Aggregate formation in the intercalation of long-chain fatty acid esters into liposomes

Article abstract of DOI:10.1016/j.chemphyslip.2008.07.005

Various hydrophobic benzenediacetic esters, the corresponding benzenedipropionic esters, and branched alkyl esters were intercalated into DMPC liposomes, where the molar ratio (n/n) of ester:DMPC was 1:5. In the case of the very long-chain derivatives, double carbonyl peaks were observed in the ¹³C NMR spectrum. This doubling phenomenon was observed only for the carbonyl peaks, whose chemical shift is most sensitive to solvent polarity, and disappeared when the ester:DMPC molar ratio drops below 1:15. This doubling reflects the presence of two populations in these samples: one group includes those molecules which are intercalated within the liposome and feel the polarity corresponding to the liposomal microenvironment; the other consists of aggregates of these long-chain derivatives located in the extra-liposomal aqueous phase.

Full text of DOI:10.1016/j.chemphyslip.2008.07.005

Chemistry and Physics of Lipids 155 (2008) 120–125
Contents lists available at ScienceDirect
Chemistry and Physics of Lipids
journal homepage: www.elsevier.com/locate/chemphyslip
Aggregate formation in the intercalation of long-chain fatty
acid esters into liposomes
Yael Cohena, Michal Afria, Aryeh A. Frimerb,∗
a Department of Chemistry, Bar-Ilan University, Ramat Gan 52900, Israel
b The Ethel and David Resnick Chair in Active Oxygen Chemistry, Bar-Ilan University, Ramat Gan 52900, Israel
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 3 April 2008
Received in revised form 7 July 2008
Accepted 8 July 2008
Available online 18 July 2008
Various hydrophobic benzenediacetic esters, the corresponding benzenedipropionic esters, and branched
alkyl esters were intercalated into DMPC liposomes, where the molar ratio (n/n) of ester:DMPC was 1:5.
In the case of the very long-chain derivatives, double carbonyl peaks were observed in the 13 C NMR
spectrum. This doubling phenomenon was observed only for the carbonyl peaks, whose chemical shift
is most sensitive to solvent polarity, and disappeared when the ester:DMPC molar ratio drops below
1:15. This doubling reflects the presence of two populations in these samples: one group includes those
molecules which are intercalated within the liposome and feel the polarity corresponding to the liposomal
microenvironment; the other consists of aggregates of these long-chain derivatives located in the extra-
liposomal aqueous phase.
Keywords:
Liposome
13 C NMR
ET(30)
© 2008 Elsevier Ireland Ltd. All rights reserved.
DMPC
Membrane
Aggregate
1. Introduction
2. Materials and methods
In a companion paper (Cohen et al., 2008), we described the syn-
thesis of benzenediacetic and dipropionic esters 1 and 2 (Scheme 1)
and their intercalation within the lipid bilayer of dimyristoylphos-
phatidylcholine (DMPC) liposomes. 13C NMR chemical shift data for
carbonyls I and II was used to determine the depth and orientation
of these compounds within the bilayer. As a rule, these carbonyls
were separated by about 1 ppm, with the less lipophilic carbonyl I
further downfield.
In these NMR studies, the molar ratio (n/n) of diester:DMPC was
generally 1:5. This rather high concentration of intercalant was nec-
essary in order to observe the desired peaks via NMR. Interestingly,
however, in the 13C NMR spectra of the very long-chain derivatives
1e–g and 2e–g, where the total number of carbons in the ester
group “R” was 16 or above, we observed double carbonyl peaks.
One pair was sharp, absorbing at a relatively higher field, while
the other pair was broad and situated at lower field. This doubling
phenomenon was observed only for the carbonyl peaks (C–I and
C–II; see Scheme 1), whose chemical shift is most sensitive to sol-
vent polarity. We decided to study this doubling phenomenon a bit
more in depth, and this paper reports on our findings.
2.1. General
The NMR spectra of the esters in CDCl3 were recorded at
300 MHz (1H), 75 MHz (13C) or 50.3 MHz (13C) at 25 1 ◦C. The
DMPC liposome solutions were run at 150.9 MHz (13C) at 45 ◦C,
above the phase transition temperature (TC) of DMPC (Mabrey and
Sturtevant, 1978; Zachariasse et al., 1981). Below this temperature,
the mobility of the intercalated molecules is low and the resulting
NMR absorptions are very broad. Raising the temperature sharpens
the peaks but does not seem to affect the chemical shifts. The chem-
ical shifts were measured relative to the solvent, except in the case
of aqueous vesicle solutions in which we calibrated the spectrum
according to the trimethylammonium peak of choline at 54.6 ppm.
