Environmental effects dominate the folding of oligocholates in solution, surfactant micelles, and lipid membranes

Article abstract of DOI:10.1021/ja103694p

Oligocholate foldamers with different numbers and locations of guanidinium-carboxylate salt bridges were synthesized. The salt bridges were introduced by incorporating arginine and glutamic acid residues into the foldamer sequence. The conformations of these foldamers were studied by fluorescence spectroscopy in homogeneous solution, anionic and nonionic micelles, and lipid bilayers. Environmental effects instead of inherent foldability were found to dominate the folding. As different noncovalent forces become involved in the conformations of the molecules, the best folder in one environment could turn into the worst in another. Preferential solvation was the main driving force for the folding of oligocholates in solution. The molecules behaved very differently in micelles and lipid bilayers, with the most critical factors controlling the folding-unfolding equilibrium being the solvation of ionic groups and the abilities of the surfactants/lipids to compete for the salt bridge. Because of their ability to fold into helices with a nonpolar exterior and a polar interior, the oligocholates could transport large hydrophilic molecules such as carboxyfluorescein across lipid bilayers. Both the conformational properties of the oligocholates and their binding with the guest were important to the transport efficiency.

Full text of DOI:10.1021/ja103694p

Published on Web 06/24/2010
Environmental Effects Dominate the Folding of Oligocholates
in Solution, Surfactant Micelles, and Lipid Membranes
Hongkwan Cho and Yan Zhao*
Department of Chemistry, Iowa State UniVersity, Ames, Iowa 50011-3111
Received April 30, 2010; E-mail: zhaoy@iastate.edu
Abstract: Oligocholate foldamers with different numbers and locations of guanidinium-carboxylate salt
bridges were synthesized. The salt bridges were introduced by incorporating arginine and glutamic acid
residues into the foldamer sequence. The conformations of these foldamers were studied by fluorescence
spectroscopy in homogeneous solution, anionic and nonionic micelles, and lipid bilayers. Environmental
effects instead of inherent foldability were found to dominate the folding. As different noncovalent forces
become involved in the conformations of the molecules, the best folder in one environment could turn into
the worst in another. Preferential solvation was the main driving force for the folding of oligocholates in
solution. The molecules behaved very differently in micelles and lipid bilayers, with the most critical factors
controlling the folding-unfolding equilibrium being the solvation of ionic groups and the abilities of the
surfactants/lipids to compete for the salt bridge. Because of their ability to fold into helices with a nonpolar
exterior and a polar interior, the oligocholates could transport large hydrophilic molecules such as
carboxyfluorescein across lipid bilayers. Both the conformational properties of the oligocholates and their
binding with the guest were important to the transport efficiency.
Introduction
Intrigued by the exquisite conformational control of biomol-
ecules by nature, chemists began to design and synthesize
synthetic molecules (i.e., foldamers) adopting ordered, control-
lable conformations.3 The field of foldamers has undergone rapid
development in recent years. As new folding motifs4 and
building blocks5 were created, novel properties and functions,
sometimes unavailable from biofoldamers, began to emerge.
Naturally, the learning generated from biomimetic, synthetic
foldamers was used to understand and solve biological problems.
Foldamer-based synthetic antimicrobial agents,6 protein surface
binders and inhibitors,7 vesicles and organogellators,8 and
biomimetic enantioselective catalysts9 all appeared in the
literature.
Environments can have strong influences on the conformation
of a molecule. Proteins, for example, frequently change their
conformation as they bind to ligands, come in contact with
membranes, or interact with other proteins. As the environment
of a molecule becomes different, the molecule experiences
different noncovalent interactions with the surrounding and, in
turn, a different balance of intra- and intermolecular forces
crucial to its conformation. The strong impact of environments
on biomolecules already made some researchers wonder how
relevant the learning from conformational studies of proteins
in test tubes is to what the molecules do in natural settings.1
Unlike a dilute solution, some intracellular compartments have
macromolecules reaching 500 g/L in concentration. The con-
formation of proteins and other macromolecules sometimes
differs greatly in these crowded environments from those in
dilute, homogeneous solution. Human amyloidogenic proteins
related to neurodegenerative diseases such as Alzheimer’s and
Parkinson’s diseases, for example, were found to be strongly
affected by cell-like crowded environments.2
Despite the impressive progress made in synthetic foldamers,
the research so far has been focused almost exclusively in
homogeneous solution and in the solid state. Although some
foldamers were prepared to interact with lipid bilayers,6 their
(4) (a) Delsuc, N.; Godde, F.; Kauffmann, B.; Le´ger, J.-M.; Huc, I. J. Am.
Chem. Soc. 2007, 129, 11348–11349. (b) Kolomiets, E.; Berl, V.; Lehn,
J.-M. Chem.sEur. J. 2007, 13, 5466–5479. (c) Bradford, V. J.; Iverson,
B. L. J. Am. Chem. Soc. 2008, 130, 1517–1524. (d) Gong, B. Acc.
Chem. Res. 2008, 41, 1376–1386. (e) Li, X.; Wu, Y.-D.; Yang, D.
Acc. Chem. Res. 2008, 41, 1428–1438. (f) Meudtner, R. M.; Hecht,
S. Angew. Chem., Int. Ed. 2008, 47, 4926–4930. (g) Haldar, D.;
Schmuck, C. Chem. Soc. ReV. 2009, 38, 363–371.
(1) Minton, A. P. J. Cell Sci. 2006, 119, 2863–2869.
(2) Zhou, Z.; Fan, J.-B.; Zhu, H.-L.; Shewmaker, F.; Yan, X.; Chen, X.;
Chen, J.; Xiao, G.-F.; Guo, L.; Liang, Y. J. Biol. Chem. 2009, 284,
30148–30158.
(5) (a) Li, C.; Zhu, Y.-Y.; Yi, H.-P.; Li, C.-Z.; Jiang, X.-K.; Li, Z.-T.;
Yu, Y.-H. Chem.sEur. J. 2007, 13, 9990–9998. (b) Meudtner, R. M.;
Ostermeier, M.; Goddard, R.; Limberg, C.; Hecht, S. Chem.sEur. J.
2007, 13, 9834–9840. (c) Shin, S. B. Y.; Yoo, B.; Todaro, L. J.;
Kirshenbaum, K. J. Am. Chem. Soc. 2007, 129, 3218–3225. (d)
Chowdhury, S.; Schatte, G.; Kraatz, H.-B. Angew. Chem., Int. Ed.
2008, 47, 7056–7059. (e) Kendhale, A. M.; Gonnade, R.; Rajamoha-
nan, P. R.; Hofmann, H.-J.; Sanjayan, G. J. Chem. Commun. 2008,
2541–2543. (f) Kim, U.-I.; Suk, J.-m.; Naidu, V. R.; Jeong, K.-S.
Chem.sEur. J. 2008, 14, 11406–11414. (g) Schafmeister, C. E.;
Brown, Z. Z.; Gupta, S. Acc. Chem. Res. 2008, 41, 1387–1398. (h)
Suk, J.-M.; Jeong, K.-S. J. Am. Chem. Soc. 2008, 130, 11868–11869.
(3) (a) Gellman, S. H. Acc. Chem. Res. 1998, 31, 173–180. (b) Hill, D. J.;
Mio, M. J.; Prince, R. B.; Hughes, T. S.; Moore, J. S. Chem. ReV.
