Dynamic assessment of bilayer thickness by varying phospholipid and hydraphile synthetic channel chain lengths

Article abstract of DOI:10.1021/ja044936+

A library of "hydraphile" synthetic ion channel analogues that differ in overall length from ～28-58 A has been prepared. A new and convenient ion-selective electrode (ISE) method was used to assay Na⁺ release. Liposomes were formed from three d

Full text of DOI:10.1021/ja044936+

Published on Web 12/15/2004
Dynamic Assessment of Bilayer Thickness by Varying
Phospholipid and Hydraphile Synthetic Channel Chain
Lengths
Michelle E. Weber,† Paul H. Schlesinger,‡ and George W. Gokel*,†,§
Contribution from the Departments of Molecular Biology and Pharmacology, Chemistry, and
Cell Biology and Physiology, Washington UniVersity School of Medicine,
660 South Euclid AVenue, St. Louis, Missouri 63110
Received August 21, 2004; E-mail: ggokel@molecool.wustl.edu
Abstract: A library of “hydraphile” synthetic ion channel analogues that differ in overall length from ∼28-
58 Å has been prepared. A new and convenient ion-selective electrode (ISE) method was used to assay
Na+ release. Liposomes were formed from three different phospholipids: 1,2-dimyristoleoyl-sn-glycero-3-
phosphocholine (DMPC), 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), and 1,2-dierucoyl-sn-glycero-
3-phosphocholine (DEPC). The acyl chains of the lipids comprise cis-unsaturated 14:1, 18:1, or 22:1
residues, respectively. Sodium release was measured for each liposome system with each of the synthetic
channels. Peak activity was observed for shorter channels in liposomes formed from DMPC and for longer
channels in DEPC. A separate study was then conducted in DMPC liposomes in the presence of the putative
membrane-thickening agents cholesterol and decane. Peak activity was clearly shifted to longer channel
lengths upon addition of 20 or 40 mol % cholesterol or n-decane to the liposome preparation.
Introduction
an active area. It has a long and rich history. Many prominent
biophysicists have published in it and moved on. Users of bilayer
Gorter and Grendel recognized in 1925 that the boundary
membrane of erythrocytes is a bilayer.1 Nearly 50 years later,
Singer and Nicolson offered a “fluid mosaic” model for the
plasma membrane.2 Extensive study of the plasma membrane
has led to a detailed understanding of its components and
composition in various organisms. Aspects of membrane
structure, including in vivo thickness and phase behavior, remain
poorly defined despite numerous and careful studies. As
proposed detailed structures for membrane proteins emerge, their
interactions with membranes in general and membrane lipids
in particular remain an important and unresolved question. Lipid
rafts, which were not anticipated by theories of membrane
structure, remain a conundrum. Their importance in protein
localization, protein activation, signal transduction, and mem-
brane fusion add real immediacy to the structural details of
membrane heterogeneity and dynamics.
Despite great interest, the determination of structural details
for fluid, heterogeneous, supramolecular complexes such as the
phospholipid bilayer membrane remain incomplete. Numerous
recent studies and reviews point to the lack of a clear
understanding of membrane thickness and phospholipid head-
group surface area. Nagle, Tristram-Nagle, and their co-workers
have made the case in recent reviews3 that certain “accepted”
membrane parameters are not uniformly reported in the litera-
ture. Indeed, they note3b: “some readers may challenge our
assertion that lipid bilayer structure should still be considered
structural data have many references to choose from and each
user has a favorite.” This variability3a in membrane organization
and thickness is likely to be amplified by interactions with
peptides or proteins embedded within them.
It is important to bear in mind that much information
concerning membrane structure has been acquired from X-ray
studies of powdered membranes. Even when humidified, these
membranes are not as dynamic as they would be in liposomes.
Small angle neutron scattering (SANS) is useful in this context
and has been applied to liposomes.4 We present here an
alternative approach to the assay of membrane thickness, which
relies upon a family of synthetic, cation-conducting channel
compounds.
The unique, synthetic channel system to which we have given
the family name “hydraphile” has a number of advantages over
natural peptide or protein channels. First, a family of hydraphile
channels has now been studied in detail.5 Their ability to
transport protons and cations has been confirmed by fluores-
cence methods,6 by NMR, by bilayer voltage clamp,7 and by
whole-cell patch clamping.8 Many of the family members are
(3) (a) Nagle, J. F.; Tristram-Nagle, S. Curr. Opin. Struct. Biol. 2000, 10, 474-
480. (b) Nagle, J. F.; Tristram-Nagle, S. Biochim Biophys Acta 2000, 1469,
159-95. (c) Tristram-Nagle, S.; Liu, Y.; Legleiter, J.; Nagle, J. F. Biophys.
J. 2002, 83, 3324-35. (d) Tristram-Nagle, S.; Nagle, J. F. Chem. Phys.
Lipids 2004, 127, 3-14.
(4) (a) Pencer, J.; Hallett, F. R. Phys. ReV. E 2000, 61, 3003-3008. (b) Balgavy,
P.; Dubnickova, M.; Kucerka, N.; Kiselev, M. A.; Yaradaikin, S. P.;
Uhrikova, D. Biochim. Biophys. Acta 2001, 1512, 40-52. (c) Kucerka,
N.; Kiselev, M. A.; Balgavy, P. Eur. Biophys. J. 2004, 33, 328-34.
(5) Gokel, G. W. Chem. Commun. 2000, 1-9.
(6) Murillo, O.; Watanabe, S.; Nakano, A.; Gokel, G. W. J. Am. Chem. Soc.
1995, 117, 7665-7679.
(7) Abel, E.; Maguire, G. E. M.; Meadows, E. S.; Murillo, O.; Jin, T.; Gokel,
G. W. J. Am. Chem. Soc. 1997, 119, 9061-9062.
† Department of Molecular Biology and Pharmacology.
‡ Department of Cell Biology and Physiology.
§ Department of Chemistry.
(1) Gorter, E.; Grendel, F. J. Exp. Med. 1925, 41, 439-443.
(2) Singer, S. J.; Nicolson, G. L. Science 1972, 175, 720-731.