2.2. Preparation of esters
Diesters 1 and 2 were prepared as previously described (Cohen
et al., 2008). 2-Methylbutyl acetate (3a), 2-ethylbutyl acetate (3b)
and 2-ethylhexyl acetate (3c) are commercially available, but we
have synthesized them all with and without 13C-labeling in the
carbonyl. Most of the acetoxy derivatives are known (Shaojun et al.,
1998; Xiaoli et al., 1998), but some of the spectral data are lacking
and, hence, the complete data are given below. The acetoxy deriva-
tives were prepared with acetyl chloride following the general
∗
Corresponding author. Tel.: +972 3 5318610; fax: +972 3 7384053.
E-mail address: frimea@mail.biu.ac.il (A.A. Frimer).
0009-3084/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.chemphyslip.2008.07.005

Y. Cohen et al. / Chemistry and Physics of Lipids 155 (2008) 120–125
121
2.2.3. 2-Ethylhexyl acetate (3c)
1H NMR (300 MHz, CDCl3) ı 3.98 (2H, dd, J = 6, 1.5 Hz, H–1), 2.04
(3H, s, H–1ꢀꢀ), 1.55 (1H, septet, J = 6 Hz, H–2), 1.35 (4H, m, H–5 and
H–1ꢀ), 1.28 (4H, m, H–3 and H–4), 0.89 (3H, t, J = 7.5 Hz, H–2ꢀ or
H–6), 0.88 (3H, t, J = 7.5 Hz, H–6 or H–2ꢀ); 13C NMR (75 MHz, CDCl3)
ı 171.4 (C O), 66.9 (C–1), 38.7 (C–2), 30.4 (C–4), 28.9 (C–3), 23.7
(C–1ꢀ), 22.9 (C–5), 21.0 (C–1ꢀꢀ), 14.0 (C–6), 10.9 (C–2ꢀ).
2.2.4. 2-Butyloctyl acetate (3d)
1H NMR (300 MHz, CDCl3) ı 3.96 (2H, d, J = 7 Hz, CH2), 2.05 (3H,
s, H–1ꢀꢀ), 1.65–1.55 (1H, bs, CH), 1.35–1.10 (16H, m, methylenes),
0.90 (3H, t, J = 7 Hz, H–4ꢀ or H–8), 0.88 (3H, t, J = 7 Hz, H–8 or H–4ꢀ);
13C NMR (75 MHz, CDCl3) ı 171.4 (C O), 67.3 (CH2), 37.2 (CH), 31.8,
31.2, 30.9, 29.6, 28.9, 26.6, 23.0, 22.6 (methylenes), 21.0 (C–1ꢀꢀ), 14.0
(CH3).
Scheme 1. Derivatives of diesters 1 and 2.
2.2.5. 2-Hexyldecyl acetate (3e)
1H NMR (200 MHz, CDCl3) ı 3.96 (2H, d, J = 6 Hz, CH2), 2.05 (3H, s,
H–2ꢀꢀ), 1.61 (1H, br-s, CH), 1.37–1.16 (24H, m, methylenes and CH),
0.88 (6H, t, J = 6 Hz, H–10 and H–6ꢀ); 13C NMR (75 MHz, CDCl3) ı
171.4 (C O), 67.4 (CH2), 37.2 (CH), 31.9, 31.8, 31.2, 29.9, 29.6, 29.5,
29.3, 26., 26.6, 22.7 (methylenes), 21.0 (C–1ꢀꢀ), 14.1 (CH3); MS (CI,
CH4) 285 (MH+, 1%), 225 (MH+–CH3COOH, 60%); HRMS (CI, CH4,
MH+) m/z calcd. for C18 H37O2 285.2794; found 285.2751.
procedure for the esterification of alcohols as previously described
(Wheeler, 1963). An analog enriched with 13C at the ester carbonyl
was prepared using acetyl-1-13C chloride, and diluted with the 12
C
compound in a 1:1 ratio. 1H and 13C NMR analysis revealed that all
the esters synthesized were essentially pure.
