NMR-based molecular ruler for determining the depth of intercalants within the lipid bilayer. Part III: Studies on keto esters and acids

Article abstract of DOI:10.1016/j.chemphyslip.2014.07.007

The development of "molecular rulers" would allow one to quantitatively locate the penetration depth of intercalants within lipid bilayers. To this end, an attempt was made to correlate the ¹³C NMR chemical shift of polarizable "reporter" carbons (e.g., carbonyls) of intercalants within DMPC liposomal bilayers - with the polarity it experiences, and with its Angstrom distance from the interface. This requires families of molecules with two "reporter carbons" separated by a known distance, residing at various depths/polarities within the bilayer. For this purpose, two homologous series of dicarbonyl compounds, methyl n-oxooctadecanoates and the corresponding n-oxooctadecanoic acids (n = 4-16), were synthesized. To assist in assignment and detection several homologs in each system were prepared ¹³C-enriched in both carbonyls. Within each family, the number of carbons and functional groups remains the same, with the only difference being the location of the second ketone carbonyl along the fatty acid chain. Surprisingly, the head groups within each family are not anchored near the lipid-water interface, nor are they even all located at the same depth. Nevertheless, using an iterative best fit analysis of the data points enables one to obtain an exponential curve. The latter gives substantial insight into the correlation between polarity (measured in terms of the Reichardt polarity parameter, E_T(30)) and penetration depth into the liposomal bilayer. Still missing from this curve are data points in the moderate polarity range.

Full text of DOI:10.1016/j.chemphyslip.2014.07.007

Chemistry and Physics of Lipids 184 (2014) 105–118
Contents lists available at ScienceDirect
Chemistry and Physics of Lipids
journal homepage: www.elsevier.com/locate/chemphyslip
NMR-based molecular ruler for determining the depth of intercalants
within the lipid bilayer
Part III: Studies on keto esters and acids
Michal Afri, Carmit Alexenberg, Pinchas Aped, Efrat Bodner, Sarit Cohen,
Michal Ejgenburg, Shlomi Eliyahu, Pessia Gilinsky-Sharon, Yifat Harel,
*
Miriam E. Naqqash, Hani Porat, Ayala Ranz, Aryeh A. Frimer
The Department of Chemistry, Bar-Ilan University, Ramat Gan 5290002, Israel
A R T I C L E I N F O
A B S T R A C T
Article history:
The development of “molecular rulers” would allow one to quantitatively locate the penetration depth of
intercalantswithinlipidbilayers.Tothisend,anattemptwasmadetocorrelatethe13CNMRchemicalshiftof
polarizable “reporter” carbons (e.g., carbonyls) of intercalants within DMPC liposomal bilayers – with the
polarityitexperiences,andwithitsAngstromdistancefromtheinterface.Thisrequiresfamiliesofmolecules
with two “reportercarbons” separated by a known distance, residing at various depths/polarities within the
bilayer. For this purpose, two homologous series of dicarbonyl compounds, methyl n-oxooctadecanoates
and the corresponding n-oxooctadecanoic acids (n = 4–16), were synthesized. To assist in assignment and
detection several homologs in each system were prepared 13C-enriched in both carbonyls. Within each
family, the number of carbons and functional groups remains the same, with the only difference being the
location of thesecond ketone carbonyl along thefattyacidchain. Surprisingly, the headgroupswithineach
family are not anchored near the lipid–water interface, nor are they even all located at the same depth.
Nevertheless, using an iterative best fit analysis of the data points enables one to obtain an exponential
curve. The latter gives substantial insight into the correlation between polarity (measured in terms of the
Reichardt polarity parameter, ET(30)) and penetration depth into the liposomal bilayer. Still missing from
this curve are data points in the moderate polarity range.
Received 9 March 2014
Received in revised form 10 July 2014
Accepted 21 July 2014
Available online 24 July 2014
Keywords:
Liposome
13C NMR
ET
DMPC
Oxooctadecanoates
Molecular ruler
ã 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction
interactions. There are number of spectroscopic techniques that
have been used to study the intercalation of various molecules and
The biological membrane is the outer envelope of the living
eukaryotic cell and protects the cell by modulating the crossing of
chemicals into and out of the cell. Determining the location of
molecules within biological membranes is important for a variety
reasons. Inter alia, it allows one to understand the connection
between the depth/orientation of a substrate within a membrane
and its reactivity. Secondly, it is helpful in gaining insight into the
structure and dynamics of various lipids and protein–lipid
peptides into biomembranes or membrane-models. These include
solid-state (Hong, 2006) or solution (Prosser et al., 2007; Evanics
et al., 2006) NMR, ATR-FTIR (Mourelatou et al., 2011), fluorescence
quenching (Postupalenko et al., 2011; Herenyi et al., 2009; Caputo
and London, 2003; Menger et al., 2002; Huster et al., 2001) and EPR
spectroscopy (Altenbach et al., 1994; Fanucci and Cafiso, 2006;
Frazier et al., 2002; Hubbell and Altenbach, 1994; Nielsen et al.,
2005).
Despite, or perhaps, because of its relative lack of complexity,
the liposome is often used as a simple model for biological
membranes (Gregoriadis, 1984; Papahadjopoulos, 1978). For
decades, biophysicists have used for this purpose DMPC (dimyr-
istoyl phosphatidylcholine, di-14:0-PC), a lipid with saturated
myristoyl (14:0) chains at the sn-1 and sn-2 positions. Its
popularity as a model membrane system can be attributed to
the fact that it is stable, inexpensive, and easy to obtain. Because it
has fully saturated acyl chains, DMPC bilayers are robust and do not
require special preparatory environments. Over the years, DMPC
bilayers have been well characterized using a variety of physical
* Corresponding author at: The Ethel and David Resnick Chair in Active Oxygen
Chemistry, The Department of Chemistry, Bar-Ilan University, Ramat Gan 5290002,
Israel. Tel.: +972 3 5318610; fax: +972 3 7384053.
E-mail addresses: Michal.Afri@biu.ac.il (M. Afri), C_Eliasi@walla.com
(C. Alexenberg), Pinchas.Aped@biu.ac.il (P. Aped), Efrat.Bodner@biu.ac.il (E. Bodner)
, SaritCohen84@gmail.com (S. Cohen), MichalDomb@gmail.com (M. Ejgenburg),
ShlomiEliyohu@gmail.com (S. Eliyahu), Pessia.Sharon@biu.ac.il (P. Gilinsky-Sharon)
, IfatAsh@Yahoo.com (Y. Harel), MiriamNaqqash@gmail.com (M.E. Naqqash),
HaniPorat@gmail.com (H. Porat), Ayala.Ranz@gmail.com (A. Ranz),
Aryeh.Frimer@biu.ac.il (A.A. Frimer).
http://dx.doi.org/10.1016/j.chemphyslip.2014.07.007
0009-3084/ã 2014 Elsevier Ireland Ltd. All rights reserved.

106
M. Afri et al. / Chemistry and Physics of Lipids 184 (2014) 105–118
techniques (Kuc9erka et al., 2011; Jämbeck and Lyubartsev, 2012;
Aussenac et al., 2003), and as a result, they have been used to study
mutatis mutandis) have two resonance forms (Fig. 2), the non-polar
double-bondedformandthecharge-separatedzwitterionicform. As
the solvent polarity increases, the zwitterionic form – with lower
electron density on the carbonyl carbon – makes a concommitantly
greater contribution, and the NMR chemical shift of the carbonyl
carbons increases, i.e., moves downfield.
problems
ranging
from
drug–membrane
interactions
(Komljenovic et al., 2010), to the enhancement of cholesterol
flipflop (Kuc9erka et al., 2010), to the difference between bulk water
and interfacial polarity and pH (Voinov et al., 2009), to examining
the location, behavior and antioxidant properties of
a
-tocopherol
Due to the polarity gradient, the chemical shift values of a polar
carbon within an intercalant will differ depending on its depth
within the bilayer. Using the aforementioned correlation between
in biomimetic membranes (Takahashi et al., 1989; Serbinova et al.,
1991; Wassall et al., 1991; Afri et al., 2004a; Fukuzawa, 2008).
Despite its broad use, there have been some authors who have
questioned DMPC’s biological relevance due to its oxidative
stability (Marquardt et al., 2014) and the low level of 14:0 fatty
acid-containing phospholipids in biological systems (e.g., <1% in
lipoproteins) (Ruíz-Gutiérrez et al., 1993).
the 13C chemical shifts and solvent polarity, one can prepare a “
d/ET
(30) correlation graph” for each of the polarizable “indicator”
carbons in the molecule. When the molecule is then intercalated
into the liposomal bilayer, the chemical shift measured reflects the
micropolarity that carbon is experiencing. (Afri et al. (2002) have
previously verified that with an intercalant concentration of 0.05 M
and a intercalant:lipid molar ratio of 1:5, the intercalant itself has
only a negligible affect on the polarity gradient. In addition, cryo-
TEM studies confirm that under these conditions the liposomes
remain essentially unchanged (Afri et al., 2004a; Cohen et al.,
2012).) Interpolating backwards from the correlation graph,
converts the chemical shift value back into a polarity values,
which allows us to approximate the extent of intercalant
penetration into the bilayer. (The error in the ET(30) value is ca.
ꢁ1.0 kcal/mol (Cohen et al., 2008a).)
In light of the above wide application of DMPC liposomes,
developing convenient and readily available techniques for
determining the depth (i.e., the distance from the water–lipid
interface) of substrates intercalated (henceforth “intercalants”)
within the DMPC liposomal bilayer seemed of great value. In
previous work, we have described an NMR method for this purpose
which is based on two observations (Frimer et al., 1996; Afri et al.,
1992, 2004a,b; Cohen et al., 2008a,b,c; Shachan-Tov et al., 2010).
Firstly, there is a generally excellent linear positive correlation
between solvent polarity and the 13C chemical shift (
for a polarizable carbon (e.g., carbonyls). Perhaps the most famous
polarity measure is the dielectric constant (March, 1992) which
d) observed
It should be emphasized that these penetration values are
averages reflective of a dynamic residency of the probe molecule or
intercalant. The timescale of molecular motion is faster than that of
the NMR experiment; the NMR chemical shift data will reflect
polarity due to electrostatic fields from the time-averaged structure,
as well as from water penetration and perhaps partial exit of the
surfactants. Despite this caveat, we believe that this 13C NMR
chemical-shift/polarity correlation technique is a useful tool for
approximating the location of substrates within lipid bilayers.
Perhaps the major strength of the “NMR-technique” for locating
intercalants in a bilayer is that it allows focusing on a specific atom,
e.g., the carbon of a carbonyl (CQO) or nitonyl (CQN–R). Other
techniques (like UV or fluorescence) require much lower intercalant
concentrations,butfocusonlargechromophores;hence,positioning
of the moiety can be tricky. The NMR concentration issue can be
circumvented to some extent by 13C labeling.
The NMR technique has allowed us to approximate the location
of various substrates, such as coumarins (Afri et al., 2002), vitamin
E (Afri et al., 2004a), dichlorofluorescein (Afri et al., 2004b),
ubiquinone (Afri et al., 2004a) and curcumin (Afri et al., 2011),
within DMPC liposomes. But these results are only qualitative in
nature. It would be desirable to “translate” these qualitative ET(30)
values to quantitative Angstrom depths from the interface.
For the purpose of developing a convenient chemical ruler for
DMPC liposomes, we turned to the synthesis (outlined in Fig. 3 and
discussed below in Section 3) of two homologous series of
dicarbonyl compounds on an 18-carbon fatty acid chain: methyl
n-oxodecanoate esters 7, and the corresponding n-oxodecanoic
acids 8. The numbering of the various carbons these dicarbonyl
compounds is exemplified by 7-oxo analogs, ester 7d and acid 8d.
In these bifunctional systems, one carbonyl is that of the acid or
ester moiety, while the second is a ketone carbonyl moiety located
e
expresses the polarizability and the ordering of the dipole of the
solvent itself. This is in contradistinction to Reichardt’s ET(30)
polarity parameter which is based on the interaction between the
solvent and a polarizable solute (Reichardt, 1965, 1994; Reichardt
and Welton, 2011). In particular, it measures the solvent
dependency of the UV absorption (in kcal/mol at 25 ꢀC, 1 atm) of
2,6-diphenyl-4-(2,4,6-triphenyl-N-pyridine)-phenolate, dubbed
Reichardt’s dye (Fig. 1). The more polar the solvent, the shorter
the wavelength and the greater is its ET(30) value.
Gamliel (2005) examined the 13C chemical shifts of the carbonyl
of acetone, and correlated them with either the dielectric constant
e
or ET(30) polarity parameter over a range of 26 solvents.
A substantially better correlation (R2 = 0.80) was obtained using
the ET(30) polarity parameter than with the dielectric constant
(R2 = 0.55), and hence the former was preferred for our future
studies.
In addition to the chemical shift- ET(30) solvent polarity
correlation, the aforementioned NMR technique is based on a
second important observation. A polarity gradient exists within the
lipid membrane, ranging from that of water (ET(30) = 63.1 kcal/mol)
at the water–lipid interface, down to hexane deep in the lipid slab
(ET(30) = 31.0 kcal/mol). The extent of the ET(30)–d correlation is
represented by the square of the correlation coefficient (R2), and
experience has taught that it is generally quite high, R2 > 0.9
(Cohen et al., 2008a).
This solvent polarity–chemical shift correlation can be readily
attributed to the fact that carbonyls (and other polarizable moieties,
Zwitterion
Double-bond
O
C
O
C
R'
R
R'
R
Increasing solvent polarity
Increasing chemical shift
Fig.1. Reichardt’s dye.
Fig. 2. Resonance forms of the carbonyl.