2001, 101, 3893–4012. (c) Hecht, S.; Huc, I. Foldamers: Structure,
Properties, and Applications; Wiley-VCH: Weinheim, Germany, 2007.
(d) Cubberley, M. S.; Iverson, B. L. Curr. Opin. Chem. Biol. 2001, 5,
650–653. (e) Kirshenbaum, K.; Zuckermann, R. N.; Dill, K. A. Curr.
Opin. Struct. Biol. 1999, 9, 530–535. (f) Goodman, C. M.; Choi, S.;
Shandler, S.; DeGrado, W. F. Nat. Chem. Biol. 2007, 3, 252–262. (g)
Bautista, A. D.; Craig, C. J.; Harker, E. A.; Schepartz, A. Curr. Opin.
Chem. Biol. 2007, 11, 685–692. (h) Horne, W. S.; Gellman, S. H.
Acc. Chem. Res. 2008, 41, 1399–1408.
9
9890 J. AM. CHEM. SOC. 2010, 132, 9890–9899
10.1021/ja103694p  2010 American Chemical Society

Environmental Effects Dominate Folding of Oligocholates
A R T I C L E S
Scheme 1. Schematic Representation of the Solvophobic Folding
of an Oligocholate in a Mixture of Polar and Nonpolar Solvents
conformations within membranes generally were not studied
in detail. Lipid membranes are one of the most important
environments for biofoldamers. Membrane proteins perform vital
biological functions, such as photosynthesis, ion conduction,
signal transduction, and immune response, and account for
nearly 50% of all drug targets10 but are notoriously difficult to
study. Biochemists have studied membrane proteins in both
lipid bilayers and surfactant micelles.11 Being amphiphilic
in nature, the latter are considered good mimics of the bilayer
environment and have yielded much insight into how
membrane-associated proteins or peptides might fold in a
membrane-like environment.
In this paper, we report the conformational study of several
functionalized oligocholate foldamers in homogeneous solution,
ionic and nonionic micelles, and lipid bilayers. To the best of
our knowledge, no study has been reported to systematically
investigate how these different environments influence the
conformation of synthetic foldamers. We found that environ-
mental effects can completely overwhelm the inherent foldability
of a foldamer. Micelles are frequently used as a membrane-
like environment to investigate the conformation of membrane-
associated proteins/peptides, but our work indicates that the
conformation of the same foldamers may differ greatly in
micelles and lipid bilayers, or even in different micelles. One
of the most difficult challenges in modern bioorganic chemistry
is selective translocation of hydrophilic molecules across lipid
bilayers. We demonstrate that the oligocholate foldamers are
effective molecular transporters in common phospholipid mem-
branes. The ability to correlate structure and activity and the
straightforward functionalization of the foldamers are important
to the rational design of molecular transporters useful in
applications such as membrane separation and drug delivery.
Results and Discussion
Design and Synthesis of Oligocholates. An oligocholate folds
best in nonpolar solvents containing a small amount of a polar
solvent.12 The polar solvent is needed not only to dissolve the
oligomer but, more importantly, to provide the solvophobic
driving force to the folding (Scheme 1). When the oligocholate
folds into a helix, the rigidity of the steroidal backbone keeps
the internal cavity from collapsing. Polar solvent is microphase-
separated from the bulk solvent mixture and concentrated into
the nanometer-sized hydrophilic cavity. Such an arrangement
is favorable for two main reasons. First, the hydrophilic amide
and hydroxyl groups of the oligocholate are efficiently solvated
by the entrapped polar solvent in a largely nonpolar medium.
Second, the phase-separated polar solvent can reside in a
preferred, polar microenvironment instead of in the bulk,
nonpolar solvent. A unique feature of this folding mechanism
is its strong dependence on both the size and the shape of the
polar and nonpolar solvents, derived partly from the dimension
of a typical solvent molecule being significant in comparison
to the size of the hydrophilic cavity.13
(6) (a) Porter, E. A.; Weisblum, B.; Gellman, S. H. J. Am. Chem. Soc.
2002, 124, 7324–7330. (b) Mowery, B. P.; Lee, S. E.; Kissounko,
D. A.; Epand, R. F.; Epand, R. M.; Weisblum, B.; Stahl, S. S.;
Gellman, S. H. J. Am. Chem. Soc. 2007, 129, 15474–15476. (c)
Chongsiriwatana, N. P.; Patch, J. A.; Czyzewski, A. M.; Dohm, M. T.;
Ivankin, A.; Gidalevitz, D.; Zuckermann, R. N.; Barron, A. E. Proc.
Natl. Acad. Sci. U.S.A. 2008, 105, 2794–2799. (d) Choi, S.; Isaacs,
A.; Clements, D.; Liu, D.; Kim, H.; Scott, R. W.; Winkler, J. D.;
DeGrado, W. F. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 6968–6973.
(e) Tew, G. N.; Scott, R. W.; Klein, M. L.; Degrado, W. F. Acc. Chem.
Res. 2010, 43, 30–39.
Chart 1 shows the structures of the oligocholates used in the
current study. Because the cholate units are connected by amide
linkages, we can introduce intramolecular salt bridges conve-
niently by incorporating arginine and glutamic acid residues in
the sequence. The syntheses of the molecules followed previ-
ously reported procedures12-14 and are detailed in the Support-
ing Information. All oligocholates are labeled with a dansyl and
a naphthyl group at the chain ends, which are used as the
acceptor (A) and donor (D) for fluorescent resonance energy
transfer (FRET).15 In FRET, the energy-transfer efficiency
(7) (a) Rodriguez, J. M.; Hamilton, A. D. Angew. Chem., Int. Ed. 2007,
46, 8614–8617. (b) Wyrembak, P. N.; Hamilton, A. D. J. Am. Chem.
Soc. 2009, 131, 4566–4567. (c) Horne, W. S.; Johnson, L. M.; Ketas,
T. J.; Klasse, P. J.; Lu, M.; Moore, J. P.; Gellman, S. H. Proc. Natl.
Acad. Sci. U.S.A. 2009, 106, 14751–14756. (d) Imamura, Y.; Wa-
tanabe, N.; Umezawa, N.; Iwatsubo, T.; Kato, N.; Tomita, T.; Higuchi,
T. J. Am. Chem. Soc. 2009, 131, 7353–7359.
(8) Cai, W.; Wang, G.-T.; Xu, Y.-X.; Jiang, X.-K.; Li, Z.-T. J. Am. Chem.
Soc. 2008, 130, 6936–6937.
(9) Maayan, G.; Ward, M. D.; Kirshenbaum, K. Proc. Natl. Acad. Sci.
U.S.A. 2009, 106, 13679–13684.
(10) (a) Hopkins, A. L.; Groom, C. R. Nat. ReV. Drug DiscoVery 2002, 1,
727–730. (b) Lundstrom, K. Curr. Protein Pept. Sci. 2006, 7, 465–
470.
6
(E) is related to the D-A distance (r) by equation E ) Ro /
6
(Ro + r6), in which Ro is the Fo¨rster distance for a specific
(11) (a) Jirgensons, B.; Hnilica, L. S. J. Am. Chem. Soc. 1966, 88, 2341–
2342. (b) Luidens, M. K.; Aks, C. S.; Zhu, Q.; Smith, T. F.; MacColl,
R.; Figge, J. Peptide Res. 1993, 6, 134–139. (c) Chorev, M.; Gurrath,
M.; Behar, V.; Mammi, S.; Tonello, A.; Peggion, E. Biopolymers 1995,
36, 473–484. (d) Schibli, D. J.; Hwang, P. M.; Vogel, H. J.