In the spectral data, the carbons and attached hydrogens were
numbered (as shown below) to allow convenient comparison of
the spectral data of the various derivatives, and not necessarily
according to the rules of nomenclature. In the 13C NMR spectra
of the long-chain derivatives of 3, there are chemical shifts, which
overlap; hence, the number of the chemical shifts does not always
match the number of the carbons. On the other hand, carbon C–2
in compounds 3a and 3c–g is chiral, with the adjacent methylene
hydrogens rendered diastereotopic. This of course results in addi-
tional couplings and peaks. The numbering of the various carbons
in symmetric 3b versus that of chiral 3a and 3c–g is exemplified in
Scheme 2.
2.2.6. 2-Octyldodecyl acetate (3f)
1H NMR (300 MHz, CDCl3) ı 3.96 (2H, d, J = 6 Hz, CH2), 2.05 (3H,
s, H–1ꢀꢀ), 1.61 (1H, br-s, CH), 1.35–1.15 (32H, m, methylenes), 0.88
(6H, t, J = 6.5 Hz, H–12 and H–8ꢀ); 13C NMR (50.9 MHz, CDCl3) ı 171.2
(C O), 67.4 (CH2), 37.3 (CH), 31.9, 31.3, 29.9, 29.6, 29.6, 29.3, 26.7,
22.7 (methylenes), 20.9 (C–1ꢀꢀ), 14.1 (CH3); MS (CI) 341 (MH+, 2%),
281 (MH+–CH3COOH, 100%); HRMS (CI, CH4, MH+) m/z calcd for
C22H45O2, 341.3419; found 341.3415.
2.2.7. 2-Decyltetradecyl acetate (3g)
1H NMR (300 MHz, CDCl3) ı 3.962 (2H, d, J = 6 Hz, CH2), 2.043
(3H, s, H–1ꢀꢀ), 1.6 (1H, br-s, CH), 1.45–1.25 (40H, m, methylenes),
0.881 (6H, t, J = 6.5 Hz, H–14 and H–10ꢀ); 13C NMR (75 MHz, CDCl3) ı
171.5 (C O), 67.5 (CH2), 37.4 (CH), 32.1, 31.4, 30.2, 30.1, 29.80, 29.74,
29.5, 26.8, 22.8 (methylenes), 21.1 (C–1ꢀꢀ), 14.2 (CH3); MS (CI) 397
(MH+, 9%), 337 (MH+–CH3COOH, 100%); HRMS (CI, CH4, MH+) m/z
calcd. for C26H53O2, 397.4045; found 397.3990.
2.2.1. 2-Methylbutyl acetate (3a)
1H NMR (300 MHz, CDCl3) ı 3.93 (1 H, ABX system, dd, J = 10.5,
6.5 Hz, H–1a), 3.84 (1H, ABX system, dd, J = 10.5, 6Hz, H–1b), 2.03
(3H, s, H–1ꢀꢀ), 1.71 (1H, qtdd, J = 7, 6.5, 6, 5.5 Hz, H–2), 1.42 (1H, dqd,
J = 15, 7.3, 5.5 Hz, H–3a), 1.15 (1H, dqd, J = 15, 7.3, 6.5 Hz, H–3b), 0.90
(3H, d, J = 7 Hz, H–1ꢀ), 0.88 (3H, t, J = 7.5 Hz, H–4); 13C NMR (75 MHz,
CDCl3) ı 171.4 (C O), 69.3 (C–1), 34.2 (C–2), 26.1 (C–3), 21.1 (1ꢀꢀ),
16.5 (C–1ꢀ), 11.3 (C–4); MS (CI, CH4) 131.109 (MH+, 14%); HRMS (CI,
CH4, MH+) m/z calcd. for C7H15O2 131.1072; found 131.1093.
2.3. General procedure for preparation of DMPC liposomal
solutions
As previously described (Afri et al., 2002), all glassware was first
rinsed with conc. HCl to remove all traces of detergents and then
with doubly distilled water. In a typical liposome preparation, the
compounds to be intercalated (dubbed “intercalants”) and DMPC
in a molar ratio of 1:5 were dissolved in chloroform in a vial. The
solvent was then evaporated with a gentle steam of N2 while rotat-
2.2.2. 2-Ethylbutyl acetate (3b)
1H NMR (300 MHz, CDCl3) ı 3.96 (2H, d, J = 6 Hz, H–1), 2.01 (3H,
s, H–1ꢀ), 1.46 (1H, septet, J = 6 Hz, H–2), 1.32 (4H, quintet, J = 7 Hz,
H–3), 0.86 (6H, t, J = 7 Hz, H–4); 13C NMR (75 MHz, CDCl3) ı 171.3
(C O), 66.5 (C–1), 40.2 (C–2), 23.2 (C–3), 20.9 (C–1ꢀ), 10.9 (C–4).