M. Afri et al. / Chemistry and Physics of Lipids 184 (2014) 105–118
107
Fig. 3. General synthesis for keto esters 7 and keto acids 8. (*Position of possible 13C-labeling). (A) K*CN, DMSO, 75 ꢀC, overnight; (B) 40% KOH in EtOH, rt, overnight; (C) MeOH,
H2SO4 (cat.), sonication overnight; (D) Ba(OH)2
8H2O, MeOH or EtOH, rt, overnight; (E) SOCl2, dry CH2Cl2, 30 ꢀC, 1h; (I) KOH 2.5N, EtOH, rt, overnight; (J) n = 4; MeOH, rt, 12h;
ꢂ
n = 5: MeOH and CHCl3, reflux, 5h.
modularly along the chain at carbons 4–16. Based on literature
precedent (Bronshtein et al., 2004), it was expected that the polar
head groups of these bifunctional compounds would be anchored
at or near the lipid–water interface. The keto-groups, on the other
hand, are located at various locations along the lipophilic carbon
chains and sense various polarity environments. Because of van der
Waals force, these long chain compounds presumably lie parallel to
the DMPC lipid chains, and the average Angstrom distance
between the two carbonyls can be readily calculated. This location
can then be correlated with the difference in ET(30) polarity of
these same carbonyls. The data should permit the effective
mapping of changes in polarity as one proceed further and further
down the chain of the liposomal bilayer. The information obtained
should enable one to construct a molecular ruler, complementing
previously reported preliminary work in this field (Cohen et al.,
2008a). We chose an 18-carbon chain ruler so that we could scan
the 14-carbon myristic chain fully and still have a portion of the
carbon beyond the second DMPC carbonyl, to keep the tail aligned
with the lipid chains.
The general synthesis of dicarbonyls 7 and 8, outlined in Fig. 3,
allows the preparation of derivatives a, e and h with the carbonyl
moieties doubly enriched with 13C. This proved to be a great boon
in detection and peak assignment.
2. Materials and methods
2.1. General
Nuclear magnetic resonance (NMR) spectra were recorded on a
variety of Bruker Fourier transform spectrometers: DRX 200 MHz,

108
M. Afri et al. / Chemistry and Physics of Lipids 184 (2014) 105–118
DRX 300 MHz, DMX 600 MHz or Avance III 700 MHz instruments.
For 1D NMR spectra, a QNP probe was used. All 2D experiments
(COSY, HMQC, HMBC, and NOESY) were run and processed by
Bruker software. NMR spectra were generally run at 25 ꢁ1 ꢀC
except in the case of DMPC liposomes, which were run at
physiological temperature 37 ꢁ1 ꢀC, above the phase transition
temperature (Tm) of DMPC which is 23.6 ꢀC (Sassaroli et al., 1990).
Milli-Q columns) containing 1.7 mM NaH2PO4, 8.1 mM Na2HPO4,
2.7 mM KCl, 137 mM NaCl and 1 wt% sodium azide.
2.3. Sonic intercalation of compounds into DMPC liposomal solutions
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 (hence-
forth 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 rotating 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, leaving 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 desired molar concentration which was generally
0.05 M in intercalant. The lipid solution was vortexed for 10 min to
obtain multilamellar liposomes. The liposomal solution was then
sonicated until a clear solution of essentially unilamellar liposome
was obtained. Previous cryo-TEM work (Afri et al., 2004a) confirms
that these conditions produce primarily unilamellar liposomes.
NMR of the liposomal solution indicates that the DMPC and
intercalants remained essentially inert throughout the sonication,
undergoing neither oxidation or hydrolysis.
1H and 13C NMR chemical shifts are expressed in
d (ppm) relative to
TMS (0 ppm). Generally the sample was calibrated by internal TMS,
though sometimes relative to the solvent. In the case of aqueous
liposome solutions, the spectrum was calibrated according to the
trimethylammonium peak at 54.6 ppm. In the 13C NMR spectra of
the long chain derivatives, recorded as ppm relative to TMS, there
are overlap peaks; therefore, the number of the chemical shifts
does not always match the number of carbons. In addition, in the
hydrogen and carbon assignments below, the various carbons were
generally numbered according to the IUPAC nomenclature rules.
Mass spectra (MS) and high resolution mass spectra (HRMS)
were collected in a Waters (UK) AutoSpec Premier instrument in
DCI/CH4 or CI/CH4 mode or by Q TOF in ESI mode. MALDI Mass
spectra were obtained on with a Bruker (Germany) Autoflex III
TOF/TOF machine. Infrared Fourier transform spectra (FTIR) were
obtained using a Bruker Vector 22 instrument, with the data
processed by Opus 5.0 (Bruker) software. The spectra were
analyzed with IR MentorPro 6.5 (Bio-Rad) software. Abbrevia-
tions used for reporting
%
transmittance were: s = strong
To verify that the various intercalants lie within the lipid
bilayer and not in the aqueous phase, the liposomes were
centrifuged down (25,000 ꢃ 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
revealed that the substrate indeed resides within the lipid bilayer
exclusively.
(100–75%), m = medium (75–45%), w = weak (45–25%), v = very
and b = broad. Melting point was run on a Uni-Melt Thomas
Hoover-capillary melting point apparatus manufactured by
Thomas (Philadelphia PA, USA). Sonications were carried out
either in a bath sonicator (EW-08890-06Cole-Parmer1 ultrasonic
cleaner, 1/2 gallon, 230 VAC), or using
a probe sonicator
(Vibra-CellTM VCX130 Sonics and Materials Inc., ultrasonic
processor with 20 kHz output frequency and a titanium alloy
Ti-6Al-4V probe). The former was used for esterification
reactions, while the later was used for liposome formation.
Other standard equipment utilized included a vortex >(Winn
Vortex Genie). The correlation coefficient (R2) for the chemical
shift-solvent polarity correlation graphs were calculated by excel.
Ab initio calculations were performed with Gaussian 03
performed at the B3LYP level using 6–31G* basis set Gaussian,
Inc., Pittsburgh PA, 2003. Molecular modeling calculations were
carried out with PCMODEL version 7.50.00, Serena Software,
Bloomington, Indiana, USA – which uses the MMX force field.
2.4. General synthesis of keto esters 7 and keto acids 8
Many of the compounds described below are commercially
available or have been previously reported. Nevertheless, we
describe the synthesis and report spectral data in those cases
where the 13C-enriched analogs were prepared in order to achieve
sharp clear NMR spectra. Moreover, for many of these compounds,
the complete spectral data (including splitting constants and
assignments – which are far from trivial) are missing in the
literature, or were previously measured with low resolution
instruments. The NMR spectral data and assignments appear in
Tables 1–5 while the complete spectral data including integration
and splitting constants appear in the Supplementary information.
2.2. Chemicals
Acetyl chloride, 1-bromohexane, 1-bromodecane, EDTA, HEPES,
KCN, Mg, dimyristoyl phosphatidylcholine (DMPC), methyl iodide,
octyl bromide, octyl magnesium bromide, adipic acid, palmitic
acid, sebacic acid, suberic acid, succinic acid, thionyl chloride,
trifluoroacetic anhydride, anhydrous chloroform, anhydrous
DMSO and the deuterated solvents were obtained from
Sigma–Aldrich. Bromotetradecane, dibromodecane, dodecyl bro-
mide and Tris–HCl were purchased from ACROS. Dibromohexane
and adipoyl chloride methyl ester were acquired from Alfa Aesar.
Barium hydroxide octahydrate (Fluka), sebacic acid monomethyl
ester (ABCR Chemical Co.), cadmium chloride (Strem Chemicals),
sodium azide (Fisher Scientific) and extra dry diethyl ether (Bio Lab
Co.), were also commercially available. Regarding the 13C-labeled
reagents, acetyl chloride-1-13C, K13CN and succinic anhydride-1
,4-13C2 were obtained from Sigma–Aldrich, octanoic acid-1-13C and
palmitic acid-1-13C were acquired from ISOTEC, and succinic acid-
1,4-13C2 was purchased from Cambridge Isotope Laboratories, Inc.
Phosphate buffer saline solution pH 7.4 (PBS-N3) was prepared
using doubly distilled water (dd H2O – purified via Millipore
2.4.1. General preparation of alkanedinitriles 2
1,n-Dibromoalkane 1 were reacted with KCN (or K13CN) in dry
DMSO following the procedure of Holmes and Lightner (1996);
yielding the corresponding 1,n + 2-alkanedinitriles in generally
excellent yields. The unlabeled analogs are commercially available
(Sigma–Aldrich).
2.4.1.1. Octanedinitrile-1,8–13C2
(
13C-2e). 100% yield. MS
(DCIꢄꢄCH4) m/z: 139.1234 (MH+, 4.49%), 111.1143 (M+–HCN,
27.44%),
97.0996
(M+ꢄꢄ
HCNꢄꢄCH2,
100%),
83.0871
(M+–HCNꢄꢄC2H4, 33.61%), 69.0887 (M+–HCNꢄꢄC3H6, 66.28%).
HRMS (DCIꢄꢄCH4) m/z: calcd. (12C613C2H13N2, MH+) 139.1235,
obsd. 139.1234.
2.4.1.2. Dodecanedinitrile-1,12-13C2
(
13C-2h). 100% yield. MS
(DCIꢄꢄCH4) m/z: 195.178 (MH+, 100%), 167.160 (MH+ꢄꢄCH2N,
13.39%). HRMS (DCIꢄꢄCH4) m/z: calcd. (12C1013C2H21N2, MH+)
195.1772, obsd. 195.1777.