Biochemistry 1999, 38, 16749–16755. (e) Montserret, R.; McLeish,
M. J.; Bockmann, A.; Geourjon, C.; Penin, F. Biochemistry 2000, 39,
8362–8373. (f) Searle, M. S.; Jourdan, M. Bioorg. Med. Chem. Lett.
2000, 10, 1139–1142. (g) Sanghera, N.; Pinheiro, T. J. T. Protein Sci.
2000, 9, 1194–1202. (h) Li, H.; Li, F.; Sun, H.; Qian, Z. M. Biochem.
J. 2003, 372, 757–766. (i) Schievano, E.; Calisti, T.; Menegazzo, I.;
Battistutta, R.; Peggion, E.; Mammi, S.; Palu, G.; Loregian, A.
Biochemistry 2004, 43, 9343–9351. (j) Thundimadathil, J.; Roeske,
R. W.; Guo, L. Biopolymers 2006, 84, 317–328.
D-A pair. Because typical Ro (2-6 nm) is comparable to the
diameter of a protein, FRET is widely used in the conformational
(12) Zhao, Y.; Zhong, Z. J. Am. Chem. Soc. 2005, 127, 17894–17901.
(13) Zhao, Y.; Zhong, Z.; Ryu, E. H. J. Am. Chem. Soc. 2007, 129, 218–
225.
(14) (a) Zhong, Z.; Zhao, Y. Org. Lett. 2007, 9, 2891–2894. (b) Zhao, Y.;
Zhong, Z. J. Am. Chem. Soc. 2006, 128, 9988–9989.
(15) (a) Selvin, P. R. Methods Enzymol. 1995, 246, 300–334. (b) Lakowicz,
J. R. Principles of Fluorescence Spectroscopy, 2nd ed.; Kluwer
Academic/Plenum: New York, 1999. (c) Van Der Meer, B. W.; Coker,
G.; Chen, S. Y. S. Resonance Energy Transfer: Theory and Data;
VCH: New York, 1994.
9
J. AM. CHEM. SOC. VOL. 132, NO. 28, 2010 9891

A R T I C L E S
Cho and Zhao
Chart 1
study of biomolecules.15 Distance-dependent techniques such
as NOE, FRET, and spin coupling16 are extremely powerful in
the characterization of synthetic and natural foldamers.3 FRET
has the benefit of measuring much larger distances than NOE
and spin coupling and was found to work extremely well for
the oligocholates.12-14
Oligocholates 1 and 2 both contain a salt bridge; their
difference lies in the number of cholate units in between the
arginine and glutamic acid. Three cholates make one turn in
the helix in the folded oligocholates;12 thus, the absolute distance
between the arginine and glutamic acid in the folded conformer
is quite similar in 1 and 2. Nonetheless, the salt bridge will be
formed within the first helix turn in 1 and across two turns in
2. If one thinks forming a salt bridge over a longer distance
more effectively brings the chain to the folded conformation
(as it impacts a larger part of the chain), 2 should fold better
than 1. If, however, one considers that formation of the salt
bridge over a longer distance needs to constrain a larger section
of the foldamer chainsan entropically unfavorable processs2
might fold worse than 1. Which is true?
Oligocholate 3 contains two arginines and one glutamic acid
and thus carries a net positive charge. Examination of the CPK
model suggests that the farther arginine is more likely to be
engaged in the salt bridge, as ion-pairing over the large and
rigid steroid backbone of the cholate may introduce significant
strain to the structure. Will the extra arginine impact the
conformation of the molecule? Compound 3 is the only one
with a net charge. Many proteins are charged.17 Electrostatic
interactions operate over much longer distances than most other
noncovalent forces. How will the charge impact the conforma-
tion of the molecule? Will the impact be similar across different
environments? How will the non-participating arginine affect
its ability to protect a hydrophilic guest and transport it across
lipid membranes?
Because the same D-A pair is used in 1-4, the energy-
transfer efficiency directly reflects the distance between the
donor and the acceptor. FRET may be determined either by the
enhancement of the acceptor emission or by the weakening of
the donor emission.15 In our experience, the naphthyl emission
is weak in most solvents used to fold the oligocholates, and
FRET is best determined by monitoring the acceptor emission.
However, because dansyl itself is sensitive to solvent polarity18
and folding of an oligocholate changes the solvent composition
near the fluorophore (Scheme 1), we cannot use the emission
intensity directly to measure the energy-transfer efficiency.
Instead, we need to extract the contribution of FRET (to the
acceptor emission) from the excitation spectrum, obtained when
the acceptor emission at 500 nm was monitored while the
excitation wavelength (λex) was scanned.19 In the absence of
energy transfer, the excitation spectrum of a compound re-
sembles the absorption spectrum of the acceptor. Energy transfer
from the donor is indicated by peaks corresponding to the
donor’s absorption.15
The excitation spectra of the oligocholates differ greatly.
FRET is weak in 1 (Figure 1a) but very strong in 2 (Figure 1b)
in low methanol solutions, evidenced by the large peak at 300
nm from the naphthyl donor in the latter. Location of the salt
bridge thus has a strong effect on the energy transfer. The
contribution from the donor gradually disappears as methanol
is added. Toward the end of solvent titration, the spectra of 1
and 2 look identical, only showing the dansyl absorption near
260 and 340 nm.
Despite the large difference in structure, 3 and 4 have very
similar excitation spectra at the beginning of the solvent titration
with 2% methanol (Figure 1c,d). The difference between the
two lies in their response to methanol. Whereas the donor
absorption at 300 nm in 3 is slowly weakened upon methanol
addition, it quickly disappears in 4 with as little as 6% methanol.
Before we can correlate the FRET data with the folding, we
have to exclude intermolecular aggregation under the experi-
mental conditions, as FRET simply reflects a close D-A dis-
tance and can come from either an intramolecular or an
intermolecular process. Our previous work indicates that non-
ionic oligocholates such as 4 do not aggregate under similar
conditions.12 Oligocholates 1-3, however, contain highly polar,
ionic groups. Figure 2a shows the excitation spectra of 3 in 2:1
hexane/ethyl acetate with 2% methanol over 0.13-2.0 µM. This
is the most stringent test for aggregation, using the most polar
oligocholate in the least polar solvent mixture. Nevertheless,
the shape of the spectrum stays unchanged during the 16-fold
dilution, and the normalized excitation spectra completely
overlap with one another (Figure 1S, Supporting Information).
As a control, 4 is the only foldamer made entirely of cholates
(other than the fluorophores). Without any salt bridge, its
inherent foldability should be the lowest among the four. Will
it be the worst folder always?
Folding of the Oligocholates in Homogeneous Solution. To
answer the above questions, we first studied the molecules in
2:1 hexane/ethyl acetate with different amounts of methanol,
one of the most “folding-friendly” solvent systems identified
for the oligocholates.13 Methanol is completely miscible with
ethyl acetate but nearly immiscible with hexane. A high
percentage of hexane reduces the energetic cost in the phase
separation of methanol from the bulk to the hydrophilic cavity
of folded helix (Scheme 1) and therefore is beneficial to the
folding.
