repulsive interaction with the
carboxylates at the cavitand rim.13b
The effect of addition of cavitand 1 to the membrane bilayer
is shown by the SPR sensorgram in Figure 2. The change in
resonance angle upon addition indicates the binding of cavitand
1 to the PC membrane. The cavitand remains bound to the
membrane bilayer even after extensive rinsing, and the angular
shift indicates a surface coverage of 8.02 ( 0.93 × 1013
molecules/cm2 (see Experimental Section). The incorporation
of cavitands 2 and 3 in the PC membrane was also observed,
but with a substantially lower change in SPR signal and, thus,
surface coverage. We attribute this to their poor solubility in
water and limited exposure to the membrane in the flow cell
setting. All subsequent analyses were performed using water-
soluble cavitand 1.
guests 7 ([KD (7•1) ) 1.87 ( 0.24 mM, KA ) 535 ( 68 M-1
]
and 8 ([KD (8•1) ) 1.95 ( 0.61 mM, KA ) 513 ( 160 M-1].
These binding affinities show that the cavitand retains most of
its host abilities while incorporated in the membrane. Further-
more, control experiments in the absence of cavitand 1 show
no SPR response from addition of any of the guests 4-8 to the
(35) Far, A. R.; Shivanyuk, A.; Rebek, J., Jr. J. Am. Chem. Soc. 2002,
124, 2854–2855.
9
J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010 10387

A R T I C L E S
Liu et al.
Table 1. Binding Properties of Guest Molecules 4-8 in the PC•1
Scheme 2. A Cartoon Representation of NeutrAvidin Binding to the
Complex of Cavitand 1 and Guests 7 or 8 Incorporated in a PC
Bilayer Membrane
Complex
guest
Mw
KA/M-1
KD/mM
4
5
6
7
8
104.2
288.5
492.6
329.2
442.3
N/A
N/A
282 ( 73
535 ( 68
513 ( 160
N/A
N/A
3.54 ( 0.92
1.87 ( 0.24
1.95 ( 0.61
supported membrane. This indicates that there is no nonspecific
incorporation of the targets in the membrane even in the case
of mildly lipophilic naphthalene guest 5.
It should be noted that although the changes in resonance
angle observed upon guest binding are small, the measurements
are reproducible. Further evidence for guest incorporation was
obtained by fluorescence microscopy. Construction of the
membrane:cavitand 1:guest 6 complex was performed as before,
and the chip visualized under a confocal microscope. The
fluorescence microscopic images are shown in Figure 3. After
extensive rinsing, the boundary of the flow cell with 6 was
clearly visible, demonstrating fluorescence signal from the guest
6•cavitand 1 complex incorporated in the membrane bilayer.
The host:guest system was strong enough to retain the substrate
in the membrane for a period of many minutes under rinse
conditions. In the absence of cavitand, little fluorescence was
observed under the same excitation condition, providing further
evidence that there was little nonspecific adsorption of 6 on
the membrane.
The result also illustrates the advantage of the choline-based
guest system. Nonspecific interaction between organic molecules
and membrane bilayers is a significant problem when creating
shape-based receptors for membranes. Hydrophobic substrates
would be susceptible to incorporation in the membrane itself,
but that is not observed in this case; negligible binding of the
guests 4-8 can be observed in the absence of cavitand by either
fluorescence (if applicable) or SPR.
These results show that cavitand 1 is able to noncovalently
bind a variety of trimethylammonium-tagged species while
incorporated in the supported membrane bilayer. The open-
ended nature of the cavitand allows a wide scope of guest size;
the trimethylammonium group is bound in the cavity, and the
unencapsulated portion of the guest is displayed above the
membrane surface. This suggests that a guest that contains a
recognition element itself should be able to recognize its target
while bound to the membrane:cavitand complex. Compounds
7 and 8 were used for this purpose. Both guests display a biotin
group above the membrane surface, differing only in the space
between the biotin recognition motif and cavitand/membrane
complex.
Upon addition of NeutrAvidin to the preformed membrane:
cavitand 1:guest 7 complex, a significant change in resonance
angle (∆R ) 0.20° ( 0.02°) was observed, indicating that the
complex is capable of displaying the biotin tag above the
membrane while retaining its affinity for a suitable protein. As
would be expected due to its much greater size and charge,
NeutrAvidin displays a significantly larger ∆R upon binding
than observed for the binding of guests 5-8. Both biotin-derived
guests 7 and 8 displayed affinity for NeutrAvidin (the SPR
sensorgrams for guest 7 are shown in Figure 4, and those for
guest 8 are shown in the Supporting Information). The sequential
binding of cavitand, guest, and NeutrAvidin are all observable.