Scheme 2. The NMR numbering of carbons in symmetrical ester 3b versus that of chiral analog 3c.

122
Y. Cohen et al. / Chemistry and Physics of Lipids 155 (2008) 120–125
ing in the palm of the hand, leaving a uniformly thin layer of lipid
on the walls of the vial. The vial was finally set in a cotton-packed
RB flask and the solvent was removed by rotary evaporation, leav-
ing a uniformly thin film layer of lipid on the walls of the vial. The
vial was then charged with 0.1 M phosphate buffer with 10% D2O
(pH 7.8) yielding a cloudy solution in the exact desired molar con-
centration. The lipid solution was vortexed for 10 min to obtain
multilamellar liposomes. The liposomal solution was then probe
sonicated until a clear solution of essentially unilamellar liposome
was obtained (generally 3–4 min). Sonication was effected with a
MSE Titanium Probe Ultrasonic Disintegrator model MK2 at 20 kHz
output frequency. The final ester concentration was 0.05 M and the
intercalant:lipid molar ratio was 1:5. The liposomes were character-
ized in previous cryo-TEM work (Afri et al., 2004), which confirms
that these conditions produce primarily unilamellar liposomes.
3. Results and discussion
As noted above, in the 13C NMR spectra of the very long-chain
derivatives 1e–g and 2e–g, where the total number of carbons in
the ester group “R” was 16 or above, we observed doubling of the
carbonyl peaks (see Table 1, entries 1, 2, 4, 7 and 8 and Fig. 1).
One pair was sharp, absorbing at a relatively higher field (e.g.,
in 1g, they appeared at 170.58 and 170.98 ppm), while the other
pair was broad and situated at lower field (e.g., in 1g, they were
observed at 171.60 and 172.49 ppm). This doubling prompted the
hypothesis that in these long-chain compounds there are two popu-
lations: one includes those molecules which are intercalated within
the liposome and feel the polarity corresponding to the liposo-
mal microenvironment; the other consists of aggregates (Bird et al.,
1996; Buffy et al., 2003; D’Emiliano et al., 2007; Dietrich et al., 2001;
Katsikas and Quinn, 1981; Munro, 2003 and Tulenko et al., 1998)
of these long-chain derivatives. In these aggregates, the diesters
feel the microenvironment polarity created by the molecules them-
selves, and should be essentially phospholipid independent.
In order to explore this suggestion further, we intercalated
derivatives 1f, 1g, 2f and 2g into the liposomal bilayer in various
diester:lipid (DMPC) molar (n/n) ratios. When we increased the
diester concentration from the usual 1:5 to 5:1 (Table 1, entries 2
vs. 3, 4 vs. 5, 8 vs. 11 and 12 vs. 13), in each of these four molecules
we observed only one pair of carbonyl peaks (in addition to the
liposomal ester carbonyls). These were nearly identical in shape
and chemical shift with the sharp upfield pair peaks observed in
the 1:5 cases. Moreover, when 1g and 2g were dispersed in water,
Fig. 1. 13 C NMR spectrum of 1g: (a) within DMPC liposomes prepared in an
diester:lipid molar ratio of 1:5; (b) within DMPC Liposomes prepared in an
diester:lipid molar ratio of 5:1; (c) dispersed in water.