(16) Matsuda, K.; Stone, M. T.; Moore, J. S. J. Am. Chem. Soc. 2002,
124, 11836–11837.
(18) Li, Y.-H.; Chan, L.-M.; Tyer, L.; Moody, R. T.; Himel, C. M.;
Hercules, D. M. J. Am. Chem. Soc. 1975, 97, 3118–3126.
(19) Zhao, Y. J. Org. Chem. 2009, 74, 834–843.
(17) Gitlin, I.; Carbeck, J. D.; Whitesides, G. M. Angew. Chem., Int. Ed.
2006, 45, 3022–3060.
9
9892 J. AM. CHEM. SOC. VOL. 132, NO. 28, 2010

Environmental Effects Dominate Folding of Oligocholates
A R T I C L E S
Figure 1. Normalized excitation spectra of (a) 1, (b) 2, (c) 3, and (d) 4 in 2-30% MeOH in 2:1 hexane/ethyl acetate; [oligomer] ) 2.0 µM. The acceptor
emission at 500 nm was monitored. The intensity at 340 nm (the λmax of dansyl in the absorption spectrum) was set to 1.
Figure 2. (a) Excitation spectra of 3 in 2% MeOH in 2:1 hexane/ethyl acetate. The concentration of 3 was 2.0, 1.0, 0.5, 0.25, and 0.13 µM from top to
bottom. (b) F300/F340 of 1 (4), 2 (0), and 3 (]) as a function of the oligocholate concentration. The acceptor emission at 500 nm was monitored. F300 and
F340 represent the emissive intensity of dansyl at 500 nm in the excitation spectrum when λex was 300 and 340 nm, respectively.
Thus, the FRET observed for the compound is independent of
concentration, strongly supporting its intramolecular origin. The
same is true for compounds 1 and 2 (Figures 2S and 3S,
Supporting Information). Because the maximum absorption of
naphthyl and dansyl occurs at 300 and 340 nm, respectively,
we can use F300/F340 in the excitation spectrum as an indicator
for the FRET. A higher F300/F340 is equivalent to a larger
contribution from the donor and translates to a shorter D-A
distance. As shown by Figure 2b, F300/F340 is completely
independent of oligomer concentration, suggesting that FRET
comes from folding instead of aggregation in all these salt-
bridged oligocholates.
turation experiments for proteins20 and is often described by a
two-state model (eq 1).
Keq
folded y z unfolded
(1)
In a two-state transition, only the fully folded and fully unfolded
states exist at any given solvent composition. Such a confor-
mational change is characterized by a sigmoidal titration curve.
As shown by Figure 3a, the F300/F340 curves for 1-4 are all
sigmoidal in shape and fit well to the two-state model. (Details
of the data analysis are given in the Supporting Information.)
The data fitting allows us to calculate the folding equilibrium
Efficient FRET itself does not mean a compound is a good
folder. The information about the foldability of a compound
may be extracted from the change of FRET in response to
solvent composition. This is very similar to the solvent-dena-
(20) (a) Chan, H. S.; Bromberg, S.; Dill, K. A. Philos. Trans. R. Soc.
London B 1995, 348, 61–70. (b) Creighton, T. E. Protein Structure:
A Practical Approach, 2nd ed.; IRL Press: Oxford; 1997.
9
J. AM. CHEM. SOC. VOL. 132, NO. 28, 2010 9893

A R T I C L E S
Cho and Zhao
Figure 3. (a) F300/F340, (b) fraction of the unfolded conformer, and (c) unfolding free energies for 1 (4), 2 (0), 3 (]), and 4 (O) as a function of volume
percentage of methanol in 2:1 hexane/ethyl acetate. The theoretical curves are nonlinear least-squares fitting to a two-state transition model in (a) and (b).
Details for the curve-fitting are found in the Supporting Information; [oligomer] ) 2.0 µM.
Table 1. Values of ∆G0 and m Determined from Solvent Denaturation Curvesa
entry
compound
solvent composition
∆G0 (kcal/mol)
m (kcal/mol)
1
2
3
4
5
6
7
8
1
2
3
4
1
2
3
4
MeOH in 2:1 hexane/ethyl acetate
MeOH in 2:1 hexane/ethyl acetate
MeOH in 2:1 hexane/ethyl acetate
MeOH in 2:1 hexane/ethyl acetate
MeOH in ethyl acetate
MeOH in ethyl acetate
MeOH in ethyl acetate
MeOH in ethyl acetate
3.5 ( 0.5 (3.1)
2.1 ( 0.2 (2.0)
2.8 ( 0.3 (2.6)
1.7 ( 0.5 (1.7)
2.3 ( 0.4 (2.1)
0.6 ( 0.5 (0.5)
3.0 ( 0.4 (2.5)
0.42 ( 0.06 (0.35)
0.34 ( 0.03 (0.32)
0.20 ( 0.02 (0.18)
0.45 ( 0.09 (0.45)
0.40 ( 0.05 (0.36)
0.29 ( 0.05 (0.26)
0.19 ( 0.02 (0.16)
b
b
-
-
a ∆G0 is the unfolding free energy of the compound in the absence of the methanol, and m represents the sensitivity of the unfolding free energy to
the concentration of methanol in percentages. Data with errors are determined by nonlinear least-squares fitting to a two-state transition model. Data in
parentheses are determined by linear fitting of the unfolding free energies as a function of denaturant concentration. See the Supporting Information for
details. b Compound 4 was completely unfolded in methanol/ethyl acetate mixtures, and the unfolding free energies cannot be determined.
as a function of methanol percentage (Figure 3b). Another
hallmark of the two-sate transition is a linear dependence of
free energies on the denaturant concentration, i.e., ∆G ) ∆G0
- m[MeOH], where ∆G0 is the unfolding free energy of the
compound in the absence of the denaturant and m represents
the sensitivity of the free energy to the concentration of the
denaturant. As shown by Figure 3c, all four compounds display
such a linear relationship.
The stability of the folded conformer is reflected by the
amount of denaturant needed to induce 50% unfolding. The
value is 8.4, 6.2, 14.3, and 3.7% methanol for oligocholates
1-4, respectively (Figure 3b). Thus, although 2 has by far the
most efficient FRET among the four, it is actually less stable
than either 1 or 3. Clearly, the location of the salt bridge impacts
the folding strongly. The non-participating arginine must be
important as well, since 3 has the same number of salt bridges
as 1 and 2 but folds much better.
Table 1 summarizes the thermodynamic folding data obtained
from the curving-fitting. These numbers provide a quantitative
picture of how the folding-unfolding equilibrium is affected
by the solvent composition. For the oligocholates, methanol
plays dual roles in the folding-unfolding equilibrium. On one
hand, the polar solvent is needed to dissolve the oligocholate
and provide the preferential solvation of the hydrophilic faces
of the cholates. On the other hand, when located outside the
hydrophilic cavity and, in particular, present at higher concen-
trations, methanol can unfold or “denature” the helix. Thus, ∆G0
is the unfolding free energy of the oligocholate predicted by
the preferential solvation model extrapolated to 0% methanol.
The number is purely theoretical in nature because the presence
of methanol is necessary for the preferential solvation to operate.