M. Afri et al. / Chemistry and Physics of Lipids 184 (2014) 105–118
109
Table 1
1H NMR data for compounds 2–6.
2ea
2ha
3ea
3ha
4ea
4ha
5aa
5ha
5i
2.23
5j
5k
2.29
6c
6d
6ea
6f
6ha
6i
H-2
H-3
H-4
H-5
H-6
H-7
H-8
H-9
H-10
H-11
H-12
H-13
H-14
H-15
OH
2.38
1.88b
1.51
2.35
1.64
1.44
1.31
1.31
1.31
1.31
1.44
1.64
2.35
2.29
1.60
1.37
1.37
1.60
2.29
2.28
1.56
1.32
1.32
1.32
1.32
1.32
1.32
1.56
2.28
2.31
1.63
1.34
1.34
1.63
2.31
2.30
1.62
1.28
1.28
1.28
1.28
1.28
1.28
1.62
2.30
2.32
1.63
1.35
1.35
1.63
2.35
2.30
1.62
1.28
1.28
1.28
1.28
1.28
1.28
1.62
2.34
2.31
1.60
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.60
2.31
2.35
1.73
1.73
2.92
2.27
1.64
1.37
1.64
2.85
2.22
1.54
1.27
1.27
1.63
2.81
2.29
1.60
1.31
1.31
1.31
1.69
2.86
2.30
1.61
1.28
1.28
1.28
1.28
1.28
1.28
1.70
2.86
2.27
1.58
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.67
2.85
1.59
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.59
2.26
11.12
1.59
1.22
1.22
1.22
1.22
1.22
1.22
1.22
1.22
1.22
1.22
1.59
2.29
10.81
3.63
1.51
1.69b
2.38
2.83
2.85
OCH
3.67
3.67
3.67
3.67
3.59
3.65
3.68
3.56
3.57
3.56
3.66
3.63
a
These compounds were also prepared with 13C-enriched nitriles or carbonyls.
The assignments for these absorptions are indefinite and may be interchangeable with each other.
b
2.4.2. General preparation of diacids 3
2.4.3.1. Dimethyl octanedioate-1,8-13C2
(
13C-4e). 81% yield.
The alkanedinitriles 2 were hydrolyzed to the corresponding
diacids using 40% KOH in ethanol, according to the procedure
J = 57.5 Hz, C2 and C7), 28.82 (d, J = 3.5 Hz, C4 and C5), 24.81
(d, J = 3.0 Hz, C3 and C6). MS (DCIꢄꢄCH4) m/z: 205.135 (MH+,
Holmes and Lightner (1996). The unlabeled analogs
3
are
49.56%), 173.106
(MH+ꢄꢄCH3OH, 100%), 130.086
(MH
+
commercially available (Sigma–Aldrich).
ꢄꢄCO2CH3ꢄꢄCH3 + H+, 19.72%). HRMS (DCIꢄꢄCH4) m/z: calcd.
(
12C813C2H19O4, MH+) 205.1350, obsd. 205.1346.
2.4.2.1. Octanedioic acid-1,8-13C2 13C-3e). 77% yield; mp: 143 ꢀC.
(
MS (DCIꢄꢄCH4) m/z: 176.093 (M, 6.18%), 159.094 (MH+ꢄꢄH2O,
100%), 130.093 (MꢄꢄH2OꢄꢄCO, 20.62%). HRMS (DCIꢄꢄCH4) m/z:
calcd. (12C613C2H14O4, M) 176.0959, obsd. 176.0926.
2.4.3.2. Dimethyl dodecanedioate-1,12-13C2 (13C-4h). 45% yield, mp:
30–32 ꢀC. MS (DCIꢄꢄCH4) m/z: 275.203 (MH+ + CH5, 6.67%), 261.197
(MH+, 11.87%), 229.172 (MH+ꢄꢄCH3OH, 100%). HRMS (DCIꢄꢄCH4)
m/z: calcd. (12C1213C2H27O4, MH+) 261.1976, obsd. 261.1965. FTIR
(KBr): 2919, 2851, 1730 (s, CQO) cmꢄ1
.
2.4.2.2. Dodecanedioic acid-1,12-13C2
(
13C-3h). 99% yield; mp:
127–129 ꢀC. MS (DCIꢄꢄCH4) m/z: 233.161 (MH+, 37.36%), 215.158
(MH+ꢄꢄH2O, 100%). HRMS (DCIꢄꢄCH4) m/z: calcd. (12C1013C2H23O4,
MH+) 233.1663, obsd. 233.1613.
2.4.4. General procedure for preparation of monoester monoacid 5
Monomethyl succinate (5a) was prepared from succinic
anhydride (13a, 0.5 g) (or the commercially available
(Sigma–Aldrich) 1,4-13C2 labeled analog) and methanol (11 mL)
according to the procedure of Lodyato et al. (2004). Monomethyl
glutarate (5b) was prepared from glutaric anhydride (13b) and
methanol according to the procedure described by Oda et al.
(1995). The remaining monoesters were prepared from the
2.4.3. General preparation of dimethylesters 4
Diesters 4 were prepared according to the procedure described
by Khurana et al. (1990), by sonicating the appropriate dicarboxylic
acid with methanol and catalytic sulfuric acid. The unlabeled
analogs 4 are commercially available (Sigma–Aldrich).
corresponding diesters 4 via monohydrolysis with Ba(OH)2
ꢂ8H2O
Table 2
1H NMR data for keto esters 7 and keto acids 8.
7aa
7c
7ea
7ha
7i
7k
8aa
8b
8c
8d
8ea
8f
8g
8ha
8i
8j
8k
H-2
H-3
H-4
H-5
H-6
H-7
H-8
H-9
H-10
H-11
H-12
H-13
H-14
H-15
H-16
H-17
H-18
OH
2.72b
2.59b
2.38
1.62
1.62
2.38
2.22
1.54
1.19
1.19
1.54
2.25
2.30
1.61
1.27
1.27
1.27
1.27
1.27
1.27
1.56
2.38
2.30
1.58
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.58
2.38
2.39
1.59
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.59
2.39
2.71b
2.64b
2.42
1.90
2.39
2.41
1.62
1.62
2.41
2.37
1.58
1.24
1.58
2.37
2.28
1.51
1.19
1.19
1.51
2.32b
2.33
1.52
1.26
1.26
1.26
1.52
2.33
2.34
1.59
1.27
1.27
1.27
1.27
1.59
2.38
2.28
1.54
1.30
1.30
1.30
1.30
1.30
1.30
1.54
2.28
2.26
1.51
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.51
2.31
2.36
1.55
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.55
2.36
2.38
1.59
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.59
2.38
c
c
–
–
c
2.44
1.56
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.25
0.88
2.44
1.53
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.24
0.86
–
c
c
–
2.39
1.55
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.25
0.88
–
c
2.38
1.62
1.26
1.26
1.26
1.26
1.26
1.26
1.26
1.26
1.26
0.88
2.41
1.62
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.25
0.88
–
c
c
–
2.37
1.58
1.24
1.24
1.24
1.24
1.24
1.24
1.24
1.24
0.86
9.67
–
c
2.25
1.54
1.19
1.19
1.19
1.19
1.19
1.19
1.19
0.91
2.31a
1.51
1.91
1.91
1.91
1.91
1.91
1.91
1.91
0.81
–
c
2.33
1.52
1.26
1.26
1.26
1.26
1.26
1.26
0.88
–
2.38
1.59
1.27
1.27
1.27
1.27
1.27
0.88
c
c
–
–
c
c
2.38
1.56
1.27
1.27
1.27
0.88
–
2.28
1.54
1.30
1.30
1.30
0.87
–
c
2.38
1.58
1.58
1.24
0.89
2.31
1.51
1.18
1.18
0.81
10.64
–
2.36
1.55
1.25
0.89
c
c
–
–
2.39
1.04
2.38
1.04
OCH
3.68
3.67
3.6
3.66
3.66
3.66
a
These compounds were also prepared with 13C-enriched carbonyls.
The assignments for these absorptions are indefinite and may be interchangeable with each other.
Carbonyl position.
b
c

110
M. Afri et al. / Chemistry and Physics of Lipids 184 (2014) 105–118
Table 3
13C NMR data for compounds 2–5.
2ea
2ha
3ea
3ha
4ea
4ha
5aa
5ha
5i
5j
5k
C-1
C-2
C-3
C-4
C-5
C-6
C-7
C-8
119.57
17.08
25.08
27.87
27.87
25.08
17.08
119.57
119.86
17.07
25.31
28.58
29.11
28.66
28.66
29.11
28.58
25.31
17.07
119.86
174.63
34.11
25.50
29.52
29.52
25.50
34.11
174.63
174.62
34.16
25.66
29.84
30.18c
30.04c
174.22
34.04
24.81
28.82
28.82
24.81
34.04
174.22
174.39
34.17
25.03
29.21
29.44c
29.30c
174.03
34.10
174.51
34.17
174.5
34.13
174.59
34.21
174.55
34.18
24.85
28.86
28.86
24.85
34.00
179.62
25.01
29.19
29.41b
29.27
29.27
29.40b
29.11
24.75
34.14
180.11
24.97b
29.51d
29.42d
29.26d
29.16d
25.08d
25.06b
29.64e
29.52e
29.34e
29.25e
25.16e
25.02b
29.66f
29.51f
29.32f
29.22f
25.14f
C-9
29.84
25.66
34.16
29.21
25.03
34.17
C-10
C-11
C-12
C-13
C-14
C-15
C-16
OCH
24.71b
34.13
180.21
174.62
174.39
24.77b
34.21
180.4
24.75b
34.18
180.13
51.54
51.53
51.50
51.38
51.53
51.50
51.60
a
These compounds were also prepared with 13C-enriched nitriles or carbonyls.
The assignments for these absorptions are indefinite and may be interchangeable with each other.
These values were observed for C-5–C-8.
Five values were observed for C-4–C-10.
Five values were observed for C-4–C-11.
b
c
d
e
f
Five values were observed for C-4–C-13.
in MeOH or EtOH at room temperature overnight, following Tanaka
(1959) and Rao et al. (1987). The unlabeled analogs 5a–h are
commercially available (Sigma–Aldrich).
2.4.4.3. Monomethyl dodecanedioate-1,12-13C2 13C-5h). 51% yield,
(
mp 50–52 ꢀC. MS (DCIꢄꢄCH4) m/z: 247.185 (MH+, 9.52%), 229.172
(MH+ꢄꢄH2O, 100%), 215.159 (MH+ꢄꢄCH3OH, 30.33%). HRMS
(DCIꢄꢄCH4) m/z: calcd. (12C1113C2H25O4, MH+) 247.1820, obsd.
247.1851.
2.4.4.1. Monomethyl butanedioate-1,4-13C2 (13C-5a). 98% yield, mp:
55–59 ꢀC. HRMS (DCIꢄꢄCH4) m/z: calcd. (12C313C2H9O4, MH+)
135.0568, obsd. 135.0572.
2.4.4.4. Monomethyl tridecanedioate (5i). 72% yield, mp: 50–52 ꢀC.
2.4.4.2. Monomethyl octanedioate-1,8-13C2 (13C-5e). 26% yield, mp:
17–19 ꢀC. MS (DCIꢄꢄICH4) m/z: 173.1043 (M+ꢄꢄOH, 9.96%),159.0924
(M+ꢄꢄOCH3, 35.35%), 140.0761 (MH+ꢄCH3OHꢄꢄH2O, 87.66%).
HRMS (DCIꢄꢄCH4) m/z: calcd. (12C713C2H15O3, M+ꢄꢄOH) 173.1088,
obsd. 173.1043.
2.4.4.5. Monomethyl tetradecanedioate (5j). 75% yield. MS (CI, CH4)
m/z 273.203 (MH+, 5.67%), 255.193 (MH+ꢄꢄH2O,100%), 241.178 (MH
+
ꢄꢄMeOH, 47.66%); anal. HRMS (CI): calcd. (C15H29O4, MH+)
273.2066 obsd. 273.2032. FTIR (KBr) 3000 (vbr, s, OH), 2912,
2845, 1737 and 1710 (s, CQO) cmꢄ1
.
Table 4
13C NMR data for compounds 6–7.
6c
6d
6ea
6f
6ha
6i
7aa
7c
7ea
7ha
7i
7k
C-1
C-2
C-3
C-4
173.53
33.41
23.66
24.48
173.73
33.52
24.67b
27.81
173.81
33.69
24.44
28.40b
174.26
34.07
174.35
34.09
174.35
34.14
173.46
22.87
37.16
173.89
33.87
24.49
23.22
174.32
34.1
174.36
34.15
173.4
34.26
25.09
29.64j
29.54j
29.40j
29.28j
174.53
35.99b
25.10c
29.74k
29.57k
29.42k
29.28k
25.06b
28.81c
25.06b
29.30c
29.25c
29.21c
29.12c
29.05c
28.41c
25.11b
29.47c
29.41c
29.33c
29.27c
29.17c
29.09c
28.46c
24.91
25.01
209.34
29.47b
29.32b
C-5
46.71
24.30b
27.90b
28.89
42.98
42.90b
29.41b
29.78c
C-6
C-7
C-8
C-9
173.45
51.69
46.79
173.59
24.71
46.83
170.5
28.31b
24.87c
47.15
23.98
29.36
29.55
29.85g
210.67
42.27b
23.89
23.67
42.87
211.69
42.84c
23.67
42.87c
29.84i
29.80i
29.67i
29.45d
29.29d
29.74c
29.20b
23.94e
42.87f
211.71
42.84f
23.93e
29.02b
31.69
173.91
29.64h
29.48h
29.43h
29.36h
29.28h
C-10
C-11
C-12
C-13
C-14
C-15
C-16
C-17
C-18
OCH3
24.94b
47.13
173.83
25.00b
47.17
173.9
24.03b
42.96
211.91
42.96
23.72b
31.59
26.6
24.11c
41.6
212.26
34.26b
7.99
29.65
32.08
22.85
14.25
51.93
31.93
22.69
14.13
51.54
32.03
22.82
14.25
51.61
22.57
14.08
51.46
14.06
51.53
51.69
51.49
51.33
51.6
51.48
51.46
51.58
b–f: The assignments for these absorptions are indefinite and may be interchangeable with each other in a given set for a given compound.
a
The starred compounds were also prepared with 13C-enriched nitriles or carbonyls.
This value corresponds to C-9–C-14.
C-9–C-15 corresponds to 5 values.
C-12–C-15 corresponds to three values.
C-4–C-10 show only 5 values.
C-14–C-14 show only 3 values.
g
h
i
j
k