without any DMPC (entries 6 and 13), the results were the same
as observed in the 5:1 case. In the instance of diester 2f, we also
reduced the diester concentration from the usual 1:5 to 1:10 and
1:15 (entry 8 vs. entries 9 and 10); here, the two pairs of carbonyl
peaks remained, but their relative heights changed substantially. In
particular, as we reduced the total concentration of the substrate in
Table 1
Chemical shift of carbonyls I and II of diesters 1e–g and 2e–g within DMPC liposomes in various molar ratios and dispersed in water
Entry
Compound
Diester:DMPC (n/n)
Chemical shift (ppm)
Sharp upfield carbonyl peaks
Broad downfield carbonyl peaks
C–I
C–II
C–I
C–II
1
2
3
4
5
6
7
8
1e
1f
1f
1g
1g
1g
2e
2f
2f
2f
2f
2g
2g
2g
1:5
1:5
5:1
1:5
171.12
171.06
171.10
170.98
171.01
170.98
172.42
172.37
172.34
172.35
172.46
172.24
172.56
172.36
170.71
170.65
170.70
170.58
170.60
170.57
172.06
172.00
171.97
171.97
172.12
171.86
171.89
171.99
172.54
172.57
171.38
171.54
172.49
171.60
5:1
Dispersed in water w/o DMPC
1:5
173.67
173.68
173.52
173.49
172.37
172.45
172.62
172.72
1:5
9
1:10a
10
11
12
13
14
1:15a
5:1
1:5
5:1
b
b
b
b
b
b
Dispersed in water w/o DMPC
a
The relative heights of the sharp peaks decreased and the broad peaks increased as we reduced the concentration of the substrate in the liposome.
bThe peaks at low field could not be detected, because they have the same chemical shifts as the carbonyls of the DMPC liposome.

Y. Cohen et al. / Chemistry and Physics of Lipids 155 (2008) 120–125
123
Scheme 3. Preparation of esters 3.
the liposomal solution, the relative heights of the high-field sharp
peaks decreased and the low-field broad peaks increased.
the aggregate surface which feel a water environment are negligible
as compared to the vast majority of ester molecules that lie within
the bulk of the aggregate—surrounded by other ester molecules.
Once an ester lies within the bulk, it makes little difference whether
this is a neat ester sample or an ester lying within an aggregate
in water; hence, they have the same chemical shift. Furthermore,
since the ET(30) polarity felt by the ester in the aggregate is ca.
30–32 kcal/mol (corresponding to hexane), it would seems clear
that the ester molecules are not aligned specifically head-to-head.
Were the latter to be true, then the carbonyls would feel a more
polar environment of carbonyls (similar to ethyl acetate, with an
ET(30) value of ca. 38 kcal/mol). This suggests that, in the aggregate,
the molecules are lying in an alternating fashion, which permits
closest packing—as expected for branched long-chain esters with a
small headgroup and large branched tail (Israelachvili et al., 1980).
To study this aggregation phenomenon a bit more, we prepared
monoester analogs of diesters 1 and 2, namely, ester 3, via the
acetylation of the corresponding commercially available branched
alcohols (Scheme 3). In addition, in the case of the C20 ester 3f,
the carbonyl of the acetyl group was 13C enriched (prepared from
13C enriched acetyl chloride), which allowed us to substantially
reduce the amount of ester within the liposome required for NMR
detection.
We turn now to the chemical shift values of carbonyls C–I
and C–II observed for the lipid intercalated compound and aggre-
gated material. Maciel and Ruben (1963), Maciel and Natterstad
(1965), Menger (1979), Menger et al. (1978, 1988), Ueji and
Makamaura (1976) and Janzen et al. (1989) have reported on
the generally excellent correlation between 13C-chemical shift of
polarizable carbons (e.g., carbonyl groups) and solvent polarity.
There exist a number of polarity parameters in the literature—each
taking into account various different interactions (e.g., van der
Waals interactions, hydrogen bonding and electrostatic forces)
between the solute and solvent. Maciel and Menger used the
dielectric constant ε as their polarity parameter. However, Janzen
and coworkers found that better results were obtained using
Reichardt’s ET(30) parameter (Reichardt, 1965, 1994), and our
extensive experience confirms Janzen’s observations (Cohen et al.,
2008). The successful applicability of Reichardt’s parameter in our
NMR systems should not be surprising. After all, ET(30) values are
based on the change in the long wavelength visible absorption of
the dissolved “Reichardt’s dye” [2,4-diphenyl-4-(2,4,6-triphenyl-
1-pyridinio)phenolate] which reflects the development of charge
separation with increasing solvent polarity. This is perfectly anal-
ogous to the development of charge separation of polarizable
carbons (e.g., carbonyl groups) with solvent polarity studied in the
NMR technique.
As shown in Table 2, when these esters were intercalated within
the liposomal bilayer, in an ester:lipid molar ratio of 1:5, only one
carbonyl peak was observed in the NMR spectrum of the C5–C16
derivatives 3a–3e (Table 2, entries 1–5). However, as observed
above for diesters 1 and 2, in the very long C20 and C24 derivatives
3f and 3g (Table 2, entries 6 and 7), we observed double carbonyl
peaks. We note that Bird et al. (1996) similarly report carbonyl dou-
bling during the intercalation of 30 mol% ethyl oleate within small
unilamellar vesicles prepared from ovalecithin.