As expected, without any salt bridge, 4 has the lowest ∆G0
of 1.7 kcal/mol (Table 1, entry 4). It is somewhat surprising
that the guanidinium-carboxylate salt bridge does not bring
much stabilization to 2, as its ∆G0 is only 0.3-0.4 kcal/mol
larger than that of 4. The small increase of ∆G0 does not mean
that the salt bridge is weak in the folded 2. The association
constant between guanidinium and carboxylate in hexane/ethyl
acetate is unknown (and cannot be determined because of
insolubility) but must be extremely large in such a nonpolar
solvent mixture.21 If one assumes that the salt bridge is formed
in both the folded and unfolded states, a small contribution to
∆G0 simply means that the salt bridge does not help the folded
state relative to the unfolded. There are four cholate groups
between the salt bridge and multiple flexible methylene groups
connecting the guanidinium and carboxylate to the foldamer
backbone in 2. If the salt bridge can be accommodated equally
well by the folded and unfolded conformers, its formation will
not make much difference to the folding-unfolding equilibrium;
this is likely to be the case when the salt bridge is separated by
a long, relatively flexible segment of a chain. If, however, the
salt bridge is favored by either the folded or unfolded conformer,
whether due to geometrical constraint or something else, its
formation will help shift the equilibrium to that particular
conformer.
Consistent with the above explanation, the unfolding free
energy increases monotonically as the number of cholate units
is reduced between the arginine and glutamic acid. For 3 and
1, whose salt bridge is separated by three and two cholates,
∆G0 ) 2.8 and 3.5 kcal/mol, respectively. It appears that, the
closer the salt bridge, the more it can constrain the chain to the
folded conformation and the less able is the chain to adopt an
(21) (a) Schmidtchen, F. P.; Berger, M. Chem. ReV. 1997, 97, 1609–1646.
(b) Best, M. D.; Tobey, S. L.; Anslyn, E. V. Coord. Chem. ReV. 2003,
240, 3–15.
9
9894 J. AM. CHEM. SOC. VOL. 132, NO. 28, 2010

Environmental Effects Dominate Folding of Oligocholates
A R T I C L E S
Figure 4. (a) F300/F340, (b) fraction of the unfolded conformer, and (c) unfolding free energies for 1 (4), 2 (0), 3 (]), and 4 (O) as a function of volume
percentage of methanol in ethyl acetate. The theoretical curves are nonlinear least-squares fits to a two-state transition model in (a) and (b). [Oligomer] )
2.0 µM.
extended conformation. The largest difference in ∆G0 in these
oligocholates is between 1 and 4, amounting to 1.4-1.8 kcal/
mol, depending on the (linear or nonlinear) method of curving-
fitting.
in m (0.40 vs 0.29) is more obvious for 1 and 2 than in the
ternary solvents (0.42 vs 0.34). The data once again support
that methanol is more detrimental to the folded conformer that
benefits more from the salt bridge. On the other hand, the solvent
change makes no difference to oligocholate 3sits ∆G0 and m
are essentially the same with and without hexane. Thus, the
extra arginine makes the foldamer more resistant to adverse
solvent conditions, whether the addition of methanol or the
elimination of hexane.
Parameter m indicates how sensitive the unfolding reaction
is to methanol percentage. The most sensitive foldamer among
the four is nonionic 4, with each percent methanol causing 0.45
kcal/mol of change in the unfolding free energy (Table 1, entry
4). The salt bridge does make the foldamer more resistant to
methanol denaturation. The difference between 1 and 2 is not
large, but m is slightly larger for 1 according to both linear and
nonlinear curving-fitting. The trend is conceptually reasonable,
as an increase in methanol not only could diminish the
preferential solvation responsible for folding the oligocholate
but also weaken the salt bridge. The more the folded conformer
benefits from the salt bridge (in ∆G0), the more sensitive the
folding/unfolding equilibrium should be toward methanol.
Foldamer 3 is the least sensitive to methanol among the four.
Hence, the non-participating arginine is quite important to the
folding equilibrium. According to the preferential solvation
model (Scheme 1), the concentration of methanol in the entire
mixture is highest inside the folded helix. We previously utilized
this nanosized pool of polar solvent for size-selective catalysis.22
This local methanol undoubtedly can weaken the hydrogen-
bonded salt bridge but, in the meantime, can better solvate the
non-participating guanidinium group, a highly favorable process
in the largely nonpolar mixture. Thus, although the extra arginine
does not contribute directly to the inherent foldability of the
oligocholate (compare ∆G0 of 1 and 3, for example), it dampens
the adverse impact of methanol by exploiting the localized
methanol in the folded conformer.
To further confirm the above conclusions, we performed the
solvent titration in methanol/ethyl acetate mixtures. Without
hexane, folding is expected to be more difficult due to the higher
cost in phase-separating methanol. Indeed, the non-salt-bridged
4 completely loses its ability to fold in the binary mixture (Figure
4a). The data for 1-3 continue to fit well to the two-state model,
yielding the folding equilibria (Figure 4b) and unfolding free
energies (Figure 4c). Interestingly, the ∆G0 for 1 and 2 is
decreased by 1-1.5 kcal/mol by eliminating hexane, while m
is hardly affected (Table 1, compare entries 5 and 6 with entries
1and 2). It seems that solvent miscibility, which is the core of
the preferential solvation model, is more important to the
inherent foldability. In a more challenging solvent, the difference
Folding of the Oligocholates in Surfactant Micelles. Domi-
nated by hydrophobic groups, the oligocholates can be solubi-
lized by surfactant micelles in water. Parent oligocholates such
as 4 were found to fold well in sodium dodecyl sulfate (SDS)
micelles.23 The folding, however, follows a mechanism com-
pletely different from that in solution. The SDS micelle (ca. 3
nm in diameter) can accommodate the folded conformer much
better than the unfolded. Each cholate unit is 1.4 nm from head
to tail. According to CPK models, a fully folded cholate hexamer
is less than 2 nm in dimension, but an unfolded conformer can
stretch to several nanometers in length, depending on the exact
conformation. Because the SDS micelles prefer to maintain the
spherical shape in water, solubilizing an oligocholate within
the micelle is analogous to pushing a snake into a small cages
the result is that the snake (oligocholate) has no choice but to
coil up (fold).
How will the salt bridge and charge impact the folding in
micelles? Once again, we recorded the excitation spectra of these
compounds in 1-70 mM SDS. The F300/F340 value is 0.75, 1.54,
and 1.14 for compounds 1-3, respectively, in 1 mM SDS. These
numbers match quite well with those of the corresponding folded
conformers (F300/F340 ) 0.78, 2.01, and 1.04) in 2% methanol
in 2:1 hexane/ethyl acetate, suggesting that the oligocholates
are probably folded in low SDS solutions (vide infra). Long
oligocholates such as 4 prefer folding instead of aggregation in
the presence of SDS because the rigid and awkwardly shaped
steroid backbone prevents their tight/stable packing.23b Assum-
ing that the FRET in 1 mM SDS solutions comes from folding,
the above data suggest that a significant population of unfolded
conformer exists in 2 in 1 mM SDS.