9
636
J. AM. CHEM. SOC. 2005, 127, 636-642
10.1021/ja044936+ CCC: $30.25 © 2005 American Chemical Society

Dynamic Assessment of Bilayer Thickness
A R T I C L E S
biologically active in contexts that comport with channel
formation.9 Recent computational studies by others generally
confirm the experimental conclusions.10 The hydraphile channels
are active over a range of lengths and sidearm functions.
Accordingly, we prepared a family of hydraphiles that were
identical in all respects except for overall length. We prepared
vesicles from phospholipid monomers of various lengths and
then directly observed Na+ efflux from the liposomes as a mea-
sure of efficacy. In this way, we have obtained a dynamic mea-
sure of bilayer thickness that may be compared with previous
values, often obtained from X-ray crystallographic data.11
et al. reported the use of a K+ ISE to examine the membrane-
destabilizing activity of nisin, but the report did not include
analytical details.15 Silberstein et al. used ISE methodology to
assay the membrane-destabilizing abilities of several viral and
peptide pore formers.16 In part following their lead, we have
developed an ISE method that uses a sodium selective electrode
to assay release of Na+ from large unilamellar vesicles (LUV).
To our knowledge, this is the first report in which a sodium-
selective ISE is used to quantitate ion channel activity.
Liposomes were prepared as described below from phospho-
lipids having different fatty acyl chain lengths. The vesicles
were prepared in an aqueous solution containing 750 mM NaCl/
15 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid
(HEPES) at pH 7.0. This aqueous solution was entrapped as
the vesicles were formed. The vesicles were extruded through
a 0.2 µm polycarbonate membrane filter. The initial buffer
solution was exchanged for a sodium-free buffer by passing
the liposomes over a Sephadex G25 column. The external buffer
was choline chloride (750 mM) and HEPES (15 mM, pH 7.0).
Vesicle concentration was measured as reported,17 and vesicle
size was measured by a Coulter N4MD submicron particle
analyzer. The vesicles used in the transport studies had diameters
of ∼200-250 nm (∼2000-2500 Å), and the observed sizes
were larger when the lipid chain length was longer.
Sodium transport was measured by inserting a combination
pH/sodium electrode (see Experimental Section) into a beaker
containing buffered vesicle solution (lipid concentration 0.4 mM,
total volume 2 mL). The baseline was established by recording
for 5 min prior to addition of the channel, which was added as
a 2-propanol solution. Sodium flux was recorded for 25 min.
A value for total sodium release was obtained by n-octylglu-
coside-induced lysis of the vesicles. The data were collected in
millivolt units and converted to concentration by using the
electrode’s calibration curve. The concentration data were
normalized to the final sodium release value after vesicle lysis.
Vesicle Studies. The vesicles used in the studies presented
here were formed from the following phospholipids: 1,2-
dimyristoleoyl-sn-glycero-3-phosphocholine, 1,2-dioleoyl-sn-
glycero-3-phosphocholine, and 1,2-dierucoyl-sn-glycero-3-
phosphocholine. The acyl chains comprise cis-unsaturated 14:
1, 18:1, or 22:1 residues, respectively. Experiments were also
conducted in which either cholesterol (20 or 40 mol %) or
n-decane (25% w/w) was added to the dimyristoleoyl lipids.
Concentration Dependence of Na+ Transport by Benzyl
Channel 3. It was first necessary to demonstrate that the ISE
method gave a reproducible account of Na+ release from
vesicles. This was done in vesicles formed as described below
from 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC). The
internal Na+ concentration was 750 mM, and release was
measured in separate experiments using the following concen-
trations of dibenzyl (C12) channel 3: 2, 4, 8, 12, 16, and 20
µM. Figure 1 shows the results. Data were normalized to full
sodium release after vesicle lysis (see above) and plotted as
fractional Na+ release (1 ) full release) on the ordinate. Each
Results
Synthesis of the Hydraphile Channel Compounds. Seven
compounds that are well established as synthetic ion channel
formers6-9 were prepared for use in the present study and are
shown below. Five of them (1-5) were prepared as reported
previously.8,12 The preparation was accomplished by monode-
benzylation by partial hydrogenolysis of N,N′-dibenzyl-4,13-
diaza-18-crown-6, followed by alkylation of the secondary
macroring nitrogen atom with the appropriate dibromoalkane.
Using a shorthand that we previously reported,13 we can describe
this conversion schematically as PhCH2 N18N CH2Ph f PhCH2-
N18N H f PhCH2 N18N (CH2)nBr. Coupling of the latter
bromide with 4,13-diaza-18-crown-6 affords PhCH2 N18N -
(CH2)n N18N (CH2)n N18N CH2Ph, 1-7. Compounds 6 and 7
have not been previously described, and their syntheses are
recorded in detail in the Experimental Section.
Analytical Method Used for Sodium Ion Transport Assay.
Because the scope of the present study is rather large, we
required an analytical method that was reliable but more
convenient than either the NMR or bilayer clamp methods, both
of which are labor intensive. Although we have used ion-
selective electrodes (ISE) to measure ion receptor interactions
in solution for more than two decades,14 the use of ISEs for
dynamic measurements of ion flux is relatively recent. Breukink
(8) Leevy, W. M.; Pajewski, R.; Huettner, J. E.; Schlesinger, P. H.; Gokel, G.
W., in press.
(9) Leevy, W. M.; Donato, G. M.; Ferdani, R.; Goldman, W. E.; Schlesinger,
P. H.; Gokel, G. W. J. Am. Chem. Soc. 2002, 124, 9022-9023.
(10) a) Srinivas, G.; Lopez, C.; Klein, M. J. Phys. Chem. B 2004, 108, 4231-
4235. (b) Srinivas, G.; Klein, M. L.; Nanotechnol. 2004, 15, 1289-1295.
(11) (a) Fettiplace, R.; Andrews, D. M.; Haydon, D. A. J. Membr. Biol. 1971,
5, 277-296. (b) Caffrey, M.; Feigenson, G. W. Biochemistry 1981, 20,
1949-1961. (c) Lewis, B. A.; Englelman, D. M. J. Mol. Biol. 1983, 166,
211-217. (d) Nezil, F. A.; Bloom, M. Biophys. J. 1992, 61, 1176-1183.