M. Afri et al. / Chemistry and Physics of Lipids 184 (2014) 105–118
111
Table 5
13C NMR data for keto acids 8.
8aa
8b
8c
8d
8ea
8f
8g
8ha
8i
8j
8k
C-1
C-2
178.22
27.86
177.21
32.71
178.35
33.60
179.75
33.92
179.66
34.02
178.80
33.78
179.84
34.10
179.23
34.02
180.15
34.17
179.64
34.12
179.77
34.14
C-3
C-4
C-5
C-6
C-7
C-8
C-9
C-10
C-11
C-12
C-13
C-14
C-15
C-16
C-17
C-18
36.91
209.33
42.90
18.63
24.21
24.55
24.6
24.59
24.76
24.80
24.74
24.81
24.82b
29.73q
29.51q
29.55q
29.42q
29.37q
29.17q
29.03q
42.94b
210.43
41.31b
23.87
29.69f
29.66f
29.62f
29.48f
29.42f
29.36f
29.26f
23.99c
24.92b
210.97
42.21b
23.88c
29.71g
29.66g
29.48g
29.42g
29.36g
29.26g
29.82b
24.02c
43.02d
211.54
42.54d
23.47c
29.73h
29.60h
29.54h
29.45h
29.39h
28.72b,h
32.03
29.62b
29.61b
24.02c
43.00d
211.78
42.74d
23.71c
29.55i
29.40i
28.95b,i
29.42b,j
29.84l
29.52l
29.42l
29.31b,l
24.04c
42.99d
211.96
42.90d
23.95
29.49n
29.38n
29.32n
29.15b,n
29.78o
29.57o
29.29o
29.11o
29.66p
29.54p
29.39p
29.35p
29.18p
23.95
29.79e
29.75e
29.60e
29.52e
29.49e
29.32e
23.89c
43.85d
211.67
42.70d
23.73c
29.26k
29.01k
28.84b,k
24.00
42.96
212.06
42.96
24.00
29.09b
31.76
22.65
14.19
23.95b
42.87c
212.1
24.03
42.97b
212.04
42.67b
26.13
29.18m
29.11m
29.05b,m
31.96
22.78
14.23
42.84c
23.64b
31.15
24.11b
42.61
212.42
36.00
8.00
32.06
22.83
14.26
31.93
22.70
14.13
31.93
23.08
14.13
32.02
22.80
14.24
31.85
22.65
14.08
22.81
14.23
22.54
13.91
22.52
14.01
b–d: The assignments for these absorptions are indefinite and may be interchangeable with each other in a given set for a given compound.
a
These compounds were also prepared with 13C-enriched carbonyls.
C-7–C-15 give these six signals.
C-8–C-15 give these seven signals.
C-9–C-15 give six signals.
C-10–C-15 give six signals.
C-4–C-6 give one signal.
C-12–C-15 give three signals.
C-4–C-7 give these four signals.
C-13–C-15 give these three signals.
C-4–C-9 give these four signals.
C-4–C-10 give these four signals.
C-4–C-11 give these five signals.
e
f
g
h
i
j
k
l
m
n
o
p
C-4–C-13 give these seven signals.
2.4.4.6. Monomethyl hexadecanedioate (5k). 85% yield. MS
(CI, CH4) m/z 301.24 (MH+, 4.73%), 283.224 (MH+ꢄꢄH2O, 100%),
269.207 (MH+ꢄꢄMeOH, 99.31%); anal. HRMS (CI): calcd. (C17H33O4,
MH+) 301.2379 obsd. 301.2395.
2.4.6.2. Methyl 4-oxooctadecanoate-1,4-13 C2 13C-7a). 36% yield,
(
mp: 54–57 ꢀC.
2.4.6.3. Methyl 6-oxooctadecanoate (7c). 49% yield, mp 43 ꢀC. MS
(DCI, CH4) m/z 313.278 (MH+, 100%), 281.244 (MH+ꢄꢄMeOH, 100%),
anal. HRMS (CI, CH4): calcd. (C19H37O3, MH+) 313.2770 obsd.
313.2783. FTIR 2960 (s), 2930 (s), 2880 and 2870 (m), 1740
(m, CQOꢄꢄester), 1720 (m, CQOꢄꢄketone), 1650 (m), 1510 (w),
2.4.5. General procedure for preparation of acyl chloride monoester 6
Acyl chloride monoesters 6 were prepared in quantitative yields
according to the procedure of Moreau et al. (2005) and minimally
characterized by NMR. The acyl chlorides were stored under
reduced pressure until used. The unlabeled analogs 6a, b and g are
commercially available (Sigma–Aldrich). The doubly labeled
analogs 6a, e and h were prepared but not characterized because
of their sensitivity to moisture.
1460 (m), 1220 and 1168 (m, CꢄꢄO) cmꢄ1
.
2.4.6.4. Methyl 8-oxooctadecanoate-1,8-13C2 13C-7e). The product
(
was obtained as an orange oil (0.598 g, 1.904 mmol, 100%) but
resisted recrystallization from hexane. It was, therefore, used
without further purification.
2.4.6. General procedure for preparation of keto esters 7
Keto esters 7 were prepared via the cadmium mediated
coupling of acyl chlorides 6 with alkyl bromides 9, according to
the procedure of Menger et al. (1989). The procedure was carried
out in toluene (rather than benzene) and the compounds were
purified by recrystallization from hexane. Partial spectral data
for many of the analogs appear in the literature (Tulloch, 1977).
2.4.6.5. Methyl 1,12-13C-12-oxooctadecanoate (13C-7h). 89% yield,
mp: 46–48 ꢀC. MS (DCIꢄꢄCH4) m/z: 315.279 (MH+, 53.07%), 283.272
(MH+ꢄꢄCH3OH, 34.79%), 229.191 (MH+ꢄꢄC6H14, 100%). HRMS
Table 6
13C NMR Shifts (ppm) for C-1 and C-7 in methyl 7-oxooctadecanoate (7d) in pure
The unlabeled analogs 7b, d, e–h and
j are commercially
solvents and intercalated within DMPC liposomes.a
available (Sigma–Aldrich). The major side product in this
reaction is presumably the long chain alkane dimer 12, which
was identified by its broad overlapping multiplets in 1H NMR
absorbing at ca. 1–2 ppm. The long chain aliphatics are well
known in the literature.
Solvent
ET(30)b
Carbonyl C-7d
Carbonyl C-1
C6D6
Acetone-d6
CD3CN
CD3OD
DMPC liposomes
34.3
42.2
45.6
55.4
208.44
209.40
211.74
213.84
209.43
39.1
173.20
174.00
174.71
175.70
174.56
47.2
2.4.6.1. Methyl 4-oxooctadecanoate (7a). 44% yield, mp: 54–57 ꢀC.
MS (DCIꢄꢄCH4) m/z: 311.259 (MꢄꢄH+, 21.47%), 281.248
(MH+ꢄꢄCH3OH, 100%). HRMS (DCIꢄꢄCH4) m/z: calcd. (C19H35O3,
MꢄꢄH+) 311.2586, obsd. 311.2592. FTIR: 2960 (s), 2880 (m), 1740
(m, CQO ester), 1710 (s, CQO ketone), 1470 (m), 1380 (m), 1230,
c
Calculated ET
:
a
Experimental error is ꢁ0.05 ppm in chemical shift and c ꢁ 1 kcal/mol in
calculated ET (Cohen et al., 2008a).
b
kcal/mol at 25 ꢀC.
From Fig. 4.
See Supplementary material Fig. S2.
c
1172 (m, CꢄꢄO) cmꢄ1
.
d