Figure 5a plots F300/F340 of 1-4 as a function of SDS
concentration. With an increase in SDS, FRET becomes less
efficient and drops most dramatically over 4-10 mM for 1-3,
(22) Cho, H. K.; Zhong, Z. Q.; Zhao, Y. Tetrahedron 2009, 65, 7311–
(23) (a) Zhong, Z.; Zhao, Y. J. Org. Chem. 2008, 73, 5498–5505. (b) Zhao,
Y. J. Org. Chem. 2009, 74, 7470–7480.
7316.
9
J. AM. CHEM. SOC. VOL. 132, NO. 28, 2010 9895

A R T I C L E S
Cho and Zhao
Figure 5. (a) F300/F340 and (b) maximum emission wavelength (λem) of the dansyl in 1 (4), 2 (0), 3 (]), and 4 (O) as a function of SDS concentration. The
data for 4 (O) were taken from ref 23b; [oligomer] ) 2.0 µM; λex ) 340 nm. The data points are connected to guide the eye.
as SDS begins to form micelles (cmc ) 8 mM).24 The behavior
is very different for nonionic 4 (O), known to be folded below
and above the critical micelle concentration (cmc).23a It seems
that micellization of SDS is very detrimental to the salt-bridged
foldamer. The most likely cause for this effect is the sulfate
headgroup of the surfactant, which can compete with carboxylate
for the salt bridge. Below the cmc, although ionic surfactants
such as SDS can assemble around large hydrophobic mol-
ecules25 such as oligocholates, the number of surfactants around
the foldamers must be limited. Upon micellization, a large
number of surfactants assemble cooperatively to form a “com-
munity”. The main difference between the micelle and the
individual surfactants is the much stronger solubilizing power
of the former for hydrophobic guests and the higher local
concentration of ionic headgroups. For nonionic 4, micellization
makes no difference because the molecule is overall hydrophobic
and prefers to reside within the core of the micelle. The size of
the micelle and the strong preference of the SDS micelle to
maintain its spherical shape constrain the chain into the folded
conformation.23 For the salt-bridged 1-3, two factors could
disfavor the folded form. First, the high concentration of sulfate
headgroups in the micelle can easily break the salt bridge. The
solution studies above indicate that all three foldamers benefit
to some extent from the salt bridge. Second, ionic groups such
as guanidinium and carboxylate have a strong demand for
solvation in water. Below the cmc, even though SDS molecules
may assemble around the folded oligocholates, the assembly
may be loose enough to allow water molecules to approach the
ionic groups. Upon micellization, the oligocholate suddenly finds
itself in a hydrophobic microenvironment. If solvation of the
ionic groups is important to the overall system, the helix
probably prefers to unfold and migrate to the surface of the
micelle, where the ionic groups can be exposed.
molecules. Micellization causes a step change in the emission
wavelength of dansyl, jumping from <500 to 530-540 nm. The
data might be surprising given that micellization creates a
hydrophobic microenvironment around the oligocholate, but they
are quite reasonable if the oligocholate unfolds and migrates to
the micellar surface. Consistent with the folding of 4, its λem,
although fluctuating to some extent, shows the dansyl being
located in similar microenvironments below and above the cmc
of SDS.
Because of the interference of the sulfate headgroup of SDS
with the salt-bridge formation, we also studied the conformation
of these compounds in Brij 35 solutions. The nonionic surfactant
has the same dodecyl chain as SDS but an oligo(ethylene glycol)
headgroup instead of sulfate. It is commonly used to solubilize
membrane proteins,26 making it particularly interesting to
understand its effect on the oligocholates.
Figure 6a shows F300/F340 of 1-4 in Brij 35. The oligocholates
clearly behave differently in the presence of the nonionic
surfactant. The F300/F340 ratio for 1, for example, is completely
independent of the surfactant concentration, averaging 0.79 over
0.025-20 mM of Brij 35. The value is almost identical to that
of the folded 1 in solution (0.78, Figure 3a), suggesting that
the energy transfer comes from folding. Because the oligocholate
is insoluble in water, the nonionic surfactant is basically the
medium for the foldamer. If FRET stays unchanged over 8000-
fold dilution (from 0.0025 to 20 mM of surfactant), there is
very little likelihood that it could come from intermolecular
aggregation. The cmc of Brij 35 is about 0.1 mM.27 Hence,
micellization of the nonionic surfactant does not unfold com-
pound 1.
The starting F300/F340 for oligocholate 2 is 2.01 (Figure 6a,
0), exactly the same as that of the folded conformer in solution
(Figure 3a, 0). Thus, the salt-bridged foldamer is fully folded
in low Brij solutions. Its stability seems to be slightly lower
than that of 1, as the F300/F340 ratio does show some decrease
above the cmc of Brij 35 but still averages at about 1.6. This
The maximum emission wavelength (λem) of dansyl is
sensitive to its microenvironment.18 Figure 5b plots the λem of
1-4 in different SDS solutions. In agreement with the
folding-unfolding transition proposed above, all three salt-
bridged foldamers emit at shorter wavelength below the cmc.
The data indicate that the dansyl of these oligocholates is located
in a relatively nonpolar microenvironmentsthis should cor-
respond to the folded conformer surrounded by some SDS
value is significantly higher than that for the unfolded (F300
/
F340 ) 0.48, Figure 3a) and even larger than the beginning value
in SDS solutions (F300/F340 ) 1.54, Figure 5a). Apparently,
(26) (a) Yoshikawa, S.; Tera, T.; Takahashi, Y.; Tsukihara, T.; Caughey,
W. S. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 1354–1358. (b)
Shinzawa-Itoh, K.; Ueda, H.; Yoshikawa, S.; Aoyama, H.; Yamashita,
E.; Tsukihara, T. J. Mol. Biol. 1995, 246, 572–575. (c) Berger, B. W.;
Garcia, R. Y.; Lenhoff, A. M.; Kaler, E. W.; Robinson, C. R. Biophys.
J. 2005, 89, 452–464.
(24) Rosen, M. J. Surfactants and Interfacial Phenomena, 2nd ed.; Wiley:
New York, 1989; Chapter 3.
(25) Hydrophobic polymers are known to be highly effective at inducing
the aggregation of ionic surfactants below the cmc. See: Rosen, M. J.
Surfactants and Interfacial Phenomena, 2nd ed.; Wiley: New York,
1989; p 181.
(27) (a) Wong, J. E.; Duchscherer, T. M.; Pietraru, G.; Cramb, D. T.
Langmuir 1999, 15, 6181–6186. (b) Tran, C. D.; Yu, S. J. Colloid
Interface Sci. 2005, 283, 613–618.
9
9896 J. AM. CHEM. SOC. VOL. 132, NO. 28, 2010

Environmental Effects Dominate Folding of Oligocholates
A R T I C L E S
Figure 6. (a) F300/F340 and (b) maximum emission wavelength (λem) of dansyl 1 (4), 2 (0), 3 (]), and 4 (O) as a function of Brij 35 concentration.
[Oligomer] ) 2.0 µM; λex ) 340 nm. The data points are connected to guide the eye.
without the sulfate headgroup, the nonionic surfactant represents
a much more folding-friendly environment to the salt-bridged
oligocholates.
work, described a cholate-based “molecular umbrella” for
molecular transport across lipid bilayers.31 In this section, we
report the conformational study of the oligocholates in mem-
branes and our efforts in converting the oligocholates into useful
molecular transporters.