(e) Wiener, M. C.; White, S. H. Biophys. J. 1992, 61, 434-447.
(12) (a) Murray, C. L.; Gokel, G. W. Chem. Commun. 1998, 2477-2478. (b)
Murray, C. L.; Gokel, G. W. J. Supramol. Chem. 2001, 1, 23-30.
(13) Hernandez, J. C.; Trafton, J. E.; Gokel, G. W. Tetrahedron Lett. 1991,
6269-6272.
(14) (a) Dishong, D. M.; Gokel, G. W. J. Org. Chem. 1982, 47, 147-148. (b)
Schultz, R. A.; White, B. D.; Dishong, D. M.; Arnold, K. A.; Gokel, G.
W. J. Am. Chem. Soc. 1985, 107, 6659-6668.
(15) Breukink, E.; Kraaij, C. V.; Demel, R. A.; Siezen, R. J.; Kuipers, O. P.;
Kruijff, B. D. Biochemistry 1997, 36, 6968-6976.
(16) Silberstein, A.; Mirzabekov, T.; Anderson, W. F.; Rozenberg, Y. Biochim.
Biophys. Acta 1999, 1461, 103-112.
(17) Stewart, J. C. M. Anal. Biochem. 1980, 104, 10-14.
9
J. AM. CHEM. SOC. VOL. 127, NO. 2, 2005 637

A R T I C L E S
Weber et al.
Table 1. Phase Transition Temperatures (∆Tm) and Estimated
Bilayer Lengths
lipid name
∆Tm
(
°C)
estimated length (Å)a
DMPCb
DOPCc
DEPCd
>-30
-22
11
20-45
27-50
32-50+
a Range of bilayer widths represents the scope of values found in the
literature. b Van Dijk, P. W. M.; de Kruijff, B.; Van Deenen, L. L. M.; De
Gier, J.; Demel, R. A. Biochim. Biophys. Acta 1976, 455, 576-587.
c Tomita, T.; Wantabe, M.; Yasuda, T. Biochim. Biophys. Acta 1992, 1104,
325-330. d Reference 11b.
to three phospholipids, more than 100 series of measurements
were required to obtain the data reported here.
Figure 1. Sodium release from DOPC vesicles mediated by benzyl channel
3 at concentrations of 2, 4, 8, 12, 16, and 20 µM.
Figure 2. Release from DOPC vesicles mediated by compounds 1-7. The
data points show fractional Na+ release measured at 1500 s when 12 µM
of compound was present. Each point represents at least three replicates.
line is the average of at least three separate determinations,
conducted at each concentration. Concentration-dependent
release was observed, and full Na+ release by 3 occurred in
<900 s at [3] ) 20 µM.
Dependence of Na+ Transport on Spacer Chain Length
in DOPC Vesicles. All transport studies were conducted at least
in triplicate. Transport studies were conducted under identical
conditions of concentration, pH, temperature, etc., to the extent
these could be controlled. Data are reported as percent transport
monitored at 1500 s as an arbitrarily selected basis for
comparison. The error bars in Figure 2 show the variation in
the data.
Choice of Phospholipid Monomers for Liposome Forma-
tion. The scope of this study required that each of the seven
channels was assayed in at least three sets of liposomes. Thus,
one series required a minimum of 21 separate experiments. In
the work reported here, three closely related phospholipids were
used. These had different chain lengths, a single cis-double bond,
and identical headgroups. The phospholipids chosen were 1,2-
dimyristoleoyl-sn-glycero-3-phosphocholine (Z14:1, DMPC),
1,2-dioleoyl-sn-glycero-3-phosphocholine (Z18:1, DOPC), and
1,2-dierucoyl-sn-glycero-3-phosphocholine (Z22:1, DEPC). The
three phospholipid structures chosen for the experimental survey
are illustrated. Table 1 records phase transition temperatures
(∆Tm) and estimated hydrophobic bilayer widths for each of
the lipids used in this study. The lipids were chosen so that
room-temperature experiments could be conducted above ∆Tm
(i.e., in the gel phase). The study also included examining the
influence of n-decane and cholesterol on the effective thickness
of bilayers formed from DMPC. Although the effort was limited
Effect of Channel Spacer Chain and Lipid Length on
Sodium Transport from Liposomes. Vesicles of ∼200-250
nm diameter were formed from DMPC, DOPC, or DEPC as
described below. Release of Na+ from the vesicles was
monitored for 25 min by using a sodium-selective ISE. The
concentration of 1-7 used in each case was 12 µM, and the
ionophore was introduced as a solution in 2-propanol. The
results are shown in Figure 3.
The effect of membrane-thickening agents on Na+ transport
from DMPC liposomes by 1-7 was also assayed. Cholesterol
9
638 J. AM. CHEM. SOC. VOL. 127, NO. 2, 2005

Dynamic Assessment of Bilayer Thickness
A R T I C L E S
Table 2. Estimated Overall Span Lengths for 1-7
compd No.
spacer chain length
estimated length (Å)a
1
2
3
4
5
6
7
C8
28.3
33.3
38.4
43.4
48.5
53.5
58.6
C10
C12
C14
C16
C18
C20
a See text for method used.
uncomplexed crown structures measured (CSD descriptors:
BADDOZ, BEXMUM, EFEDAU, EFEDIC, GOXWAR), the
N-N distance varied from 6.799 to 7.736 Å; the average N-N
distance was 7.292 Å. A typical C-C bond length (e.g., ethane)
is 1.54 Å. We measured the chain length of several saturated
lipid chains to determine the lineal distance spanned by the
saturated C-C bonds. The average length spanned by a C-C
bond was found to be 1.262 Å. The C-N bond of carbon
attached to macroring nitrogen is expected to be shorter than a
C-C bond, owing to the greater electronegativity of nitrogen.
Measurements of alkylated macrocycles in the solid state
consistently gave a C-N bond length of 1.47 Å. Estimates of
hydraphile length were made by summing the N-N distance
across the central relay, adding four times the C-N bond
distance, and then the appropriate number of C-C bonds. The
estimated lengths are shown in Table 2.