As before, we focused on the 13C-labeled carbonyl of the C20
analog 3f, and varied the ratio between ester and the DMPC (see
Table 3 and Fig. 2). In Table 3, we see that under the usual condi-
tions of a 1:5 ester:DMPC molar ratio (Table 3, entry 4), two carbonyl
peaks were observed at 170.49 and 169.16 ppm, with stronger and
weaker relative intensities, respectively. When this molar ratio is
reversed to 5:1 (Table 3, entry 3), the chemical shifts remained
essentially unchanged, but the relative intensities were exchanged.
As in the case of diesters 1 and 2 discussed above, we attribute
In a companion paper (Cohen et al., 2008), we studied this cor-
relation in the case of compounds 1 and 2, dissolved in various
pure solvents and intercalated within liposomes. The chemical shift
values observed in Table 1 indicate that carbonyls C–I and C–II of
diesters 1f–g and 2f–g are experiencing similar micropolarities. In
addition, the chemical shifts reported in Table 1 entries 3 and 5
(diester:lipid (DMPC) molar (n/n) ratios of 5:1 for 1f and 1g, respec-
tively) correlate with an ET(30) values of ca. 30–32 kcal/mol, which
is a very non-polar microenvironment similar to that observed
for alkanes (Reichardt’s ET(30) value for hexane is 31.0 kcal/mol,
Reichardt, 1965, 1994). The above data is indeed what we would
expect for aggregates, with the carbonyls finding themselves sur-
rounded by long-chain fatty acid groups.
Table 2
Chemical shifts of the carbonyl of methyl esters (3) in DMPC liposomes
Entry
Compound
Carbons in
alcohol 4
Chemical shift of
carbonyl(s) (ppm)
1
2
3
4
5
6
7
3a
3b
3c
3d
3e
3f
5
6
8
12
16
20
24
171.49
171.23
171.00
170.64
170.44
One further comment: we unfortunately have little information
about the exact structure of the aggregates or how many molecules
are involved. This is certainly worthy of future in-depth investi-
gation (Lundberg et al., 2007). We believe, however, that these
aggregates are large collections of molecules. Thus, the esters on
170.23
170.39
169.16
169.06
3g

124
Y. Cohen et al. / Chemistry and Physics of Lipids 155 (2008) 120–125
Table 3
Chemical shifts of the carbonyl of ester 3f in DMPC liposomes in various concentrations and molar ratios of ester:DMPC
Entry
Ester 3f:DMPC (n/n)
[M]
Chemical shift (relative intensitya)
Sharp upfield carbonyl peak
Broad downfield carbonyl peak
1
2
3
4
5
6
7
8
Neat
168.85
169.55
169.33 (strong)
169.16 (weak)
169.20 (weak)
Dispersed in water w/o DMPC
5:1
1:5
1:10
1:20
1:20
1:40
0.05
0.05
0.01
0.01
0.005
0.005
170.87 (very weak)
170.49 (strong)
170.72 (strong)
170.81
170.69
170.82
a
Relative intensities based on peak integration.
the sharp upfield carbonyl peak to aggregated ester molecules out-
side the liposomal bilayer, while the broader downfield carbonyl
corresponds to ester molecules intercalated within the bilayer and
surrounded by liposomal material. Clearly the latter should pre-
dominate when the ester:DMPC molar ratio is low (1:5), while
aggregates should be more prominent when the ester:DMPC molar
ratio is high (5:1).
Confirmation of the assignment of the sharp upfield peak at
169 ppm to aggregated molecules comes from a spectrum of a sam-
ple of the C20 analog 3f dispersed in water (also presumably as
aggregates) in the absence of DMPC (Table 3, entry 2). The latter
showed single sharp upfield carbonyl absorption at 169.55.
One fundamental question still requires elucidation: where
exactly are the aggregates located, in the lipid phase of the lipo-
some, the aqueous phase, or both? To this end, monoester 3f was
intercalated within DMPC liposomes in an ester:lipid molar ratio of
1:5. The liposomes were centrifuged down (25,000 × g for 15 min)
to a lipid pellet, the supernatant liquid was decanted and replaced
by buffer, and the pellet was redispersed by vortexing the sample.