Our initial assumption was that the salt-bridge would continue
to be important to the conformation of oligocholates 1-3 in
lipid membranes. Although the guanidinium-carboxylate in-
teraction is weak in water, its association constant can be as
high as 106 M-1 at the air-water or lipid-water interface.32
None of these oligocholates have any significant solubility in
water. Thus, any salt-bridge interactions for the oligocholates
must happen in or on the surface of lipid bilayers.
We employed the detergent dialysis method to incorporate
the oligocholates into lipid bilayers.33 This procedure is often
employed to reconstitute membrane proteins into liposomes.34
Briefly, we first solubilized the oligocholates in mixed micelles
formed between Brij 35 and dilauroylphosphatidylcholine
(DLPC). DLPC was chosen because it has the same C12 chains
as SDS and Brij 35; we reasoned that the dimension of the
hydrophobic environment should be important to the folding.
As Brij 35 was removed from the mixed micelles by hydro-
phobic Bio-Beads, the remaining DLPC spontaneously forms
liposomes with the oligocholates embedded in the membranes.
Table 2 summarizes the F300/F340 ratio of these compounds
in DLPC bilayers. To understand whether the observed FRET
comes from folding or intermolecular aggregation, we varied
the ratio between the oligocholate and DLPC from 1/50 to
1/1000. If folding is the dominant contribution to the observed
The charged oligocholate 3 is the only one being unfolded
upon micellization. The F300/F340 curve shows a distinct drop
from 1.20 to about 0.6 around the cmc of Brij 35. When the
three salt-bridged foldamers are compared, the effect of micel-
lization on the conformational stability follows the order of 3 >
2 > 1. The order can be rationalized by the increase in envi-
ronmental hydrophobicity during micellization mentioned ear-
lier. Compound 3 has a net positive charge and thus the strongest
need for water solvation. Keeping it folded inside the micelle
will be most difficult. Both 1 and 2 are overall neutral, but the
former has higher inherent foldability according to the studies
in homogeneous solutions (Table 1).
The F300/F340 curve for 4 is completely flat (Figure 6a, O).
The concentration-independent FRET is strong evidence for the
folding of the oligocholate. The average F300/F340 (1.17) is
significantly lower than that (1.68) in the SDS micelles. It is
not very clear why the nonionic oligocholate folds worse in
the Brij micelles, but the result is still in line with the
environmental hydrophobicity. Folding of 4 creates a hydrophilic
cavity that prefers to be filled with water. It is clearly easier to
put water molecules inside SDS instead of Brij micelles.
Nonionic micelles are “drier” than ionic micelles in the inte-
rior.28 Water is actually known to penetrate into SDS micelles
appreciably.29
Folding of Oligocholates in DLPC Bilayers and Transloca-
tion of Carboxyfluorescein across Lipid Membranes. Although
surfactant micelles are frequently used to study membrane-
associated proteins and peptides,11 lipid bilayers differ signifi-
cantly from surfactant micelles. They are liquid crystalline,
anisotropic, and much more hydrophobic than most micelles.
Because of the steroidal backbone, cholate derivatives are
“naturally designed” to be compatible with lipids. Cholate
derivatives, for example, are used in nature in fat digestion and
lipid transport.30 Regan and co-workers, in a series of seminal
(31) For other cholate-based molecular transporters, see: (a) Janout, V.;
Di Giorgio, C.; Regen, S. L. J. Am. Chem. Soc. 2000, 122, 2671–
2672. (b) Janout, V.; Staina, I. V.; Bandyopadhyay, P.; Regen, S. L.
J. Am. Chem. Soc. 2001, 123, 9926–9927. (c) Janout, V.; Jing, B. W.;
Staina, I. V.; Regen, S. L. J. Am. Chem. Soc. 2003, 125, 4436–4437.
(d) Janout, V.; Regen, S. L. J. Am. Chem. Soc. 2005, 127, 22–23. (e)
Mehiri, M.; Chen, W.-H.; Janout, V.; Regen, S. L. J. Am. Chem. Soc.
2009, 131, 1338–1339.
(32) Ariga, K.; Kunitake, T. Acc. Chem. Res. 1998, 31, 371–378.
(33) We initially employed the membrane extrusion method to prepare
oligocholate-containing liposomes. The F300/F340 ratio in some oligo-
cholates was higher at lower [oligocholate]/[lipid] ratio. This result
was contrary to what intermolecular aggregation would produce and
had been thought to support the folding of the oligocholates in the
DLPC bilayers. Later, we found that, at high [oligocholate]/[lipid] ratio,
the DPLC bilayers did not have the capacity to incorporate all
oligocholate molecules into the bilayers, and the oligocholates were
removed by the polycarbonate membrane used during extrusion. The
detergent dialysis method had no such problems. .
(28) Kano, K.; Ueno, Y.; Hashimoto, S. J. Phys. Chem. 1985, 89, 3161–
3166.
(29) SDS micelles do contain an appreciable amount of water. See: (a)
Menger, F. M. Acc. Chem. Res. 1979, 12, 111–117. (b) Martens, F. M.;
Verhoeven, J. W. J. Phys. Chem. 1981, 85, 1773–1777. (c) Turro,
N. J.; Okubo, T. J. Am. Chem. Soc. 1981, 103, 7224–7228. (d)
Fadnavis, N.; Engberts, J. B. F. N. J. Org. Chem. 1982, 47, 152–154.
(e) Szajdzinska-Pietek, E.; Maldonado, R.; Kevan, L.; Jones, R. R. M.
J. Am. Chem. Soc. 1984, 106, 4675–4678.
(30) Danielsson, H.; Sjo¨vall, J. Sterols and Bile Acids; Elsevier: Amsterdam,
1985.
(34) Smith, S. A.; Morrissey, J. H. J. Thromb. Haemost. 2004, 2, 1155–
1162.
9
J. AM. CHEM. SOC. VOL. 132, NO. 28, 2010 9897

A R T I C L E S
Cho and Zhao
Table 2. F300/F340 Values for 1-4 in DLPC Bilayersa
F300/F340
[oligocholate]/[lipid] )
1/50
[oligocholate]/[lipid] )
1/200
[oligocholate]/[lipid] )
1/1000
compound
1
2
3
4
0.62
1.29
0.68
1.07
0.43
0.52
0.43
0.85
0.49
0.59
0.46
0.47
a Determined from the excitation spectra of the liposomal solutions.
FRET, F300/F340 should be independent of oligocholate concen-
tration in the lipid bilayer. If, on the other hand, aggregation is
the dominant process, F300/F340 should decrease upon lowering
the [oligocholate]/[lipid] ratio.
Figure 7. Percent leakage of CF from LUVs for 1 (4), 2 (0), 3 (]), and
4 (O) as a function of oligocholate concentration.
Unfortunately, all compounds seem to aggregate at high
concentrations but dissociate and unfold at low concentrations
in DLPC bilayers. FRET becomes less efficient in all oligo-
cholates at lower [oligocholate]/[lipid] ratio. The spectra for
[oligocholate]/[lipid] ) 1/1000 were noisier than others due to
the extremely low concentration of the fluorophores; thus, the
small increase from 1/200 to 1/1000 in some F300/F340 could
very well be an artifact. The 1/1000 liposomes represent the
least likely environment for aggregation (and thus most likely
for folding), yet F300/F340 ranged from 0.46-0.59, clearly
corresponding to the unfolded structures.