Addition of NeutrAvidin to the membrane:cavitand 1:guest 8
complex caused a much larger change in resonance angle (∆R
) 0.42° ( 0.06°).
Control experiments were performed by substituting biotin
itself for guests 7/8 (Figure 4b) and the addition of guests 7/8
to the membrane followed by NeutrAvidin in the absence of
cavitand 1. All controls were performed with the same
concentration of guests. The sensorgram in Figure 4b displays
no increase in resonance angle upon addition of biotin or
NeutrAvidin, indicating minimal nonspecific interactions of the
protein or biotin with the cavitand or the membrane itself. In
the absence of cavitand 1 (Figure 4c), neither guest 7 nor
NeutrAvidin show significant incorporation in the membrane.
The presence of both cavitand 1 and the trimethylammonium
binding handle are required to attach the protein to the surface;
a four-component binding event is seen, with only noncovalent
interactions binding the cavitand, guest, and NeutrAvidin to the
Figure 4. SPR sensorgrams for NeutrAvidin interaction with a supported
PC membrane containing (a) cavitand:guest complex 1•7, (b) cavitand 1
and biotin control, and (c) derivatized biotin guest 7 in the absence of
cavitand.
Figure 3. Fluorescence microscopic images of the calcinated chip
containing (a) PC bilayer, cavitand 1, and guest 6 and (b) PC bilayer and
guest 6 alone (control).
9
10388 J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010

Recognition Properties of Deep Cavitands
A R T I C L E S
supported membrane bilayer. The binding affinities are strong
enough to hold the four-component complex together during
the final washing phase.
The NeutrAvidin surface coverage was calculated by SPR
response as before. When immobilized by the cavitand 1•guest
7 complex, NeutrAvidin surface coverage was estimated to be
2.94 ( 0.62 × 1012 molecules/cm2, [cf. θ (cavitand 1) ) 8.02
( 0.93 × 1013 molecules/cm2]. The cavitand:NeutrAvidin ratio
is 27.3:1, which is to be expected due to the vast difference in
sizes between cavitand 1 (MW ) 1377 Da) and NeutrAvidin
(MW ) 60 kDa). The millimolar binding affinity of guest 7
will also contribute to this difference, as not all host molecules
will be occupied.
Although there was minimal change in the binding affinity
for guests 7 and 8 in cavitand 1, a significant difference in SPR
response was detected upon NeutrAvidin binding. The surface
coverage of NeutrAvidin when bound to the PC•1•guest 8
complex is 4.94 ( 0.71 × 1012 molecules/cm2, which is 2-fold
greater than that observed for the shorter guest 7. This is most
likely due to reduced steric interactions between the protein and
the membrane bilayer when the longer guest 8 is used as the
recognition motif.
In addition to the surface coverage experiments, the binding
affinity of NeutrAvidin for the cavitand•guest complex could
be measured. This is unusual, as the biotin:NeutrAvidin binding
affinity is ∼1015 M-1, far too high for analysis by SPR. In this
case, the saturation binding mode gave the biotin:NeutrAvidin
binding constant of 3.64 × 105 M-1. This lowered binding
affinity is most likely not due to the biotin:NeutrAvidin
interaction but the interaction between NeutrAvidin•8 complex
and the membrane:cavitand complex; the “weakest link” is the
cavitand:trimethylammonium interaction, and some of the guest
may be pulled out of the cavitand upon NeutrAvidin coordina-
tion. Given that the binding affinity is at least partially driven
by the hydrophobic effect,36 it is unsurprising that the affinity
decreases upon conversion of the small 8 to the large, extremely
hydrophilic NeutrAvidin•8 complex. The binding affinity was
determined by regression analysis, and the best fit was obtained
by assuming a 1:1 biotin:avidin binding motif. It is possible
that multivalent binding between NeutrAvidin and the 1•8
complex can occur,37 but our binding data suggests this is not
dominant in this case, most probably due to the low density of
1•8 in the membrane.
Figure 5. Illustration of the relative sizes of the cavitand 1•guest 6 complex
and POPC lipids (host:guest complex minimized by SPARTAN, AM1 force
field).
lipids that can act as competitive substrates, in buffered solution
that lowers substrate affinity, and despite shear forces from the
flow cell, a host the diameter of only two POPC lipid molecules
is able to immobilize a 60 kDa protein. This protein is held in
place solely by cation-π interactions between eight aromatic
rings and a single trimethylammonium cation. Cation-π
interactions are a vital component of protein:substrate recogni-
tion events,38 but this shows their potential as the “mortar”
holding together quaternary nanoscale constructs.