Methodology Development. The Na+ release assay requires
that several variables be controlled. In particular, three variables
were optimized as part of this study. These include: (1) internal
and external salt concentrations, (2) pH, and (3) solvent for
channel inoculation. The salt concentrations for both internal
and external buffer systems were varied from 500 mM to 1.0
M. At the lower end of this concentration range, electrode
response became sluggish. At the highest concentrations, Na+
release occurred in the electrode’s dead time. In the studies
presented here, internal [Na+] ) 750 mM.
Each hydraphile has six nitrogen atoms that may be proto-
nated. The pKa values for closely related structures were deter-
mined and reported previously.6 The hydraphiles function in
the pH 7-8 range. Acidity values in this range were studied,
and pH ) 7.0 was chosen for the buffer solutions. At this pH
value, both electrode response and channel function were
acceptable.
The macrocyclic subunits of each hydraphile are crown ethers,
which can complex a variety of cations. We wished to avoid
competing cations. We tested both KOH and NH4OH, which
decreased the transport activity of the hydraphiles. We ultimately
found that (CH3)4NOH could be used effectively to adjust the
buffer solutions to pH 7.0.
Figure 3. Sodium release monitored at 1500 s for compounds 1-7 (12
µM) in DMPC, DOPC, and DEPC vesicles.
was added at concentrations of 20 or 40 mol % relative to
DMPC. The experiments with n-decane used a ratio of n-decane/
phospholipid that was 1:3 (w/w). In both cases, liposomes were
prepared by mixing the diluent with the phospholipids prior to
vesicle formation so that uniform distribution of the excipient
would be achieved.
Discussion
Use of Hydraphiles as a Membrane Probe. The hydraphiles
are now a well-established and well-characterized class of syn-
thetic channels.6-9,12 These channels show open-shut behavior
characteristic of protein channels. They exhibit selectivity for
cations over anions and for Na+ over K+. The hydraphiles are
all symmetrical structures so they are inherently nonrectifying.
The aggregation state was assessed for two fluorescent hydra-
philes and found to be ∼1.7 This implies that they function as
monomers. They have also shown a fascinating range of biolog-
ical activity. For many years, peptide pore formers such as gram-
icidin and melittin have been favored as channel models in the
biological community. This was in part because synthetic chan-
nels were not available and in part because the pore formers
were naturally occurring and therefore inherently biological.
More recently, a variety of synthetic peptides have been studied,
and these may be obtained in chemical purity and with specific
lengths and sequences. Synthetic channels offer the prospect that
they may be more finely manipulated by the chemist and thus
may serve as a more sensitive probe of a given biological effect.
The hydraphiles used in the present study all are members
of the benzyl channel family. The shorthand we use to represent
compound 1 (structure shown above) is PhCH2 N18N (CH2)8-
N18N (CH2)8 N18N CH2Ph. Results obtained to date show that
the channel’s span is defined by the width of the central
macrocycle and the two oligomethylene chains appended to it.
Measurements of Approximate Hydraphile Lengths. We
attempted computerized molecular modeling of 1-7 to measure
the distance between proximal nitrogens in the two distal
macrocycles. The calculations we attempted gave energy-
minimized structures in which the molecule was folded back
upon itself rather than in an extended conformation. We
abandoned this strategy and turned instead to the Cambridge
Structural Database (CSD) to obtain dimensions of substructural
units that are reported in the solid state.
A final concern is the solvent used to add the hydraphiles to
the aqueous buffer. The solvent chosen should not inhibit ion
transport nor contain ionizable protons at pH ) 7. Water/
compound solubility was required for compound delivery. In
addition, the solvent could not compromise the structural
integrity of the liposomes. Compounds 1-7 are not soluble in
DMSO, the solvent commonly used for peptide delivery into
an aqueous medium. A range of solvents was surveyed,
including acetonitrile, ethanol, n-propanol, n-butanol, trifluo-
roethanol, and others. 2-Propanol was ultimately adopted for
these studies, but care was exercised to keep the solvent amount
Since we have assumed the hydraphiles are fully extended
across the lipid bilayer and that the central relay is only
transiently complexed, we measured the N-N distance across
the macrocycle for several substituted diazacrowns in the CSD,
using the included software package ConQuest 1.4. For the five
9
J. AM. CHEM. SOC. VOL. 127, NO. 2, 2005 639

A R T I C L E S
Weber et al.
below 15 µL. When amounts greater than 20 µL were used,
isopropyl alcohol caused slight ion leakage from the vesicles.
There are literature reports16,18 of detergent interference with
sodium-selective electrode response. Triton X-100 proved
problematic in this respect. Use of a 10% aqueous solution of
n-octylglucoside eliminated these problems and was found to
afford complete lysis of the liposomes in <10 s. A thorough
rinse of the electrode with an aqueous buffer solution proved
to be sufficient to remove residual detergent from the electrode.
Channel Response as a Function of Lipid Chain Length.
Figure 3 plots the Na+ release rate observed at 1500 s for
compounds 1-7 in liposomes prepared from DMPC, DOPC,
and DEPC. Each phospholipid (structures shown above) contains
a cis (Z) double bond that prevents the extended conformation
from being as long as would be the case if both chains were
saturated. The individual chain lengths are 14 carbons (Z14:1,
DMPC), 18 carbons (Z18:1, DOPC), and 22 carbons (Z22:1,
DEPC). The shortest chains are present in DMPC, the activity
profile of which is shown as the blank line in Figure 3.
Compound 1 (eight-carbon spacer chains) is inactive, but
essentially complete release is obtained when any of 2-5 (10-
16 carbon spacer chains) is present. Activity decreases for the
two longest hydraphiles, 6 and 7, although both release Na+
from the liposomes.
Figure 4. Sodium release monitored in DMPC membranes at 1500 s for
1-7 in the presence of 0, 20, or 40 mol % cholesterol.
Table 3. Percent Activities Observed for 1-7 in DMPC, DOPC,
and DEPC Membranes
phospholipid
compd No.