112
M. Afri et al. / Chemistry and Physics of Lipids 184 (2014) 105–118
(DCIꢄꢄCH4) m/z: calcd. (12C1713C2H37O3, MH+) 315.2810, obsd.
299.2569. FTIR (KBr) 3061 (vbr, s, OH), 2926, 2845 (s, ꢄꢄCH2ꢄꢄ),
1703 (s, CQO) cmꢄ1
315.2788.
.
2.4.6.6. Methyl 13-oxooctadecanoate (7i). 37% yield, mp: 43 ꢀC. MS
(DCI, CH4) m/z 313.275 (MH+,100%), 281.265 (MH+ꢄꢄMeOH, 6.35%),
197.071 ((CH2)11CO, 78.06%); anal. HRMS (CI): calcd. (C19H37O3, MH
+) 313.2743 obsd. 313.2754. FTIR (KBr) 2983 (w, ꢄꢄCH2ꢄꢄ), 2954
(m, ꢄꢄCH2ꢄꢄ), 2931, 2918 and 2849 (s, ꢄꢄCH2ꢄꢄ), 1738 and 1705
(s, CQO), 1470, 1438, 1415, 1380 and 1321 (m, CꢄꢄH), 1260, 1229,
2.4.7.9. 12-Oxooctadecanoic acid (8h). 84% yield, mp: 110–112 ꢀC.
MS (DCIꢄꢄCH4) m/z: 299.254 (MH+). HRMS (DCIꢄꢄCH4) m/z: calcd.
(C18H35O3, MH+) 299.2586, obsd. 299.2542. FTIR (KBr): 3052
(vbr, s, OH), 2912, 2845 (s,ꢄꢄCH2ꢄꢄ), 1696 (s, CQO) cmꢄ1
.
2.4.7.10. 12-Oxooctadecanoic acid-1,12-13C2
(
13C-8h). 87% yield,
1203 and 1176 (m, CꢄꢄO) cmꢄ1
.
mp: 110–112 ꢀC. MS (DCIꢄꢄCH4) m/z: 301.262 (MH+, 8.62%),
284.252 (MH+ꢄꢄOH, 18.44%), 283.251 (MH+ꢄꢄH2O, 100%), 215.153
2.4.6.7. Methyl 16-oxooctadecanoate (7k). 25% yield. MS (CI, CH4)
m/z 313.272 (MH+, 34.73%), 283.217 ((MHꢄꢄEt)ꢄ, 100%), 281.249
(MH+ꢄꢄMeOH, 61.89%); anal. HRMS (CI): calcd. (C19H37O3, MH+)
313.2743 obsd. 313.2724.
(MH+ꢄꢄC6H14
,
40.82%). HRMS
(DCIꢄꢄCH4) m/z:
calcd.
12
(
C
13C2H35O3, MH+) 301.2653, obsd. 301.2617.
16
2.4.7.11. 13-Oxooctadecanoic acid (8i). 60% yield, mp: 82 ꢀC. MS
(DCI, CH4) m/z 299.25 (MH+, 15.38%), 281.24 (MH+ꢄꢄH2O, 100%),
99.052 (CH3(CH2)4CO,12.13%), 71.07 (CH3(CH2)4, 6.5%); anal. HRMS
(CI): calcd. (C18H35O3, MH+) 299.2586 obsd. 299.2539. FTIR (KBr)
3115 (vbr, s, OH), 2954, 2930, 2916 and 2849 (s, ꢄꢄCH2ꢄꢄ), 1699
(s, CQO), 1469, 1440, 1418 and 1317 (m, CꢄꢄH), 1266, 1236, 1214
2.4.7. General procedure for preparation of keto acid 8
Keto acids 8 were prepared from the corresponding keto esters
7 by saponification according to the procedure of Menger et al.
(1989). The products were crystalline.
(w, CꢄꢄO) cmꢄ1
.
2.4.7.1. 4-Oxooctadecanoic acid (8a). 90% yield, mp: 84 ꢀC. MS
(DCIꢄꢄCH4) m/z: 299.258 (MH+, 23.36%), 281.239 (MH+ꢄꢄH2O,
100%). HRMS (CIꢄꢄCH4) m/z: calcd. (C18H35O3, MH+) 299.2586,
obsd. 299.2585. FTIR (KBr): 3405 (vbr, s, OH), 2919 (s), 2851 (m),
2.4.7.12.14-Oxooctadecanoic acid (8j). 35% yield, mp: 82 ꢀC. MS (CI,
CH4) m/z 299.260 (MH+, 7.71%), 281.248 (MH+ꢄꢄH2O, 100%); anal.
HRMS (CI): calcd. (C18H35O3, MH+) 299.2586 obsd 299.2596. FTIR
(KBr) 3061 (vbr, s, OH), 2912, 2851(s, ꢄꢄCH2ꢄꢄ), 1703
1703 (s, CQO) cmꢄ1
.
(s, CQO) cmꢄ1
.
2.4.7.2. 4-Oxooctadecanoic acid-1,4-13C2 13C-8a). 90% yield, mp:
(
84 ꢀC. 5-Oxooctadecanoic acid (8b):89% yield, mp: 80 ꢀC; MS (EI+)
m/z 298 (M+, 2.4%), 211 (M+ꢄꢄ(CH2)3COOH,15.6%), 84 (C6H12,100%).
2.4.7.13. 16-Oxooctadecanoic acid (8k). 40% yield, mp: 84 ꢀC. MS
(CI, CH4) m/z 299.258 (MH+, 5.45%), 281.252 (MH+ꢄꢄH2O, 100%);
anal. HRMS (CI): calcd. (C18H35O3, MH+) 299.2586 obsd. 299.2583.
FTIR (KBr) 3074 (vbr, s, OH), 2919, 2851 (s,ꢄꢄCH2ꢄꢄ), 1703
2.4.7.3. 6-Oxooctadecanoic acid (8c). 60% yield, mp 84 ꢀC. MS
(DCI, CH4) m/z 299.259 (MH+, 9.8%), 281.243 (MH+ꢄꢄH2O, 100%).
FTIR 3599 (vbr, s, OH), 2976 (m), 2917 (s), 2846 (m), 1715 and 1688
(s, CQ) cmꢄ1
.
(s, CQO), 1473 (m), 1481 (w), 1260 and 1105 (m, CꢄꢄO) cmꢄ1
.
2.5. NMR conditions for preparing correlation graphs and liposomal
samples
2.4.7.4. 7-Oxooctadecanoic acid (8d). 13% yield, mp: 78 ꢀC. MS
(DCI, CH4) m/z 299.259 (MH+, 48.19%), 281.249 (MH+ꢄꢄH2O, 100%),
158.095 ((CH2)10CH3, 18.84%); anal. HRMS (CI): calcd. (C18H34O3,
MH+) 299.2586 obsd. 299.2587. FTIR (KBr) 3105 (vbr, s, OH), 2954,
2928, 2917, 2848 (s, ꢄꢄCH2ꢄꢄ), 1715, 1699 (s, CQO), 1470, 1419 and
For the correlation graph, samples (ca. 10 mg) of ketoesters 7 or
ketoacid 8 were dissolved in 0.6 mL of pure solvent to give a final
concentration of 0.05 M. This is approximately the same concen-
tration as that of the intercalants within the liposomal solution.
The samples in pure solvents were scanned for 20 min. or
1000 scans. Liposomal samples were prepared as described above
in Section 2.3 and were scanned overnight (40,000 scans). D2O
served as the internal lock.
1382 (m, CꢄꢄH), 1250 and 1106 (m, CꢄꢄO) cmꢄ1
.
2.4.7.5. 8-Oxooctadecanoic acid (8e). 82% yield, mp: 79 ꢀC. MS
(DCIꢄꢄCH4) m/z: 299.258 (MH+, 13.22%), 281.261 (MH+ꢄꢄH2O,
100%), 157.084 (MH+ꢄꢄC10H22, 88.61%). HRMS (DCIꢄꢄCH4) m/z:
calcd. (C18H35O3, MH+) 299.2586, obsd. 299.2576. FTIR: 3737 (vbr,
m, OH), 2969, 2939 (s, ꢄꢄCH2ꢄꢄ), 2882 (m), 1749, 1737 (s, CQO),
3. Results and discussion
1541, 1457, 1376 (w), 1217 (w, CꢄꢄO) cmꢄ1
.
3.1. General synthesis of keto esters 7 and keto acids 8
2.4.7.6. 8-Oxooctadecanoic acid-1,8-13C2
(
13C-8e). 62% yield, mp:
As pointed out in the introduction, the first goal of this research
was to synthesize two related families of compounds containing
two carbonyl groups, long chain fatty keto esters 7 and keto acids 8.
Each of these families would be comprised of the same number of
carbons and functional groups, with the ester or acid carbonyls
presumably anchored in the polar region of the membrane. The
only difference between the various members in each set would be
the location of the second ketone carbonyl along the fatty acid
chain.
As outlined in Fig. 3 above, the first step in the synthesis of these
families was the quantitative conversion of dibromoalkanes 1 to
the corresponding dicyanoalkane derivatives 2 using KCN,
followed by basic hydrolysis of 2 to diacids 3. Monoesterification
of 3 to give monoacid monoester 5 proved to be a low-yield
approach. More effective was the initial diesterification of diacid 3
to diesters 4 followed by monohydrolysis to the desired acid esters
79 ꢀC. MS (DCIꢄꢄCH4) m/z: 283.257 (MH+ꢄꢄOH, 0.79%). HRMS
(DCIꢄꢄCH4) m/z: calcd. (12C1613C2H33O2, MH+ꢄꢄOH) 283.2548, obsd.
283.2573.
2.4.7.7. 9-Oxooctadecanoic acid (8f). 95% yield, mp: 74 ꢀC. MS
(DCI, CH4) m/z 298.2 (MH+, 7.22%), 281.247 (MH+ꢄꢄH2O, 100%),
155.108 (CO(CH2)7CO, 21.79%), 125.063 ((CH2)7CO, 12.8%), 98.047
((CH2)7, 21.74%). anal. HRMS (CI): calcd. (C18H34O3, M+) 298.2508
obsd. 298.2437. FTIR (KBr) 3200–3014 (vbr, s, OH), 2955, 2929,
2919 and 2848 (s,ꢄꢄCH2ꢄꢄ), 1717, 1699 (s, CQO), 1470, 1437, 1380,
1306 and 1332 (m, CꢄꢄH), 1106 (w, CꢄꢄO) cmꢄ1
.
2.4.7.8. 10-Oxooctadecanoic acid (8g). 26% yield, mp: 80 ꢀC. MS
(DCIꢄꢄCH4) m/z: 299.257 (MH+, 18.1%), 281.251 (MH+ꢄꢄH2O, 100%).
HRMS (DCIꢄꢄCH4) m/z: calcd. (C18H35O3, MH+) 299.2586, obsd.

M. Afri et al. / Chemistry and Physics of Lipids 184 (2014) 105–118
113
5. Esterification was accomplished by ultra-sonication in acidified
C-7 (keto)
R2=0.96
220
210
200
190
180
170
160
EtPenOH
dMePenOH MeCN
acetone
methanol, which has various advantages: (a) simplicity; (b) the
absence of hazardous reagents; (c) high yields; (d) the absence of
side products; and (e) relatively short reaction times (generally, a
few hours). The monohydrolysis of diesters 4 was mediated by Ba
(OH)2, yielding the desired monoester monoacids 5. The shorter
analogs, monomethyl succinate (5a) and monomethyl glutarate
(5b), were conveniently prepared via the methanolysis of succinic
(13a) and glutaric (13b) anhydrides, respectively. Monoester
monoacids 5 were quantitatively converted with thionyl chloride
to the acyl chlorides 6, which were alkylated in turn with the
appropriate dialkylcadmium reagent 11 to give the 18-carbon
methyl keto esters 7 accompanied by variable amounts of the
Wurtz coupling product 12. The keto ester underwent base
catalyzed hydrolysis to the desired keto acid 8. By beginning with
13C enriched K13CN (Fig. 3, step A) or doubly labeled succinic
anhydride (13a, step J), one could prepare the corresponding
labeled dicarbonyls compounds 7 and 8, homologs a, e, and h.
The 1H and 13C NMR spectral data of compounds 2–8 appear in
Tables 1–5.
benzene
MeOH
EtOH
Liposome MeBuOH
C-1 (acid)
R2 = 0.04
R2 = 0.94
R2 = 0.98
30
35
40
45
50
55
60
E T(30) (kcal/mol)
Fig. 5. The change in 13C NMR chemical shift of carbonyls C-1 and C-7 in 7-
oxooctadecanoic acid (8d) as a function of the ET(30) polarity of pure solvents. The
hollow symbols represent the corresponding liposome value. (Errors are as detailed
in Fig. 4.)
polarity. As expected, the correlation graph of keto ester 7d
(Table 6 and Fig. 4) revealed a very good direct correlation between
the polarity of the solvent and the chemical shifts of the ketone
(R2 = 0.94) and the ester carbonyls (R2 = 0.98). These results repeat
themselves closely for all eleven members of keto ester family 7
synthesized (7a–k).
3.2. Correlation graphs of keto esters 7 and keto acids 8
With keto esters 7 and keto acids 8 now in hand, we proceeded
to measure the chemical shifts of the various carbons in each
different derivatives in solvents of various ET(30) polarities. As
expected, based on extensive previous experience, only the
polarizable carbonyls were substantially and conasistently respon-
sive to polarity changes. A 13C NMR chemical shift–solvent polarity
correlation graph was created for each homolog in each family, and
the correlation coefficient R2 was determined to confirm reliability
and consistency of our results. As noted in our introductory
comments, these correlation graphs make it possible to determine
the polarity of the microenvironment experienced by a specific
carbonyl based on the observed chemical shift. From this, the
carbonyl’s qualitative depth within the liposome can be estimated
using the correlation graph.
This process will be exemplified by describing the formation
and use of a correlation graph of methyl 7-oxooctadecanoate (7d;
Fig. 3) and the corresponding keto acid 7-oxooctadecanoic acid
(8d). Ketoeser 7d was dissolved in four solvents (ET(30) values in
kcal/mol from Reichardt, 1994): benzene (34.3), acetone (42.2),
acetonitrile (45.6) and methanol (55.5). The 13C NMR chemical
shifts for each carbonyl carbon were plotted as a function of solvent
Turning now to the keto acids 8, and 7-oxooctadecanoic acid
(8d) is emblematic of the whole family. When drawing up our
initial correlation graph (not shown) in the abovementioned four
solvents (benzene, acetone, acetonitrile and methanol), we
observed an excellent correlation (R2 = 0.98) between the solvent
polarity and the chemical shifts of the ketone carbonyl. However,
the corresponding line for the acid carbonyls gave a very poor
correlation coefficient value of R2 = 0.11. For this reason, a new
correlation graph was prepared based on a much wider range of
eight solvents (abbreviation in Fig. 5; ET(30) values in kcal/mol
from Reichardt, 1994): benzene (34.3), 3-ethylpentanol
(EtPenOH; 38.5), 2,4-dimethyl-3-pentanol (dMePenOH; 40.1), 2-
methyl-2-butanol (MeBuOH; 41.0), acetone (42.2), acetonitrile
(MeCN; 45.6), ethanol (EtOH; 51.9), and methanol (MeOH; 55.5).
The new correlation graph obtained (Fig. 5) again showed an
excellent correlation for the keto group (R2 = 0.96) but a poor
correlation coefficient of R2 = 0.04 for the acid carbonyl (see the
dashed line in Fig. 5).
Further examination of this latter curve reveals that there is a
break in this line at ca. 43 kcal/mol (just above acetone). As one
proceeds to more polar solvents from this point (increasing ET(30)
values), one observes the expected direct relationship between
C-7 (keto)
220
R2 = 0.94
Liposome
methanol
acetonitrile
benzene
acetone
200
180
160
O
C
O
C
OH
R
OH
R
C-1 (ester)
R2 = 0.98
Increasing solvent polarity
Increases acyl carbon chemical shift
O
C
O
C
30
35
40
45
50
55
60
+ H
OH
R
O
R
E T(30) (kcal/mol)
Fig. 4. The change in 13C NMR chemical shift of carbonyls C-1 and C-7 in methyl 7-
oxooctadecanoate (7d) as a function of the ET(30) polarity of pure solvents. The
hollow symbols represent the corresponding liposome value. (Errors: ca. ꢁ1.0 kcal/
mol in the ET(30) value and ꢁ0.05 ppm in chemical shift. Errors are smaller than the
data points.)
Increasing solvent polarity
Decreases acyl carbon chemical shift
Fig. 6. Opposing effects of solvent polarity on the acid carbonyl chemical shift.