Conclusions
In this work, we have shown that a deep water-soluble
cavitand can be incorporated in a membrane bilayer attached
to a nanoglassified surface, while still retaining its host
properties. The open-ended nature of the cavitand allows binding
of a variety of guest sizes; the trimethylammonium binding
“handle” is incorporated inside the cavity while the rest of the
molecule is displayed above the cavity in the aqueous phase.
Competitive binding from the membrane lipids is minimal, and
association constants on the order of 103 M-1 are observed with
a variety of trimethylammonium-derived substrates. The binding
events can be observed in real-time by SPR spectroscopy and
the binding of fluorescent derivatives observed by fluorescence
microscopy. No nonspecific binding of the guest molecules in
the membrane itself can be observed in the absence of cavitand;
the host can selectively recognize its targets in a biorelevant
setting. The membrane:cavitand:guest complexes can subse-
quently be used to sense proteins at the membrane surface. By
the use of a suitable biotin-derived guest molecule, the biotin
motif is displayed above the membrane bilayer and is able to
immobilize NeutrAvidin at the surface, a process detectable by
SPR. The surface coverage is dependent on the spacer used to
derivatize the biotin: an increased distance from the bilayer
allows a higher concentration of protein to be immobilized.
These binding studies showcase the potential of this system
as a host for larger, more hydrophilic aggregates; the trimethy-
lammonium:cavitand interaction is not just a product of the
hydrophobic effect, so it is tolerant of the binding of highly
hydrophilic species such as proteins. Lipid/steroid “anchors”
can be rendered useless if the external group is too hydrophilic
Figure 5 shows an illustration of the relative sizes of the
cavitand, guest, and lipid bilayer (as determined by molecular
modeling; SPARTAN, AM1 force field). The cavitand is
relatively small, with a vertical height of approximately 13 Å.
When compared to the POPC length (∼26 Å), it becomes clear
that the cavitand is most likely positioned toward the top of the
lipid bilayer, incorporated in the first layer such that the
negatively charged carboxylates are in proximity with the
aqueous exterior, and the hydrophobic cavitand body is incor-
porated among the lipid hydrocarbon chains. This allows the
host to present suitable guests to the exterior milieu.
The binding abilities of the membrane:cavitand:trimethylam-
monium complexes are quite remarkable: in the presence of
(36) Hooley, R. J.; Biros, S. M.; Rebek, J., Jr. Angew. Chem., Int. Ed.
2006, 45, 3517–3519.
(37) (a) Jung, H.; Robison, A. D.; Cremer, P. S. J. Struct. Biol. 2009, 168,
90–94. (b) Yang, T.; Baryshnikova, O. K.; Mao, H.; Holden, M. A.;
Cremer, P. S. J. Am. Chem. Soc. 2003, 125, 4779–4784. (c) Pe´rez-
Luna, V. H.; O’Brien, M. J.; Opperman, K. A.; Hampton, P. D.; Lo´pez,
G. P.; Klumb, L. A.; Stayton, P. S. J. Am. Chem. Soc. 1999, 121,
6469–6478.
(38) Ma, J. C.; Dougherty, D. A. Chem. ReV. 1997, 97, 1303–1324.
9
J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010 10389

A R T I C L E S
Liu et al.
(as has been observed for lipophilic oligonucleotides39); exploit-
ing nonhydrophobic interactions to anchor species in a mem-
brane can allow a far greater range of target species. Further
studies on the properties of membrane-bound hosts are underway.
Acknowledgment. We are grateful to Profs. Julius Rebek,
Jr. and Michael Schramm for helpful discussions. Q.C. acknowl-
edges financial support from NSF (CHE-0719224) and NIH
(1R21EB009551).
Supporting Information Available: SPR sensorgrams and
binding calculations. This material is available free of charge
via the Internet at http://pubs.acs.org.
(39) (a) Bunge, A.; Kurz, A.; Windeck, A.-K.; Korte, T.; Flasche, W.;
Liebscher, J.; Herrmann, A.; Huster, D. Langmuir 2007, 23, 4455–
4464. (b) Jakobsen, U.; Simonsen, A. C.; Vogel, S. J. Am. Chem.
Soc. 2008, 130, 10462–1046.
JA102252D
9
10390 J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010