DMPC
DOPC
DEPC
1
2
3
4
5
6
7
4
98
1
17
73
86
82
32
12
0
9
109a
105a
104a
85
47
64
96
77
52
43
a Ratio of observed release to value after lysis is slightly greater than 1.
The activity profile is significantly different for 1-7 for the
intermediate length DOPC phospholipid vesicles (red line).
Again, 1 is inactive, but 2 (10 carbon spacers) releases Na+
only to the extent of ∼20%. Maximal ion release reaches only
80-90% for 3-5 (12-16 carbon spacers) and then falls off
rapidly. A similar shift in activity is apparent for the longest
chain phospholipids, DEPC (blue line). In the latter case, a
gradual increase in activity is apparent for 1 (inactive), 2 (∼10%
release), 3 (∼50%), and 4 (60%). Maximal release in this lipo-
some system is observed for 5 (∼95%), but activity diminishes
rapidly as chain length increases. The leveling of activity ob-
served for the longest hydraphiles [6 (C18) and 7 (C20)] may
result from the channels being longer than the membrane is
thick. Indeed, independent molecular dynamics simulations10
of the hydraphile channels show that 7 resides in a single leaflet
of the bilayer and is therefore unable to adopt a stable trans-
membrane conformation. Clearly, these channels continue to
function but do not increase in efficacy beyond a certain length.
Overall, it is clear that longer chained hydraphiles are required
for activity in liposomes prepared from longer chained phos-
pholipids. Compound 1 is essentially inactive in all three
liposome preparations. Compound 2 exhibits near maximal
activity in DMPC liposomes but shows <20% release in DOPC.
Even so, this is twice the 9% release apparent in DEPC vesicles.
A similar trend is observed for 3. Table 3 records the average
Na+ release as a percentage of release after vesicular lysis for
1-7 in all three liposome preparations. The data shown in the
table are an alternate representation of the graph of Figure 3.
Membrane-Thickening Agents. The ability of cholesterol
to thicken a bilayer has been documented.19 It is proposed that
cholesterol partitions into the PC bilayer with the 3-hydroxyl
group hydrogen bonding to the ester carbonyl group of the
phospholipid (recently reviewed20). When the lipids are in the
liquid crystalline phase, above Tm, the downward placement of
cholesterol is thought to increase acyl chain ordering by
restricting the flexibility of adjacent phospholipid acyl chains.
Changes in cholesterol levels have been demonstrated to affect
many different biological processes, including ion channel
function,21 membrane enzyme activity,22 the accessibility of
membrane bound substrates,23 and receptor display.24
We have used our ISE-based method to probe the effect that
cholesterol incorporation has on hydraphile-mediated ion trans-
port. Figure 4 (above) shows the influence of added cholesterol
(20 or 40 mol %) on transport by 1-7 in DMPC liposomes.
An increase in apparent membrane thickness was observed when
either 20 or 40 mol % cholesterol was added, but the effect
was greater when more steroid was present. The profile shown
in Figure 3 for DMPC is shifted to the right (see Figure 4),
indicating an effective increase in membrane thickness. More-
over, the putative membrane-thickening agent n-decane shows
a similar shift in activity (Figure 5), confirming its effect on
the bilayer.
It seems unlikely to us that the results presented here reflect
changes in membrane fluidity. It is known that the presence of
cholesterol in a bilayer to the extent of >25 mol % promotes
the formation of an ordered liquid phase. However, this
concentration of cholesterol does not limit the mobility of
integral membrane proteins above Tm.25 Cornelius has examined
the effects of both membrane fluidity and membrane thickness
(20) Ohvo-Rekila, H.; Ramstedt, B.; Leppimaki, P.; Slotte, J. P. Prog. Lipid
Res. 2002, 41, 66-97.
(21) (a) O’Connell, K. M. S.; Martens, J. R.; Tamkun, M. Trends CardioVasc.
Med. 2004, 14, 37-42. (b) Martens, J. R.; O’Connell, K.; Tamkun, M.
Trends Pharmacol. Sci. 2003, 25, 16-21. (c) Venegas, B.; Gonzalez-
Damian, J.; Celis, H.; Ortega-Blake, I. Biophys. J. 2003, 85, 2323-2332.
(22) Cornelius, F. Biochemistry 2001, 40, 8842-8851.
(23) (a) Pilot, J. D.; East, J. M.; Lee, A. G. Biochemistry 2001, 40, 8188-
8195. (b) Lehto, M. T.; Sharom, F. J. Biochemistry 2002, 41, 1398-1408.
(24) Pike, L. J.; Casey, L. Biochemistry 2002, 41, 10315-10322.
(25) Ipsen, J. H.; Karlstrom, G.; Mouritsen, O. G.; Wennerstrom, H. W.;
Zuckermann, M. J. Biochim. Biophys. Acta 1987, 905, 162-172.
(18) Gruber, H. J.; Wilmsen, H. U.; Schurga, A.; Pilger, A.; Schindler, H.
Biochim. Biophys. Acta 1995, 1240, 266-276.
(19) Smondyrev, A. M.; Berkowitz, M. L. Biophys. J. 1999, 77, 2075-2089.
9
640 J. AM. CHEM. SOC. VOL. 127, NO. 2, 2005

Dynamic Assessment of Bilayer Thickness
A R T I C L E S
1,2-Dimyristoleoyl-sn-glycero-3-phosphocholine, 1,2-dioleoyl-sn-
glycero-3-phosphocholine, and 1,2-dierucoyl-sn-glycero-3-phospho-
choline were purchased from Avanti Polar Lipids as chloroform
solutions. Decane, cholesterol, HEPES, and the inorganic salts NaCl
and cholineCl were all purchased from Sigma-Aldrich. The water that
was used for all buffer preparations was of Milli-Q Plus quality, which
is essential to avoid salt contamination in the buffer systems. N-
Octylglucoside was purchased from CalBioChem.