114
M. Afri et al. / Chemistry and Physics of Lipids 184 (2014) 105–118
Table 7
Chemical shifts (ppm) and corresponding ET(30) values (kcal/mol)a of carbons C1 and Cn of keto esters 7a–k intercalated within the DMPC liposomes.
Cmpd. (n)
d
C1 (ppm)
ET(30) C1
d
Cn (ppm)
ET(30)a
Cn
D
d(C1 ꢄ Cn)Calc.
(Å)b
7a(n = 4)
7b(n = 5)
7c(n = 6)
7d (n = 7)
7e(n = 8)
7f (n = 9)
7g(n = 10)
7h(n = 12)
7i(n = 13)
7j(n = 14)
7k(n = 16)
n.d.c
–
208.92
209.83
209.85
209.44
209.1
208.73
208.4
208.1
40.5
41
3.75
5
6.25
7.5
8.75
174.38
174.6
174.57
174.4
174.31
173.8
173.8
173.83
n.d.c
45.7
45.2
45.8
43.5
41
40.3
37.5
37.2
–
40.9
38.1
36.1
33.2
32.2
31.8
31.6
30.8
31.3
10
11.25
13.75
15
16.25
18.75
208.15
208.7
206.8
n.d.c
–
a
These ET(30) values are calculated from the correlation curve and have an error of ca. ꢁ1.0 kcal/mol (Cohen et al., 2008a).
Calculated with PCMODEL 7.50.00 Software (Bloomington, Indiana) using an MMX force field.
n.d = not determined (Section 3.3.1).
b
c
chemical shift and solvent polarity. Thus, as solvent polarity rises
the chemical shift-polarity correlation is a positive one. However,
at low polarity, polarization of the carbonyl becomes minimal, and
the determinant factor becomes the amount of acid dissociation.
It is clear, then, that no single line equation is appropriate for
the acid carbonyl correlation curve over the whole range of polarity
values. However, truth be told, a complete correlation graph is not
required. This is because, the acid groups in all the derivatives of 8
lie in the upper polar region of the liposomal membrane
(vide infra), not far from the interface. Hence, the correlation
curve obtained using just the polar solvents (acetone, acetonitrile,
ethanol and methanol) can be used – for which region, as noted
above, the correlation coefficient is excellent (R2 = 0.98).
so does the chemical shift of the carboxy carbonyl, with an
excellent correlation coefficient of R2 = 0.98. However, below the
break, the trend is inverted: as solvent polarity drops, the chemical
shift of the carboxy carbonyl also rises – again with an excellent
correlation coefficient of R2 = 0.94.
A similar correlation graph line-break has been previously
observed in other carboxylic acid systems (Weitman et al., 2001;
Afri et al., 2004). In the latter work it was thought that the break
had to do with whether or not the solvent was protic. However,
from the present study, it would seem clear that this is not the case,
since there are protic and non-protic solvents on both sides of the
break. Rather, it is now believed that the problematics of the
carboxylic acid system can be attributed to the presence of two
competing phenomena (Fig. 6).
3.3. Intercalation of keto esters 7 and keto acids 8 into liposomes
As already noted in the introduction, as the solvent polarity
increases, the charge-separated zwitterionic resonance form of the
carbonyl makes a concomitantly greater contribution. As a result,
the NMR chemical shift of the carbonyl carbon rises. However, in
the case of carboxylic acids, there is an additional contravening
factor at play: the equilibrium between the associated (COOH) and
dissociated forms (COOꢄ) of the acid moiety. As the solvent
polarity increases, there is concomitantly greater dissociation of
the carboxylic acids with a build up of negative charge on the
carboxylate oxygen (Reichardt and Welton, 2011 p. 109). This
build-up spills over to the acid carbonyl carbon, whose positive
charge is increasingly neutralized by the electron rich oxyanion.
This in turn results in a lowering of the chemical shift of the acid
carbonyl. Clearly, the amount of positive charge on the acid
carbonyl carbon at each polarity value will depend on the relative
contributions of each of these factors. At higher ET(30) values, the
polarization of the carbonyl is the predominant factor and, hence,
With suitable correlation graphs in hand, the next step was to
intercalate the substrates within DMPC liposomes, as previously
described (Frimer et al., 1996; Afri et al., 2002, 2004a,b; Bronshtein
et al., 2004c; Cohen et al., 2008a,b,c; Bodner et al., 2010), and
measure the 13C NMR chemical shifts for the acyloxy carbonyl
carbons. The data from the respective correlation graphs allowed
us to calculate the ET(30) polarity value corresponding to each
chemical shift observed for the carbonyl carbons of keto esters 7
and keto acids 8 intercalated in the liposome (see Tables 7 and 8).
In addition these tables contain
D
d(C1 ꢄ Cn)Calc, which is the aerial
gap difference (in Å) between the upper acid or ester carbonyl (C1)
and the deeper lying ketone carbonyl (Cn) – calculated using
PCMODEL software.
At this juncture, it would be appropriate to discuss the use of
PCMODEL MMX calculations for measuring distances in these
long-chain systems. One could well argue that static molecular
models of these long dicarbonyl fatty acid chains and
Table 8
Chemical shifts (ppm) and corresponding ET(30) values (kcal/mol) of carbons C1 and Cn of keto acids 8a–k intercalated within the DMPC liposomes.a
Compd. (n)
d
C1 (ppm)
ET(30)
C1
d
Cn (ppm)
ET(30)a
Cn
D
(Å)
d(C1 ꢄ Cn)Calc.
8a (n = 4)
8b(n = 5)
8c(n = 6)
8d(n = 7)
8e(n = 8)
8f(n = 9)
8g(n = 10)
8h(n = 12)
8i(n = 13)
8j(n = 14)
8k(n = 16)
176.76
177.17
177.56
177.21
177
176.95
177
176.71
176.6
176.69
176.92
56.8
55.5
56.7
55.6
51.9
53.5
51.3
51.1
50.7
51.2
52.3
210.35
210.65
210.53
209.51
209.2
208.41
209.2
207.51
207.6
42.7
43.6
42.6
37.5
38.7
34.2
33.7
30
3.75
5
6.25
7.5
8.75
10
11.25
13.75
15
16.25
18.75
33
29.3
33.3
207.38
208.42
a
See notes to Table 7.

M. Afri et al. / Chemistry and Physics of Lipids 184 (2014) 105–118
115
phospholipids generated by PCMODEL MMX calculations are an
oversimplification of reality and unlikely to yield a proper model of
a fluid bilayer. This is because lipid hydrocarbon chains in the fluid
bilayer phase have a certain probability of trans/gauche isomeri-
zation. Furthermore, lipid molecules undergo rapid wobble. Both
effects are temperature dependent and they reduce bilayer
thickness and increase area per molecule.
Noteworthy, however, are recent diffraction studies by Kuc9erka
et al. (Kuc9erka et al., 2011) which have shown that the
aforementioned effects are relatively small, 4% over a 20ꢀ range,
and hence should certainly not have any dramatic affect on the
preliminary results reported herein. Not surprisingly, these studies
also show that in saturated fatty acid chains trans/gauche
isomerization is substantially tempered by attractive chain–chain
van der Waals interactions which increase with increasing chain
length. Most importantly, PCMODEL MMX calculations estimated
the length of one-half of a bilayer – from the choline nitrogen
down to the midplane – as ca. 22 Å. This corresponds nicely to
bilayer thickness values of ca. 44 Å reported for DMPC liposomes
(Kuc9erka et al., 2004, 2005), and other experimental data reviewed
by Cohen et al. (2008a). All this strongly confirms the reliability of
the MMX calculations.
value for the ester carbonyl head group holds steady for keto esters
7b–d (n = 5–7) at ca. 45.5 kcal/mol. But then it drops some 8.5 units
down to 37 kcal/mol for keto ester 7i (n = 13), spanning much of
the non-polar region of the liposome. From previous studies
(Cohen et al., 2008a), it is known that the two glycerol carbonyls in
the DMPC liposome are situated at a polarity of ca. 50–52 kcal/mol.
Thus, an ET(30) value of 45.7 kcal/mol places keto ester 7b
somewhat below the glycerol ester carbonyls of the DMPC
liposomes toward the top of the non-polar region. By contrast,
37 kcal/mol places keto ester 7i substantially down along the
14-carbon DMPC lipid chain. Perforce, one must conclude that the
C1 ꢄ Cn proximity has a dramatic affect on the “buoyancy” of the
upper part of the keto ester chain.
Turning now to the ketone carbonyl carbons, as expected their
chemical shifts are substantially lower, spanning some 9 ET units
from 40.5 kcal/mole in compound 7a down to 31.3 kcal/mol in keto
ester 7k. The former value indicates that the ketone of 4-keto ester
7a lies toward the upper portion of the lipid fatty acid chain.
Regarding the lower value of 31.3 kcal/mol in keto ester 7k, this is
well along the fatty acid chain considering that 31 kcal/mol is the
lowest possible value on the ET(30) scale. Indeed, the ketone
carbonyl ET(30) values of Table 7 suggest that, as this moiety
penetrates deeper into the bilayer, the polarity it senses
approaches the minimum value 31 kcal/mol asymptotically – as
would be expected for an exponential function. We will return to
this point momentarily.
There is, however, one assumption which may well deserve
further investigation, and that is, that the thickness of the DMPC
lipid bilayer does not change dramatically upon the intercalation of
intercalants 7 and 8.
3.3.1. The ET(30) polarity values of keto esters 7
3.3.2. ET(30) polarity values of keto acids 8
We begin with several preliminary comments. Firstly, a quick
initial survey of the data in Table 7 for keto esters 7 reveals that no
chemical shifts were determined for the ester carbonyl of keto
esters 7a, j and k. This is because the DMPC glycerol esters of the
liposome have very large and broad absorptions at ca.
173–174 ppm, which sometimes obscure other ester peaks falling
in the same range.
The results for keto acids 8 appear in Table 8. Unlike the ester
derivatives, the 13C chemical shifts of all the acyloxy carbons were
detected, because the carbonyl peaks fall downfield (>176.6 ppm)
from the liposomal DMPC ester carbonyls (173–174 ppm). The
results indicate that as expected the carboxylate carbonyls are
located in the polar region of the liposome; however, they are not
anchored at the water–lipid interface (at ca. 63 kcal/mol) or even at
some other very polar value, as generally assumed in the literature
(Bronshtein et al., 2004). Rather, they range over ET(30) polarity
values of 51–57 kcal/mol. These results are not inconsistent with
previous work from our lab (Cohen et al., 2008a) which has shown
that the interface region between the lipid chains and the aqueous
media is actually a broad area, which spans about 12 ET(30) units
(ꢅ63–51 kcal/mol). This area extends from the charged choline
group, past the zwitterionic phosphate moiety, down to the
glycerol ester carbonyls. Some ions and small polar molecules such
as water have been shown to readily penetrate into this area.
(Gamliel et al., 2008).
Interestingly, the ET values for the acid carbonyls remains
rather constant at ca. 56 kcal/mol for the shorter derivatives 8a–d
(n = 4–7). However, this is followed by a sudden drop of 4 ET units to
52 kcal/mol for 8e (n = 8), and the polarity value again remains
essentially constant at 51–52 kcal/mol for the remaining deriva-
tives 8e–k (n = 8–16). Interestingly, as in keto esters 7, the breaking
point for the acid carbonyl is in-between n = 7 (8d) and n = 8 (8e).
As before, the drop is presumably related to the relative
hydrophilicity of the upper portion of the keto acid.
The results for the ketone carbonyls of keto acids 8 are similar
mutatis mutandis to those observed in the keto ester 7 system. The
polarity of the keto-carbonyl carbon decreases as the distance from
the polar head group increase. Here too, there is a sudden sharp
drop in ET(30) value in-between n = 7 (8d) and n = 8 (8e). But again
this is not surprising assuming that the chain is lying perpendicular
to the interface and the head and tail portions of the chain move
together. Interestingly, in the shallow region (4 ꢆ n ꢆ 6) the average
polarity of the ketone carbonyl is 43.1 kcal/mol, while in the deeper
area (12 ꢆ n ꢆ 16) the average polarity is ca. 33 kcal/mol, placing it
deep within the DMPC lipid slab.
Secondly, in interpreting the results of Table 7, it seems
appropriate to assume that keto esters 7a–k lie essentially
vertically in the liposomes, parallel to the liposomal lipids, based
on the following reasoning. As mentioned above, there exists a
polarity gradient as one moves deeper into the lipid
bilayer – ranging from that of water (ET(30) = 63 kcal/mol) at the
water–lipid interface down to hexane (ET(30) = 31 kcal/mol) deep
in the non-polar lipid slab. Different molecules in the lipid bilayer
feel different polarities in accordance to their location and
orientation. The keto esters 7a–k are amphiphilic, having
dissimilar hydrophobic and hydrophilic regions. When intercalat-
ed into the liposome, the acyloxy moiety of the ester head group is
expected to rise toward a rather polar region within the bilayer. On
the other hand, the long fatty acid tail is expected to pull the
intercalant toward the more hydrophobic lipid slab, allowing for
greater van der Waals interactions with the long-chain lipid tails of
the liposome. As a result, keto esters 7a–k are expected to lie
vertically in the liposomes, parallel to the lipid chains and
perpendicular to the water/lipid interface.
Considering that all the keto ester derivatives 7a–k have the
same number of carbons, and differ only in the location of the
ketone group, it would have been expected that all the ester
carbonyls would be anchored at approximately the same value. The
ET(30) value for the ketone carbonyls would be expected to
decrease as the “n” value increases and the ketone moves down the
chain approaching the lipid slab. This, in turn, should render the
construction of a chemical ruler quite trivial.
Much to our frustration, the behavior and observed values for
keto esters 7 given in Table 7 are not as predicted. In particular, the
ester carbonyls are not anchored near the lipid–water interface,
nor are they even all located at the same depth. Indeed, the ET(30)