NMR spectra of the starting and final liposomal and supernatant
solutions were then run. These spectra revealed the presence of
aggregate and liposome intercalated peaks in each case, but in very
different ratios. Indeed, the ratio of aggregated substrate to lipid
intercalated material is 3:1 in the starting liposomal solution; 10:1
in the supernatant solution; and 1:1 in the redispersed liposomal
solution. This piece of evidence suggests that, at least the majority
of the aggregate, lies in the aqueous phase. However, we believe that
the aggregate observed in the redispersed liposomes lies primarily,
if not completely, in the internal water of the liposomes.
In addition, we have noted above (see Tables 1 and 3), that sub-
strates 1g, 2g and 3f dispersed in water have the same chemical
shift as that of aggregated material. By contrast a neat sample of
3f (Table 3, entry 1) – arguably analogous to an aggregate of 3f
in a non-polar organic solvent – showed single sharp upfield car-
bonyl absorption at 168.85 ppm. In addition, when we lowered the
ester:DMPC ratio to below 1:10 (entries 6–8), the upfield carbonyl
disappeared completely.
We suggest, therefore, that in this case the aggregates form and
reside in the aqueous phase when the lipid becomes saturated with
substrate. Based on the data in Tables 1–3, this presumably occurs
when the substrate:lipid molar ratio rises above ∼1:20. In this light,
predominance of aggregates in the extra-liposomal supernatant
solution is to be expected. The fact that liposome intercalated peaks
are observed at all in the supernatant solution results from resid-
ual liposomes that were not pelleted; indeed, DMPC lipid peaks are
also evident in the NMR spectrum of this supernatant solution. By
contrast, in the spectrum of the redispersed liposomal solution, the
ratio of aggregate to lipid intercalated material is ca. 1:1. We suggest
that the aggregates observed in the latter instance are located in the
encapsulated internal water of the liposome, which is not removed
upon pellet formation and redispersal (Katsikas and Quinn, 1981).
4. Conclusion
In conclusion, these studies suggest that when the substrate
to lipid molar ratio rises above ∼1:20, the lipid becomes sat-
urated with substrate. In the case of long-chain esters, the
non-incorporated substrate forms aggregates in the aqueous phase.
The varying environments and chemical shifts of the ester carbonyls
can be detected by 13C NMR.
Fig. 2. NMR spectra 13 C-labeled carbonyl of 3f in DMPC liposomes prepared under
various conditions: (a) 1:5 molar ratio of 3f:DMPC; (b) 5:1 molar ratio of 3f:DMPC;
(c) neat sample; (d) 1:20 molar ratio of 3f:DMPC.

Y. Cohen et al. / Chemistry and Physics of Lipids 155 (2008) 120–125
125
Acknowledgements
Katsikas, H., Quinn, P.J., 1981. The interaction of coenzyme Q with dipalmitoylphos-
phatidylcholine bilayers. FEBS Lett. 133, 230–234.
Lundberg, D., Unga Johan Galloway, A.L., Menger, F.M., 2007. Studies on an ester-
modified cationic amphiphile in aqueous systems: behavior of binary solutions
and ternary mixtures with conventional surfactants. Langmuir 23, 11434–11442.
Mabrey, S., Sturtevant, J.M., 1978. High-sensitivity differential scanning calorime-
try in the study of biomembranes and related model systems. In: Korn, E.D.D.
(Ed.), Methods in Membrane Biology, vol. 9, 1st ed. Plenum Press, New York, pp.
237–274.
We acknowledge the kind and generous support of the Israel
Science Foundation (Grants Number 327/02 and 437/06) founded
by The Israel Academy of Sciences and Humanities, and The Ethel
and David Resnick Chair in Active Oxygen Chemistry.
Maciel, G.E., Natterstad, J.J., 1965. Carbon-13 chemical shifts of the carbonyl group
III. Solvent effects. J. Chem. Phys., 2752–2759.
References
Maciel, G.E., Ruben, G.C., 1963. Solvent effects on the 13 C chemical shift of the car-
bonyl group of acetone. J. Am. Chem. Soc. 85, 3903–3904.