Vesicle Preparation. The vesicles used were prepared using the
reverse evaporation method of Szoka and Papahadjopoulos.26 Vesicles
comprised cis-unsaturated 14:1, 18:1, or 22:1 lipids. 1,2-Dimyristoleoyl-
sn-glycero-3-phosphocholine, 1,2-dioleoyl-sn-glycero-3-phosphocho-
line, and 1,2-dierucoyl-sn-glycero-3-phosphocholine were obtained as
CHCl3 solutions, which were dried to lipid films and stored at ambient
temperature under high vacuum. The vesicles were prepared by
dissolving a dry lipid film in 0.3 mL of diethyl ether and 0.3 mL of
buffer (750 mM NaCl/15 mM HEPES, pH 7.0). The mixture was
sonicated for ∼20 s to give an opaque solution. The organic solvent
was removed under reduced pressure, and the solution was passed
through a mini extruder containing a 0.2-µm polycarbonate membrane
filter. The residual, external buffer solution was exchanged for a sodium-
free buffer (750 mM cholineCl/15 mM HEPES, pH 7.0) via passage
over a Sephadex G25 column. Vesicle concentration was measured as
reported,17 and vesicle size was measured by a Coulter N4MD
submicron particle analyzer. The vesicles used in the transport studies
had diameters of ∼200-250 nm, depending on lipid tail length.
For experiments using dimyristoleoyl lipids and either cholesterol
or decane, the appropriate amount of additive (20 or 40 mol % for
cholesterol or 25% w/w for n-decane) was added to the dry lipid film.
The lipid and additive were then dissolved in 0.3 mL of dry Et2O, and
the vesicles were formed and characterized as outlined above.
Na+ Transport Measurements. All data were collected by Axo-
scope 7.0 using a Digidata 1322A series interface. Sodium transport
was measured using a Micro-Combination pH/sodium electrode
(Thermo-Orion) in aqueous sodium-free buffer (750 mM cholineCl/
15 mM HEPES, pH 7.0). After equilibration in external buffer, the
electrode was placed in a 5-mL disposable beaker containing external
buffer and vesicle solution to achieve a lipid concentration of 0.4 mM
and a total volume of 2 mL. To set the baseline, the voltage output
was recorded for 5 min prior to addition of the channel. The channel
solution in 2-propanol was added, and channel conductivity was
measured over an appropriate time range, typically 25 min. Addition
of 200 µL of 10% aqueous n-octylglucoside induced complete lysis of
the vesicles to achieve total sodium release. All experiments were
performed at room temperature.
Figure 5. Sodium release monitored at 1500 s from DMPC liposomes for
1-7 in the presence of 40 mol % cholesterol or 25% w/w decane.
and shown that membrane thickness rather than lipid fluidity is
critical for ATPase activity.22 Further, decane gives results (see
Figure 5) similar to those observed for cholesterol but is not
known to significantly impact membrane fluidity. Finally, if
fluidity was the key determinant of these results, all lengths of
channels should be affected. Taken together, it appears that
bilayer thickness, not changes in fluidity, account for the
observed results.
Conclusion
We report here a methodology that permits a direct measure
of Na+ release from phospholipid vesicles. The method, which
uses an ion-selective electrode to measure Na+ transport,
provides a rapid and reproducible technique to obtain concentra-
tion-dependent ion release data. We have used this methodology
to measure the Na+ transport of a series of synthetic hydraphiles
that differed only in their overall length. The results thus
obtained were validated by comparison with length-dependence
data previously obtained by NMR methods. We have further
tested lipids that vary in length, from C14 to C22 acyl chains,
and observed significant changes in the activity profiles of the
hydraphiles. When liposomes of shorter chain lengths were
prepared, shorter hydraphiles showed increased Na+ transport.
As the lipid length increased, the shorter hydraphiles became
less active or inactive in the thicker membranes. This trend was
confirmed by using liposomes formed from dimyristoleylphos-
phtidylcholine diluted with the putative membrane-thickening
agents cholesterol and n-decane. In all cases, the activity profile
for Na+ release was shifted to the right, indicating that the
shorter hydraphiles cannot form active channels in wider
membranes.
Data Analysis. All data were analyzed using OriginPro 7. The data,
which were collected in units of millivolts, were first converted to units
of concentration, using the calibration curve for the electrode. A new
calibration curve was generated every few weeks to account for the
changing temperature of the room, diminishing electrode performance
with age, and any other condition that might affect its performance.
After converting from millivolt to concentration, the data were
normalized to the total sodium concentration, as determined by lysing
the vesicles with n-octylglucoside.
Experimental Section
1H, 13C, and 23Na NMR spectra (in ppm (δ) downfield from internal
Me4Si) were recorded at 300, 75, and 132.3 MHz respectively, in
CDCl3, unless otherwise stated. Data are presented as follows: chemical
shift, peak multiplicity (b ) broad, s ) singlet, d ) doublet, t ) triplet,
m ) multiplet, bs ) broad singlet, etc.), assignment, integration. Melting
points were determined in open capillaries and are uncorrected. Thin
layer chromatographic (TLC) analyses were performed on aluminum
oxide 60 F-254 neutral (type E) or on silica gel 60-F-254 (0.2-mm
thickness). Preparative chromatography columns were packed with acti-
vated aluminum oxide (MCB 80-325 mesh, chromatographic grade,
AX 611) or Kieselgel 60 (70-230 mesh). Flash chromatography col-
umns were packed with silica gel Merck grade 9385, 230-400 mesh,
60 Å.
Hydraphile Preparation. N,N′-Bis[10-(N′-benzyldiaza-18-crown-
6)octyl]diaza-18-crown-6, [C6H5CH2 N18N (CH2)8 N18N (CH2)8-
N18N CH2C6H5], 1, was prepared as reported in detail in ref 12b.
N,N′-Bis[10-(N′-benzyldiaza-18-crown-6)decyl]diaza-18-crown-
6, [C6H5CH2 N18N (CH2)10 N18N (CH2)10 N18N CH2C6H5], 2, was
prepared as reported in detail in ref 8.
N,N′-Bis[12-(N′-benzyldiaza-18-crown-6)dodecyl]diaza-18-crown-
6, [C6H5CH2 N18N (CH2)12 N18N (CH2)12 N18N CH2C6H5], 3, was
prepared as previously reported.8
All reagents were the best grade commercially available and were
distilled, crystallized, or used without further purification, as appropriate.
THF and diethyl ether were distilled from sodium metal and benzophe-
none.