116
M. Afri et al. / Chemistry and Physics of Lipids 184 (2014) 105–118
This process was repeated iteratively several times, until obtaining
the best fit required no further corrections. Fig. 7 shows the “best
fit” curve obtained for the data points of keto esters 7 and keto
acids 8.
The exponential curve gives us two mathematical relationships
between PD and the corresponding ET(30) value, as shown in Eqs.
(2) and (3).
ETð30Þ ¼ 31:5 þ 31:5 ꢃ expðꢄ0:137PDÞ
(2)
ꢀ
ꢁ
31:5
ETð30Þ ꢄ 31:5
PD ¼ 7:30 ꢃ ln
(3)
From a practical perspective, using Eq. (3) to calculate PD values
from the corresponding ET(30) numbers works well (i.e., gives good
results with only small errors) only in the more polar area of the
bilayer (ET(30) 0 37). However, as one moves to the very lipophilic
areas of the liposomal bilayer – as is the case for the PD of the Cn
very long chain keto esters and keto acid and, a slight error in ET
leads to a very large error in the calculated PD. As a result, a more
reliable means for determining the PD of Cn, is to calculate the PD
of the headgroup C1 using Eq. (3), and then adding to it the
Fig. 7. ET(30) values sensed by the keto ester 7 and keto acid 8 carbonyls as a
function of penetration depth (PD).
3.4. The change in ET(30) as a function of penetration depth (PD)
PCMODEL calculated distance from C1 to Cn,
D
d(C1 ꢄ Cn)Calc. This
For the purpose of constructing a molecular ruler, it is
important to determine how the ET value changes as a function
of the penetration depth (PD) within the liposome. As previously
explained, the original plan was to develop a molecular ruler using
families of isomeric compounds containing two polarizable
functional groups; the upper moiety would be anchored in or at
the interphase, while the lower one would progress gradually
down the chain. Since the ester and acid functionalities of
derivatives 7 and 8 were not anchored as hoped for, a less
straight-forward more sophisticated approach for gleaning the
desired information was necessary.
Because the lipids comprising the liposomal bilayer are
amphiphilic, two rather different areas of polarity, one polar and
the other apolar, would be a priori expected. As noted in our
introductory comments, a gradient of polarities was indeed
observed – ranging from that of water (ET(30) = 63 kcal/mol) at
the water–lipid interface, down to hexane (ET(30) = 31 kcal/mol)
deep in the non-polar lipid slab. Sharp transitions in polarity do not
normally exist in nature on the bulk scale, unless one is dealing
with different sides of a membrane or other partition. This
in effect is Eq. (1). Table 9 lists the PDCalc of carbons C1 keto esters 7
and 8, calculated from the corresponding ET(30)exp values of
Tables 7 and 8 using Eq. (3). The PDCalc of carbons Cn was calculated
using Eq. (1), as just described. It should be noted that since keto
esters 7a, j and k lack points for C1, they were not included in Fig. 7
and Table 9. Also not included are keto acids 8h and j which yield ET
(30) ET values for the ketone carbonyl which are below 31 kcal/mol,
the minimal value possible.
The curve in Fig.
7 raises several issues which require
clarification. Firstly, while the data points for keto esters 7 and
keto acids 8 all lie in close proximity to the “best fit” curve, the keto
ester points as a rule tend to lie on or above the curve, while those
of the keto acid seem to lie on or somewhat below. This is
presumably related to the fact that the substrate (intercalant) is
present in a substrate:lipid molar ratio of 1:5. As a result, the
polarity sensed by the carbonyls are to some small extent a result of
the substrate itself, and not merely the solvent. Having a large
number of points from different substrates should allow us to
cancel this effect out to some extent.
corresponds best to an exponential function, which has
a
Table 9
penetration depth (PD) (in Å) zero at ca. ET(30) = 63 kcal/mol at
one end, and approaches the minimum ET(30) value at ca. 31 kcal/
mol asymptotically at the other.
To calculate this exponential function, as a first approximation,
all the head group carbonyls (C1) of the keto esters 7 and keto acids
8 were lined up on the Y-axis according to their determined ET(30)
values (see Tables 7 and 8) at zero penetration (X-axis). The tail
group carbonyls were then placed on the graph based on their
determined ET(30) values and an approximate PD, assuming that
the latter value is governed by Eq. (1):
ET(30) value (kcal/mol) and calculated penetration depth (PD in Å)of carbons C1 and
Cn of keto esters 7 and 8 intercalated within the DMPC liposomes.
Cmpd. (n)
ET(30) C1
PD C1
Calc.b
PD Cn
Calc.c
ET(30) Cn
Exp.a
Exp.a
Calc.d
8a (n = 4)
8b(n = 5)
8c(n = 6)
8d(n = 7)
8e(n = 8)
8f(n = 9)
8g(n = 10)
8i(n = 13)
8k(n = 16)
7b(n = 5)
7c(n = 6)
7d (n = 7)
7e(n = 8)
7f (n = 9)
7g(n = 10)
7h(n = 12)
7i(n = 13)
56.8
55.5
56.7
55.6
51.9
53.5
51.3
50.7
52.3
45.7
45.2
45.8
43.5
41
1.64
2.03
1.67
2
3.25
2.68
3.47
3.7
5.39
7.03
7.92
9.5
12
12.68
14.72
18.7
21.85
10.95
12.47
13.4
15.96
18.95
20.77
26.13
27.76
42.7
43.6
42.6
37.5
38.7
34.2
33.7
33
46.6
43.5
42.1
40.1
37.6
37
35.7
33.9
33.1
38.5
37.2
36.5
35
PDCn ¼ PDC1 þ
DdðC1 ꢄ CnÞ
(1)
3.1
33.3
41
where
D
d(C1 ꢄ Cn)Calc is the PCMODEL calculated distance from C1
5.95
6.22
5.9
7.21
8.95
9.52
12.38
12.76
to Cn in Å, listed in Tables 7 and 8. The curve drawing program was
then asked to find the exponential curve that would give the “best
fit.” Once this was done, the head groups were then moved to the
curve and the “best fit” process repeated. It should noted that in the
case of the keto acids, since the headgroup carbonyls lie at the
beginning of the curve where the slope is rather steep, only small
corrections were required. For the keto esters, however, the initial
correction was a bit more substantial. Clearly, a revision of the
initial PD value of the C1 headgroup carbonyls had a concomitant
effect on the location of the tailgroup carbonyls, based on Eq. (1).
40.9
38.1
36.1
33.2
32.2
31.8
31.6
33.8
33.3
32.4
32.2
40.3
37.5
37.2
a
From Tables 7 and 8.
From Eq. (3).
From Eq. (1), where PD C1 was calculated from Eq. (3).
From Eq. (2). The standard deviation for 17 points is 1.90.
b
c
d