Afri, M., Ehrenberg, B., Talmon, Y., Schmidt, J., Cohen, Y., Frimer, A.A., 2004. Active oxy-
gen chemistry within the liposomal bilayer, Part III: Locating vitamin E, ubiquinol
and ubiquinone and their derivatives in the lipid bilayer. Chem. Phys. Lipids 131,
107–121.
Afri, M., Gottlieb, H.E., Frimer, A.A., 2002. Superoxide organic chemistry within the
liposomal bilayer, Part II: A correlation between location and chemistry. Free
Radic. Biol. Med. 32, 605–618.
Bird, D.A., Laposata, M., Hamilton, J.A., 1996. Binding of ethyl oleate to low density
lipoprotein, phospholipid vesicles, and albumin: a 13 C NMR study. J. Lipid Res.
37, 1449–1458.
Buffy, J.J., Waring, A.J., Lehrer, R.I., Hong, M., 2003. Immobilization and aggregation of
the antimicrobial peptide protegrin-1 in lipid bilayers investigated by solid-state
NMR. Biochemistry 42, 13725–13734.
Cohen, Y., Bodner, E., Richman, M., Afri, M., Frimer, A.A., 2008. NMR-based molecular
ruler for determining the depth of intercalants within the lipid bilayer Part I:
discovering the guidelines. Chem. Phys. Lipids 155, 98–113.
D’Emiliano, D., Casieri, C., Paci, M., De Luca, F., 2007. Detection of ganglio-
side clustering in DOPC bilayers by 1H NMR spectroscopy. Physica A 374,
293–303.
Dietrich, C., Bagatolli, L.A., Volovyk, Z.N., Thompson, N.L., Levi, M., Jacobson, K.,
Gratton, E., 2001. Lipid rafts reconstituted in model membranes. Biophys. J. 80,
1417–1428.
Israelachvili, J.N., Marcelja, S., Horn, R.G., 1980. Physical principles of membrane
organization. Quart. Rev. Biophys. 13 (2), 121–200.
Janzen, E.G., Haire, D.L., Coulter, G.A., Stronks, H.J., Krygsman, P.H., Towner, R.A.,
Hilborn, J.W., 1989. Locating spin traps in heterogeneous media by carbon-13
NMR spectroscopy. Investigations in SDS micelles, DMPC vesicles, and rat liver
microsomes. J. Org. Chem. 54, 2915–2920.
Menger, F.M., 1979. The structure of micelles. Acc. Chem. Res. 12, 111–117.
Menger, F.M., Aikens, P., Wood Jr., M., 1988. Water content of phospholipid bilayers.
J. Chem. Soc., Chem. Commun., 180–182.
Menger, F.M., Jerkunica, J.M., Johnston, J.C., 1978. The water content of a micelle
interior. The fjord vs. reef models. J. Am. Chem. Soc. 100, 4676–4678.
Munro, S., 2003. Lipid rafts: elusive or illusive? Cell 115, 377–388.
Reichardt, C., 1965. Empirical parameters of solvent polarity. Angew. Chem. 77,
30–40.
Reichardt, C., 1994. Solvatochromic dyes as solvent polarity indicators. Chem. Rev.
94, 2319–2358.
Shaojun, W., Xiaoli, D., Qingda, A., Bin, Z., Yan, W., 1998. Research on catalyst of syn-
thetic Guerbet alcohol acetate. People Rep. China. Jingxi Huagong 15, 206–208.
Tulenko, T.N., Chen, M., Mason, P.E., Mason, R.P., 1998. Physical effects of choles-
terol on arterial smooth muscle membranes: evidence of immiscible cholesterol
domains and alterations in bilayer width during atherogenesis. J. Lipid Res. 39,
947–956.
Ueji, S., Makamaura, M., 1976. The effects of solvent on carbon-13 chemical shifts of
the carbonyl system. Tetrahedron Lett., 2549–2552.
Wheeler, T.S., 1963. Flavone (method 1). Org. Synth. Coll. 4, 478–481.
Xiaoli, D., Shaojun, W., Qingda, A., Zhonglin, H., 1998. The synthesis of the Guerbet
alcohol acetate and study of it’s wave spectrum. People Rep. China Guangpu
Shiyanshi 15, 20–22.
Zachariasse, K.A., Van Phuc, N., Kozankiewicz, B., 1981. Investigation of
micelles, microemulsions, and phospholipid bilayers with the pyridinium-N-
phenolbetaine ET(30), a polarity probe for aqueous interfaces. J. Phys. Chem. 85,
2676–2683.