(26) Szoka, F. J.; Papahadjopoulos, D. Proc. Natl. Acad. Sci. U.S.A. 1978, 75,
4194-4198.
9
J. AM. CHEM. SOC. VOL. 127, NO. 2, 2005 641

A R T I C L E S
Weber et al.
N,N′-Bis[14-(N′-benzyldiaza-18-crown-6)tetradecyl]diaza-18-crown-
6, [C6H5CH2 N18N (CH2)14 N18N (CH2)14 N18N CH2C6H5], 4, was
prepared as reported in detail in ref 8.
N,N′-Bis[16-(N′-benzyldiaza-18-crown-6)hexadecyl]diaza-18-crown-
6, [C6H5CH2 N18N (CH2)16 N18N (CH2)16 N18N CH2C6H5], 5, was
prepared as reported in detail in ref 8.
(56 mg, 7% yield) as a white solid, mp 52-53 °C. 1H NMR: 1.17 and
1.37 (s, bs, 64H), 2.42 (bs, 8H), 2.73 (t, 24H), 3.54 (pt, s, 52H). 13C
NMR: 227.18, 29.69, 53.22, 53.38, 53.79, 59.62, 69.84, 70.53, 127.04,
128.24, 128.90. IR (νmax cm-1): 2924, 2853, 1643, 1454, 1113. Calcd:
C, 70.35; H, 10.85; N, 5.72. Found: C, 70.12; H, 10.55; N, 5.67%.
Preparation of the N,N′-Bis[20-(N′-benzyldiaza-18-crown-6)-
eicosyl]diaza-18-crown-6, [C6H5CH2 N18N (CH2)20 N18N (CH2)20-
N18N CH2C6H5, 7. CH3OCO(CH2)18COOCH3. Methanol (100 mL)
was dried over 4 Å molecular sieves for ∼30 min prior to use. 1,18-
Octadecanedicarboxylic acid (4.99 g, 14.6 mmol) was dissolved with
heating in MeOH. Concentrated H2SO4 (0.3 mL) was added, and the
reaction was heated under reflux (2 h) and cooled to room temperature,
whereupon a white precipitate formed. Evaporation of the solvent gave
a white solid that was dissolved in EtOAc, washed with saturated
NaHCO3 (3 × 75 mL), and dried over Na2SO4, and the solvent was
removed in vacuo. The resulting diester (4.6 g, 86%) was isolated as
a white solid after drying in vacuo overnight, mp 63-64 °C (lit. 65-
Preparation of the N,N′-Bis[18-(N′-benzyldiaza-18-crown-6)-
octadecyl]diaza-18-crown-6, [C6H5CH2 N18N (CH2)18 N18N (CH2)18-
N18N CH2C6H5], 6. CH3OCO(CH2)16COOCH3. MeOH (100 mL)
was dried over 4 Å molecular sieves for ∼30 min prior to use. 1,16-
Octadecanedicarboxylic acid (5.02 g, 16.0 mmol) was dissolved with
heating in dry MeOH. Concentrated H2SO4 (1.2 mL) was added to the
solution, and the reaction was heated to reflux under N2 for 2 h. The
reaction was cooled to room temperature, and a white precipitate formed
upon cooling. The solvent was removed in vacuo to afford a white
solid. The product was dissolved in EtOAc, washed (3 × 75 mL) with
saturated NaHCO3, and dried over Na2SO4, and the solution was evapor-
ated to dryness. The crude diester was dried under high vacuum over-
night to give 5.29 g (97%) of a white powder, mp 57-58 °C (lit. 58-
1
66 °C). H NMR: 1.25 (s, 28H), 1.61 (bt, 4H), 2.30 (t, 4H), 3.66 (s,
6H). 13C NMR: 24.95, 29.14, 29.25, 29.45, 29.66, 29.63, 34.11, 51.42.
HO(CH2)20OH. The diester CH3O(CH2)20OCH3 (3.27 g, 8.83 mmol)
was dissolved in freshly distilled THF (N2), Red-Al (10.6 mL, 35.3
mmol) was added, and the reaction was stirred for 3 h. The reaction
was quenched by careful addition of H2O and stirred for an additional
30 min. The mixture was filtered through Celite, EtOAc was added to
the mother liquor, and the resulting mixture was washed with H2O (3
× 50 mL). The organic layer was dried (Mg2SO4) and then evaporated
to afford the diol (1.73 g, 62.5%) as a white solid after drying, mp
1
59 °C). H NMR: 1.25 (s, 24H), 1.61 (bt, 4H), 2.30 (t, 4H), 3.66 (s,
6H). 13C NMR: 24.95, 29.14, 29.25, 29.45, 29.58, 29.63, 34.12, 51.42.
HO(CH2)18OH. The diester CH3O(CH2)18OCH3 (3.56 g, 10.4 mmol)
was dissolved in freshly distilled THF (under N2), Red-Al (12.5 mL,
41.6 mmol) was added, and the reaction was stirred at room temperature
for 3 h. Water was added dropwise until gas evolution ceased; a white
solid formed at this time. The mixture was stirred for an additional 30
min and then filtered through a pad of Celite. EtOAc was added, and
the reaction was washed 3× with H2O to remove the white solid, which
dispersed into the aqueous layer. The organic layer was dried over
Mg2SO4 and evaporated in vacuo to afford 2.43 g (82%) of a white
solid, mp 99 °C (lit. 98-99 °C) after drying under high vacuum
1
101-102 °C (lit. 102-103 °C). H NMR: 1.25 (s, 32H), 1.56 (mt,
4H), 3.64 (t, 4H). 13C NMR: 25.73, 29.41, 29.60, 29.66, 63.11.
Br(CH2)20Br. Acetic anhydride (25 mL) was added (dropwise, 0
°C) to concentrated HBr (12 mL) and HO(CH2)20OH. The reaction
mixture was stirred under reflux for 5 h, cooled, and concentrated to a
brown solid. The dibromide was chromatographed (SiO2, hexanes) to
1
overnight. H NMR: 1.25 (s, 28H), 1.56 (m, 4H), 3.64 (t, 4H). 13C
NMR: 25.73, 29.41, 29.63, 32.80, 63.11.