It is not difficult to understand why the oligocholates do not
want to fold in DLPC bilayers. From the conformational studies
in micelles, we know that both salt-bridge-breaking groups and
high environmental hydrophobicity are detrimental to the
folding. A phospholipid bilayer certainly represents the worst
combination. The surface of the membrane is densely packed
with phosphate groups, which are very effective at hydrogen-
bonding with the guanidinium group.21 The interior of a lipid
bilayer is much more hydrophobic than that of a micelle.
Solvation of the introverted polar groups (hydroxyl and amide
on the cholate and the amino acid side chains) of the folded
helix will be significantly more difficult in a lipid bilayer as a
result.
Undeterred by the apparent failure, we decided to investigate
the potential of these oligocholates as molecular transporters.
Even though the FRET studies indicate that the molecules are
unfolded in the lipid bilayers, they do not mean that the
oligocholates cannot fold in a transient fashion. The folded helix
(Scheme 1) has a hydrophobic exterior and a nanometer-sized
hydrophilic cavity, perfect for shielding a hydrophilic guest and
escorting it across a hydrophobic barrier. This is a highly
attractive application, as selectively transporting hydrophilic
molecules across cell membranes is extremely important to
applications such as drug delivery.
To our delight, the oligocholates are quite effective transport-
ers of CF, and the leakage is strongly affected by their
functionality. The compound with the extra arginine (3) is the
best transporter among the four, causing nearly 90% of the
entrapped CF to leak out at 0.25 µM concentration (Figure 7).
The strong transporting ability probably derives from its ability
to bind CF, which contains two carboxylic acid groups. Binding,
however, is not absolutely necessary, as the nonionic 4 shows
about half of the activity. Presumably, the oligocholate can fold
into the helix and the hydrophilic cavity is large enough to
accommodate CF. The oligocholate does not have to be folded
permanently for the transport; as long as the molecule can fold
and migrate across the membrane, it should be able to transport
hydrophilic guests. According to the CPK model, a fully folded
cholate hexamer is about 1.5 nm in length. The hydrophobic
thickness of the POPC bilayer is about 2.6 nm.36 The nearly
linear relationship between the efflux and the oligocholate
concentration suggests that 4 probably acts as a molecular
carrier. It should be mentioned that foldability of the oligocho-
lates was important to the transport, as methyl cholate and
shorter oligomers (e.g., dimer and trimer) displayed no activity
at all (data not shown). Other mechanisms such as liposome
fusion were excluded by lipid mixing assays (Figure 4S,
Supporting Information).
The most striking difference in Figure 7 is between 1 and 2.
Although compound 1 displays transport comparable to that of
the non-salt-bridged 4, 2 is completely incompetent. The
conformational studies in homogeneous solution indicate that
2 has a lower inherent foldability than 1. The micellar studies
also suggest that 2 is more susceptible to salt-bridge-breaking
groups and environmental hydrophobicity. The difference must
be large enough to devoid 2 of its ability to fold even transiently
within the membrane.
The membrane transport was based on the well-established
carboxyfluorescein (CF) leakage assay.35 CF is a water-soluble
dye that displays self-quenching at high concentrations. Large
unilamellar vesicles (LUVs) consisting of POPC/POPG lipids
were prepared by standard methods in the presence of 50 mM
CF. DPLC could not be used in this experiment because the
liposomes were too unstable to trap CF. The extravesicular CF
was removed by gel filtration. Leakage was monitored by the
increase of CF fluorescence upon injection of the oligocholates
in DMSO. The efflux was followed for 60 min, at which the
LUVs were lysed by the addition of Triton X-100.
Conclusions
Although it is convenient to study the conformation of
proteins and peptides in homogeneous solution and/or micelles,
great care must be taken to directly extrapolate the results
obtained in one environment to another. Such a treatment
assumes that the inherent foldability of foldamers stays the same
across different environments, which clearly is not the case for
the oligocholates. Environmental effects can easily overwhelm
the inherent foldability of the molecules. As the surrounding
of the molecule is changed and different noncovalent forces
(35) New, R. R. C. Liposomes: A Practical Approach; Oxford University
Press: Oxford, 1990; pp 131-134.
(36) (a) Lewis, B. A.; Engelman, D. M. J. Mol. Biol. 1983, 166, 203–210.
(b) Nezil, F. A.; Bloom, M. Biophys. J. 1992, 61, 1176–1183.
9
9898 J. AM. CHEM. SOC. VOL. 132, NO. 28, 2010

Environmental Effects Dominate Folding of Oligocholates
A R T I C L E S
are involved in the conformational control, the best folder (i.e.,
3) could become the worst, and the worst (i.e., 4) could become
the best.
The FRET studies indicate that the molecules aggregate at higher
concentrations (e.g., at [oligocholate]/[lipid] ) 1/50) and the
aggregates dissociate at low concentrations (e.g., at [oligocho-
late]/[lipid] ) 1/200-1/1000). It is not exactly clear what the
molecules do on or in the membrane. The FRET shows that
the unfolded conformers dominate. The molecules probably are
embedded to a certain extent in the membrane, considering their
strong hydrophobicity. The molecules, however, can readily fold
and migrate to the other side of the membrane, evidenced by
their ability to transport large water-soluble guests across lipid
bilayers. The transport efficiency depends on both the foldability
and the binding with the guest. The modular synthesis of the
oligocholates, the predictability of their conformational behavior,
and the ease of their functionalization by natural R-amino acids
potentially will make them very useful, selective molecular
transporters.
Folding of the oligocholates in homogeneous solutions is
dominated by the preferential solvation. This mechanism is an
interesting variation of solvophobic interactions. In conventional
solvophobic interactions, poor solvation of a molecular surface
(e.g., hydrocarbon by water) causes the molecule to aggregate
to minimize solvent exposure. Although the same is true for
the hydrophilic faces of the cholate groups in 1-4, the
hydrophilic faces cannot aggregate directly due to the rigidity
of the backbone. Instead, the phase-separated methanol acts as
a bridge to hydrogen-bond with the hydroxyl and amide groups.
It is interesting that even a non-participating arginine can
strongly influence the conformation of the molecules.
As the molecules move into micelles, preferential solvation
by different solvents is no longer available, but the need for
the amphiphilic molecules to interact with appropriate environ-
ments stays the same. The solvation of ionic groups by water
is a strong factor controlling the molecular conformation, and
the salt bridge continues to be important. The folded helix
demands its interior to be solvated by a pool of water molecules.
Folding, as a result, becomes more difficult as the oligocholates
move from ionic SDS micelles to nonionic Brij 35 micelles.
Although the zwitterionic 1 and 2 can still fold in the latter, the
cationic 3 lost its ability to fold in the more hydrophobic
microenvironment.
Acknowledgment. This paper is dedicated to Prof. Joseph B.
Lambert for his 70th birthday. We thank the U.S. Department of
Energy, Office of Basic Energy Sciences (grant DE-SC0002142),
for supporting the research.
Supporting Information Available: Experimental Section,
including the general experimental details, synthesis and
characterization of the compounds, data analysis, and fluores-
cence and lipid-mixing data. This material is available free of
charge via the Internet at http://pubs.acs.org.
The phosphate groups and hydrophobicity of the lipid bilayer
make it the most challenging environment for the oligocholates.
JA103694P
9
J. AM. CHEM. SOC. VOL. 132, NO. 28, 2010 9899