M. Afri et al. / Chemistry and Physics of Lipids 184 (2014) 105–118
117
Secondly, the thickness of one-half of the DMPC bilayer is
estimated to be ca. 22–23 Å (Cohen et al., 2008a). According to
the curve in Fig. 6 this occurs at an ET(30) value of about
32–33 kcal/mol Yet the curve places the tail carbonyl of the longest
keto esters 7h (n = 12) and 7i (n = 13) at ca. 27 Å. This data can be
readily resolved in one of two ways. The first is that the tail carbons
cross the midplane and begin to penetrate the complementary half
of the bilayer. In this lower region of the second half of the lipid
layer, the ET(30) polarity is also expected to be ca. 32–33 kcal/mol.
An alternative suggestion is that the long tails situate themselves
in the midplane. This latter approach has been widely used to
describe the location of the ubiquinone tails, or even rings, within
the bilayer (see: Afri et al., 2004a for a review).
The final issue relates to the completeness of our results.
Combining keto ester 7 and keto acids 8 together to construct the
molecular ruler curve in Fig. 7 was a serendipitous choice. The
former family supplies a good number of points in the apolar
section of the curve, while the latter affords many points for the
more polar region. Although the points nicely fit the curve with a
standard deviation of 1.9 kcal/mol, experimental points are lacking
for the region between ET(30) values of 51 and 45.5 kcal/mol. It
would, therefore, seem appropriate to add more points in this
region of the curve. To this end, the companion paper describes the
preparation of a series of ketophosphatidylcholines, whose ester
carbonyls fall exactly in this region (Afri et al., 2014).
Afri, M., Ehrenberg, B., Talmon, Y., Schmidt, J., Cohen, Y., Frimer, A.A., 2004a. Active
oxygen 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., Frimer, A.A., Cohen, Y., 2004b. Active oxygen chemistry within the
liposomal bilayer Part IV: locating 20,70-dichlorofluorescein (DCF),
20,70-dichlorodihydrofluorescein (DCFH) and 20,70-dichlorodihydrofluorescein
diacetate (DCFH-DA) in the lipid bilayer. Chem. Phys. Lipids 131, 123–133.
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.
Afri, M., Naqqash, M.E., Frimer, A.A., 2011. Using fluorescence to locate intercalants
within the lipid bilayer of liposomes, bioliposomes and erythrocyte ghosts.
Chem. Phys. Lipids 164, 759–765.
Altenbach, C., Greenhalgh, D.A., Khorana, H.G., Hubbell, W.L., 1994. A collision
gradient method to determine the immersion depth of nitroxides in lipid
bilayers: application to spin-labeled mutants of bacteriorhodopsin. Proc. Natl.
Acad. Sci. U. S. A. 91, 1667–1671.
Aussenac, F., Laguerre, M., Schmitter, J.-M., Dufourc, E.J., 2003. Detailed structure
and dynamics of bicelle phospholipids using selectively and per-deuterated
labels: a 2H NMR and molecular mechanics study. Langmuir 19, 10468–10479.
Bodner, E., Afri, M., Frimer, A.A., 2010. Determining radical penetration into
membranes using ESR splitting constants. Free Radic. Biol. Med. 49, 427–436.
Bronshtein, I., Afri, M., Weitman, H., Frimer, A.A., Smith, K.M., Ehrenberg, B., 2004.
Porphyrin depth in lipid bilayers as determined by iodide and parallax
fluorescence quenching methods and its effect on photosensitizing efficiency.
Biophys. J. 87, 1155–1164.
Caputo, G.A., London, E., 2003. Using a novel dual fluorescence quenching assay for
measurement of tryptophan depth within lipid bilayers to determine
hydrophobic
a-helix locations within membranes. Biochemistry 42,
3265–3274.
Cohen, Y., Afri, M., Frimer, A.A., 2008a. NMR-based molecular ruler for determining
the depth of intercalants within the lipid bilayer. Part II: quantitative values.
Chem. Phys. Lipids 155, 114–119.
Cohen, Y., Afri, M., Frimer, A.A., 2008b. Aggregate formation in the intercalation of
long chain fatty acid esters into liposomes. Chem. Phys. Lipids 155, 120–125.
Cohen, Y., Bodner, E., Richman, M., Frimer, A.A., 2008c. 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.
Cohen, Y., Weitman, H., Afri, M., Yanus, R., Rudnick, S., Talmon, Y., Schmidt, J., Aped,
P., Shatz, S., Ehrenberg, B., Frimer, A.A., 2012. The effect of intercalants on the
host liposome. J. Liposome Res. 22, 306–318.
4. Conclusions
We had originally hoped to develop a molecular ruler using
families of isomeric compounds containing two polarizable
functional groups, where the upper moiety was anchored high
in the bilayer, and the lower group would progress gradually down
then chain. Although keto ester and keto acid derivatives 7 and 8
were not anchored as originally thought, use of an iterative best fit
analysis of the data points obtained allows one to prepare an
exponential curve. This curve provides substantial insight into the
correlation between ET(30) micropolarity and penetration depth
into the liposomal bilayer. Still lacking for the curve are data points
in the ET(30) range of 51–45.5 kcal/mol. Once suitable substrates
are prepared, we should be closer to a molecular ruler for testing
and application.
Evanics, F., Hwang, P.M., Cheng, Y., Kay, L.E., Prosser, R.S., 2006. Topology of an outer-
membrane enzyme: measuring oxygen and water contacts in solution NMR
studies of PagP. J. Am. Chem. Soc. 128, 8256–8264.
Fanucci, G.E., Cafiso, D.S., 2006. Recent advances and applications of site-directed
spin labeling. Curr. Opin. Struct. Biol. 16, 644–653.
Frazier, A.A., Wisner, M.A., Malmberg, N.J., Victor, K.G., Fanucci, G.E., 2002.
Membrane orientation and position of the C2 domain from cPLA2 by site-
directed spin labeling. Biochemistry 41, 6282–6292.
Frimer, A.A., Strul, G., Hameiri-Buch, J., Gottlieb, H.E., 1996. Can superoxide organic
chemistry be observed within the liposomal bilayer? Free Radic. Biol. Med. 20,
843–852.
Fukuzawa, K., 2008. Dynamics of lipid peroxidation and antioxidion of alpha-
tocopherol in membranes. J. Nutr. Sci. Vitaminol. 54, 273–285.
Gamliel, A., 2005. Spin-Trapping with New Lipophilic Spin-Traps, Ph.D. Thesis. Bar-
Ilan: Ramat Gan, Israel.
Transparency document
Gamliel, A., Afri, M., Frimer, A.A., 2008. Determining radical penetration of lipid
bilayers with new lipophilic spin traps. Free Radic. Biol. Med. 44, 1394–1405.
Gregoriadis, G., 1984. Liposome Technology, Vol. I–III. CRC Press, Boca Raton.
Herenyi, L., Veres, D., Békási, S., Voszka, I., Módos, K., Csík, G., Kaposi, A.D., Fidy, J.,
2009. Location of mesoporphyrin in liposomes determined by site-selective
fluorescence spectroscopy. J. Phys. Chem. B 113, 7716–7724.
The Transparency document associated with this article can be
found in the online version.
Acknowledgements
Holmes, D.L., Lightner, D.A., 1996. Synthesis and acidity constants of 13CO2H-
labelled dicarboxylic acids. pKas from 13C NMR. Tetrahedron 52, 5319–5338.
Hong, M., 2006. Solid-state NMR studies of the structure, dynamics, and assembly of
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 is gratefullyacknowleged.
b-sheet membrane peptides and a-helical membrane proteins with antibiotic
activities. Acc. Chem. Res. 39, 176–183.
Hubbell, W.L., Altenbach, C., 1994. Investigation of structure and dynamics in
membrane proteins using site directed spin labeling. Curr. Opin. Struct. Biol. 4,
566–573.
Huster, D., Muller, P., Arnold, K., Herrmann, A., 2001. Dynamics of membrane
penetration of the fluorescent 7-nitrobenz-2-oxa-1,3-diazol-4-yl (NBD) group
attached to an acyl chain of phosphatidylcholine. Biophys. J. 80, 822–831.
Jämbeck, J.P.M., Lyubartsev, A.P., 2012. Derivation and systematic validation of a
refined all-atom force field for phosphatidylcholine lipids. J. Phys. Chem. B 116,
3164–3179.
Khurana, J.M., Sahoo, P.K., Maikap, G.C., 1990. An international journal for rapid
communication of synthetic organic chemistry. Syn. Commun. 20, 2267–2271.
Komljenovic, I., Marquardt, D., Harroun, T.A., Sternin, E., 2010. Location of
chlorhexidine in DMPC model membranes: a neutron diffraction study. Chem.
Phys. Lipids 163, 480–487.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.chemphy-
slip.2014.07.007.
References
Afri, M., Alexenberg, C., Aped, P., Bodner, E., Cohen, S., Eigenberg, M., Eliyahu, S.,
Harel, Y., Naqqash, M.E., Porat, H., Frimer, A.A., 2014. NMR-based molecular
ruler for determining the depth of intercalants within the lipid bilayer. Part
IV: Studies on KetoPhospholipids (Companion paper).
Kuc9erka, N., Kisekev, M.A., Balgavy, P., 2004. Determination of bilayer thickness and
lipid surface area in unilamellar dimyristoylphosphatidylcholine vesicles from
small-angle neutron scattering curves: a comparison of evaluation methods.
Eur. Biophys. J. 33, 328–334.

118
M. Afri et al. / Chemistry and Physics of Lipids 184 (2014) 105–118
Kuc9erka, N., Liu Chu, Y.N., Petrache, H.I., Tristram-Nagle, S., Nagle, J.F., 2005.
Structure of fully hydrated fluid phase DMPC and DLPC lipid bilayers using X-ray
scattering from oriented multilamellar arrays and from unilamellar vesicles.
Biophys. J. 88, 2626–2637.
Prosser, R.S., Evanics, F., Kitevski, J.L., Patel, S., 2007. The measurement of immersion
depth and topology of membrane proteins by solution state NMR. Biochim.
Biophys. Acta 1768, 3044–3051.
Rao, A.V.R., Mysorekar, S.V., Yadav, J.S., 1987. Synthesis of 1,3-diols: its application to
Kuc9erka, N., Marquardt, D., Harroun, T.A., Nieh, M.-P., Wassall, S.R., de Jong, D.H.,
Schafer, L.V., Marrink, S., Katsaras, J., 2010. Cholesterol in bilayers with PUFA
chains: doping with DMPC or POPC results in sterol reorientation and
membrane-domain formation. J. Biochem. 49, 7485–7493.
the synthesis of a-lipoic acid. Synth. Commun. 17, 1339–1347.
Reichardt, C., 1965. Empirical parameters of the polarity of solvents. Angew. Chem.
Int. Ed. 4, 29–40.
Reichardt, C., 1994. Solvatochromic dyes as solvent polarity indicators. Chem. Rev.
94, 2319–2358.
Kuc9erka, N., Nieh, M.-P., Katsaras, J., 2011. Fluid phase lipid areas and bilayer
thicknesses of commonly used phosphatidylcholines as
temperature. Biochim. Biophys. Acta 2761–2771.
Lodyato, V.I., Yurkova, I.L., Sorokin, V.L., Shadyro, O.I., Dolgopalets, V.I., Kisel, M.A.,
2004. Novel (3,5-di-tert-butyl-2-hydroxy-phenylcarbamoyl)-alkanoic acids as
potent antioxidants. Bioorg. Med. Chem. Lett. 14, 4253–4256.
a
function of
Reichardt, C., Welton, T., 2011. Solvents and Solvent Effects in Organic Chemistry,
fourth ed. Wiley-VCH Weinhein, Germany Chapter 6 and 7.
Ruíz-Gutiérrez, V., Prada, J., Pérez-Jiménez, F., 1993. Determination of fatty acid and
triacylglycerol composition of human very-low-density lipoproteins.
J. Chromatogr. B 622, 117–124.
March, J., 1992. Advanced Organic Chemistry: Reactions, Mechanisms, and
Sassaroli, M., Vauhkonen, M., Perry, D., Eisinger, J., 1990. Lateral diffusivity of lipid
analogue excimeric probes in dimyristoylphosphatidylcholine bilayers. Bio-
phys. J. 57, 281–290.
Serbinova, E., Kagan, V., Han, D., Packer, L., 1991. Free radical recycling and
intramembrane mobility in the antioxidant properties of alpha-tocopherol and
alpha-tocotrienol. Free Radic. Biol. Med. 10, 263–275.
Structures, fourth ed. Wiley, New York.
Marquardt, D., Williams, J.A., Kinnun, J.J., Kuc9erka, N., Atkinson, J., Wassall, S.R.,
Katsaras, J., Harroun, T.A., 2014. Dimyristoyl phosphatidylcholine: a remarkable
exception to a-tocopherol’s membrane presence. J. Am. Chem. Soc.136, 203–210.
Menger, F.M., Keiper, J.S., Caran, K.L., 2002. Depth-profiling with giant vesicle
membranes. J. Am. Chem. Soc. 124, 11842–11843.
Shachan-Tov, S., Afri, M., Frimer, A.A., 2010. A reinvestigation of the reaction of
coumarins with superoxide in the liposomal bilayer correlation between depth
and reactivity. Free Radic. Biol. Med. 49, 1516–1521.
Takahashi, M., Tsuchiya, J., Niki, E., 1989. Scavenging of radicals by vitamin E in
the membranes as studied by spin labeling. J. Am. Chem. Soc. 111,
6350–6353.
Menger, F.M., Richardson, S.N., Wood Jr., M.G., Sherrod, M.J.,1989. Chain-substituted
lipids in monomolecular films. Effect of polar substituents on molecular
packing. Langmuir 5, 833–838.
Moreau, J.J.E., Vellutini, L., Man, M.W.C., Bied, C., Dieudonne, P., Bantignies, J.-L.,
Sauvajol, J.L., 2005. Lamellar bridged silsesquioxanes: self-assembly through a
combination of hydrogen bonding and hydrophobic interactions. Chem.
Eur. J. 11, 1527–1537.
Tanaka, A., 1959. Anodic synthesis of fatty acids IV. Synthesis of chemical
constituents of Conifer wax. Yakugaku Zasshi 79, 1327–1331 cited by Chem.
Abstr. 1960, 54, 4381.
Mourelatou, E.A., Libster, D., Nir, I., Hatziantoniou, S., Aserin, A., Garti, N., Demetzos,
C., 2011. Type and location of interaction between hyperbranched polymers and
Tulloch, A.P., 1977. Deuterium isotope effects and assignment of 13C chemical shifts
in spectra of methyl octadecanoate and the sixteen isomeric oxooctadeca-
noates. Can. J. Chem. 55, 1135–1142.
liposomes. Relevance to design of
a potentially advanced drug delivery
nanosystem (aDDnS). J. Phys. Chem. B 115, 3400–3408.
Nielsen, R.D., Che, K., Gelb, M.H., Robinson, B.H., 2005. A ruler for determining the
position of proteins in membranes. J. Am. Chem. Soc. 127, 6430–6442.
Oda, H., Kobayashi, T., Kosugi, M., Migita, T.,1995. Synthesis of 5,12-dihete derivative
by palladium-catalyzed ternary coupling between vinylic halide, a vinylic tin
compound, and norbornadiene. Tetrahedron 51, 695–702.
Papahadjopoulos, D., 1978. Liposomes and Their Uses in Biology and Medicine.
Academy of Sciences, New York.
Postupalenko, V.Y., Shvadchak, V.V., Duportail, G., Pivovarenko, V.G., Klymchenko, A.
S., Mély, Y., 2011. Monitoring membrane binding and insertion of peptides by
two-color fluorescent label. Biochim. Biophys. Acta 1808, 424–432.
Voinov, M.A., Kirilyuk, I.A., Smirnov, A.I., 2009. Spin-labeled pH-sensitive
phospholipids for interfacial pKa determination: synthesis and characterization
in aqueous and micellar solutions. J. Phys. Chem. B 113, 3453–3460.
Wassall, S., Wang, L., McCabe, R., Ehringer, W., Stillwell, W., 1991. Electron spin
resonance study of the interaction of alpha-tocopherol with phospholipid
model membranes. Chem. Phys. Lipids 60, 29–37.
Weitman, H., Roslaniec, M., Frimer, A.A., Afri, M., Freeman, D., Mazur, Y., Ehrenberg,
B., 2001. Solvatochromic effects in the electronic absorption and nuclear
magnetic resonance spectra of hypericin in organic solvents and in lipid
bilayers. Photochem. Photobiol. 73, 110–118.