1
afford a white solid (1.94 g, 79.7%), mp 65 °C (lit. 66-67 °C). H
Br(CH2)18Br. Acetic anhydride (30 mL) was slowly added to
concentrated HBr (15 mL) and HO(CH2)18OH at 0 °C. The reaction
mixture was brought to reflux and stirred for 5 h before being cooled
to room temperature. The reaction was concentrated in vacuo under
reduced pressure to afford a brown solid upon drying. The product
was chromatographed by SiO2 (Hexanes) to afford a white solid (2.95
NMR: 1.25 (s, 28H), 1.42 (bt, 4H), 1.85 (m, 4H), 3.41 (t, 4H). 13C
NMR: 28.17, 28.6, 29.43, 29.54, 29.66, 32.85, 34.06.
PhCH2 N18N (CH2)20Br. PhCH2 N18N H (750 mg, 2.13 mmol) and
1,20-dibromoeicosane (886 mg, 2.02 mmol) were dissolved in n-PrCN
(50 mL). To this solution were added Na2CO3 (6.8 g, 63.9 mmol) and
catalytic KI. The reaction was heated (reflux, N2) for 3 h, cooled, and
evaporated to an orange oil, which was dissolved in CH2Cl2 and washed
with 5% NaHCO3 (2 × 25 mL). The organic layer was dried (Mg2-
SO4) and evaporated to an orange oil. Chromatography (SiO2, hexanes
f 5% Et3N/acetone) gave the dibromide (640 mg, 44%) as an orange
1
g, 85.8%), mp 59 °C (lit. 59-60 °C). H NMR: 1.26 (s, 28H), 1.42
(bm, 4H), 1.85 (m, 4H), 3.41 (t, 4H). 13C NMR: 28.17, 28.76, 29.43,
29.54, 29.64, 32.83, 34.07.
PhCH2 N18N (CH2)18Br. PhCH2 N18N H (750 mg, 2.13 mmol)
and 1,18-dibromooctadecane (1.05 g, 2.55 mmol) were dissolved in
n-PrCN (50 mL). To this solution were added Na2CO3 (7.5 g, 21.3
mmol) and a catalytic amount of KI. The reaction was stirred for 3 h
at reflux (N2) and cooled to room temperature. The solvent was
evaporated in vacuo, and the residual orange oil was dissolved in
CH2Cl2. The organic solution was washed with 5% NaHCO3 (2 × 50
mL). The organic layer was dried over Mg2SO4, and the solvent was
removed in vacuo to afford an orange oil. The product was purified by
SiO2 chromatography (hexanes f 5% Et3N/acetone) to afford 730 mg
(50%) of an oily yellow-orange solid with mp near ambient temperature.
1H NMR: 1.25 (s, 28H), 1.42 (bm, 2H), 1.85 (m, 2H), 2.50 (bt, 2H),
2.80 (t, 8H), 3.40 (t, 2H), 3.58-3.67 (m, s, 18H). 13C NMR: 27.11,
27.47, 28.16, 28.75, 29.41, 29.52, 29.66, 32.82, 34.06, 53.72, 53.87,
55.99, 59.91, 70.00, 70.67, 126.80, 128.12, 128.81, 139.61. IR (νmax
cm-1): 2924, 2853, 1454, 1352, 1126.
PhCH2 N18N (CH2)18 N18N (CH2)18 N18N CH2Ph, 6. PhCH2-
N18N (CH2)18Br (730 mg, 1.07 mmol) and diazacrown (133 mg, 0.508
mmol) were dissolved in ∼30 mL of n-PrCN. To this solution were
added Na2CO3 (1.09 g, 10.16 mmol) and catalytic KI. The reaction
was stirred under reflux (N2) for 4 d, cooled, and evaporated. The
resulting orange oil was dissolved in CH2Cl2, washed with 5% NaHCO3
(2 × 25 mL), dried over Mg2SO4, and evaporated to an orange oil.
Chromatography (SiO2, 1% Et3N/acetone f 2% Et3N/acetone) gave 6
1
oil. H NMR: 1.25 (s, 28H), 1.42 (bm, 4H), 1.85 (m, 2H), 2.49 (t,
2H), 2.80 (t, 8H), 3.40 (t, 2H), 3.58-3.67 (m, s, 16H), 7.24-7.34 (m,
5H). 13C NMR: 27.50, 28.16, 28.75, 29.41, 29.52, 29.60, 29.67, 32.83,
34.05, 53.73, 53.90, 56.02, 59.92, 70.03, 70.70, 126.80, 128.14, 128.82.
IR νmax (cm-1): 2924, 2853, 1454, 1351, 1126.
PhCH2 N18N (CH2)20 N18N (CH2)20 N18N CH2Ph, 7. PhCH2-
N18N (CH2)20Br (640 mg, 0.899 mmol) and 4,13-diaza-18-crown-6
(112 mg, 0.427 mmol) were dissolved in n-PrCN (30 mL), and Na2CO3
(4.7 g, 4.27 mmol) and catalytic KI were added. After being heated
(reflux) for 5 d, the mixture was cooled and evaporated to an orange
oil. The oil was dissolved in CH2Cl2, washed with 5% NaHCO3 (2 ×
25 mL), dried (Mg2SO4), and evaporated, and the orange oil was
chromatographed (SiO2, acetone f 2% Et3N/acetone) to afford 7 (40
1
mg, 4%) as a white solid, mp 47-48 °C. H NMR: 1.18 (s, 72H),
2.50 (bs, 8H), 2.73 (t, 24H), 3.54 (pt, s, 52H). 13C NMR: 27.49, 29.67,
29.41, 34.06, 53.75, 59.91, 70.02, 70.70, 128.14, 128.82 IR (νmax cm-1):
2924, 2853, 1454, 1353, 1114. Calcd. C, 70.91; H, 10.98; N, 5.51.
Act. C, 71.14; H, 11.02; N, 5.51.
Acknowledgment. We thank the NIH for grants (GM 36262
and 63190) and for training grant support of M.E.W. through
T32GM08785.
JA044936+
9
642 J. AM. CHEM. SOC. VOL. 127, NO. 2, 2005