Membrane perturbing properties of sucrose polyesters

Article abstract of DOI:10.1016/S0009-3084(00)00222-X

Sucrose polyester (SPE), in the form of sucrose octaesters and sucrose hexaesters of palmitic (16:0), stearic (18:0), oleic (18:1_cis), and linoleic (18:2cis) acids, have many uses. Applications include: A non-caloric fat substitute, detoxification agent, and oral contrast agent for human abdominal (MRI) magnetic resonance imaging. However, it has been shown that the ingestion of SPE was shown to generate a depletion of physiologically important lipidic vitamins and other lipophilic molecules. In order to better understand, at the molecular level, the type of interaction between SPE and lipid membrane, we have, first synthesized different type of labelled and non-labelled SPEs. Secondly, we have studied the effect of SPEs on multilamellar dispersions of dielaidoylphosphatidylethanolamine (DEPE) and dipalmitoylphosphocholine (DPPC) as a function of temperature, SPE composition and concentration. The effects of SPEs were studied by differential scanning calorimetry (DSC), X-ray diffraction, ²H and ³¹P NMR spectroscopy. At low concentration (<1 mol%) all of the SPEs lowered the bilayer to the inverted hexagonal phase transition temperature of DEPE and induced the formation of a cubic phase in a composition dependent manner. At the same low concentration, SPEs in DPPC induce the formation of a non-bilayer phase as seen by ³¹P NMR. Order parameter measurements of DPPC-d62/SPE mixtures show that the SPE effect on the DPPC monolayer thickness is dependent on the SPE, concentration, chains length and saturation level. At higher concentration (≥10 mol%) SPE are very potent DEPE bilayer to H_II phase transition promoters, although at that concentration the SPE have lost the ability to form cubic phases. SPEs have profound effects on the phase behaviour of model membrane systems, and may be important to consider when developing current and potential industrial and medical applications.

Full text of DOI:10.1016/S0009-3084(00)00222-X

Chemistry and Physics of Lipids
109 (2001) 185–202
www.elsevier.com/locate/chemphyslip
Membrane perturbing properties of sucrose polyesters
Gerald G. McManus a, Gerald W. Buchanan a,*, Harold C. Jarrell b,
Richard M. Epand c, Raquel F. Epand c, James J. Cheetham d
a Department of Chemistry, College of Natural Sciences, Carleton Uni6ersity, 1125 Colonel By Dri6e, Ottawa, Ont., Canada,
K1S 5B6
b National Research Council of Canada, Institute for Biological Sciences, Ottawa, Ontario, Canada, K1A 0R6
c Department of Biochemistry, McMaster Uni6ersity, Health Sciences Center, 1200 Main St. W., Hamilton, Ont., Canada,
L8N 3Z5
d Department of Biology, College of Natural Sciences, Carleton Uni6ersity, 1125 Colonel By Dri6e, Ottawa, Ont., Canada,
K1S 5B6
Received 11 August 2000; received in revised form 16 October 2000; accepted 16 October 2000
Abstract
Sucrose polyester (SPE), in the form of sucrose octaesters and sucrose hexaesters of palmitic (16:0), stearic (18:0),
oleic (18:1cis), and linoleic (18:2cis) acids, have many uses. Applications include: a non-caloric fat substitute,
detoxification agent, and oral contrast agent for human abdominal (MRI) magnetic resonance imaging. However, it
has been shown that the ingestion of SPE was shown to generate a depletion of physiologically important lipidic
vitamins and other lipophilic molecules. In order to better understand, at the molecular level, the type of interaction
between SPE and lipid membrane, we have, first synthesized different type of labelled and non-labelled SPEs.
Secondly, we have studied the effect of SPEs on multilamellar dispersions of dielaidoylphosphatidylethanolamine
(DEPE) and dipalmitoylphosphocholine (DPPC) as a function of temperature, SPE composition and concentration.
The effects of SPEs were studied by differential scanning calorimetry (DSC), X-ray diffraction, 2H and 31P NMR
spectroscopy. At low concentration (B1 mol%) all of the SPEs lowered the bilayer to the inverted hexagonal phase
transition temperature of DEPE and induced the formation of a cubic phase in a composition dependent manner. At
the same low concentration, SPEs in DPPC induce the formation of a non-bilayer phase as seen by 31P NMR. Order
parameter measurements of DPPC-d62/SPE mixtures show that the SPE effect on the DPPC monolayer thickness is
dependent on the SPE, concentration, chains length and saturation level. At higher concentration (E10 mol%) SPE
are very potent DEPE bilayer to HII phase transition promoters, although at that concentration the SPE have lost the
ability to form cubic phases. SPEs have profound effects on the phase behaviour of model membrane systems, and
may be important to consider when developing current and potential industrial and medical applications. © 2001
Elsevier Science Ireland Ltd. All rights reserved.
Keywords: Lipid polymorphism; Bilayer; Differential scanning calorimetry; Lipid; Membrane; Nuclear magnetic resonance;
Phospholipid; Olestra
* Corresponding author. Tel.: +1-613-5203841; fax: +1-613-5203749.
E-mail address: gwbuchan@ccs.carleton.ca (G.W. Buchanan).
0009-3084/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd. All rights reserved.
PII: S0009-3084(00)00222-X

186
G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
1. Introduction
as an oral contrast agent for human abdominal
MRI studies (Ballinger et al., 1991). There are,
however, some negative aspects of the use of SPE
as a non-caloric fat substitute. SPE absorbs fat-
soluble vitamins during the course of food diges-
tion, thus depleting the availability of these
important micro-nutrients (Crouse and Grundy,
1979; Mellies et al., 1985; Koonsvitsky et al.,
1997; Kelly et al., 1998). Consequently, food-con-
taining SPE must be supplemented with these
vitamins (Akoh, 1995; Prince and Welschenbach,
1998).
Lipidic carbohydrate polyesters have been the
focus of a broad spectrum of scientific research
ranging from studies of naturally occurring cell
membrane glycolipids to synthetic carbohydrate
polyesters. Some of the research interest in these
molecules has been based on the structural varia-
tions that can be introduced into these systems,
such as, the nature of the esterified fatty acyl
chains, the degree of esterification and the type
and number of the carbohydrate moieties. In re-
cent years, extensive research on sucrose
polyesters (SPE) has led to the discovery of a
large range of applications based on their physico-
chemical properties (Jandacek and Webb, 1978).
SPE in the form of a mixture composed mainly of
sucrose octaesters with a low amount of sucrose
hepta and hexaesters of palmitic (16:0), stearic
(18:0), oleic (18:1cis), and linoleic (18:2cis) acids is
known as (Olestra*), a low-calorie fat substitute
(Akoh, 1995). The ability of SPE to behave as a
lipid-like molecule is likely due to this high degree
of esterification with long acyl chains. In addition,
these molecules are resistant to the hydrolytic
action of pancreatic lipase and hence are non-
caloric since they are not absorbed by the small
intestine and pass through the human body unme-
tabolized (Matteson and Volpenhein, 1972).
SPE is also effective as a non-absorbable
lipophilic binding agent that can reduce intestinal
absorption of certain lipophilic molecules. The
presence of SPE as a substitute for dietary triacyl-
glycerides (TAG) has been shown to decrease
levels of low density LDL cholesterol (Crouse and
Grundy, 1979; Jandacek et al., 1990). SPE has
also been used to detoxify lipophilic environmen-
tal contaminants such as polychlorinated dibenzo-
p-dioxins (PCDDs), dibenzofurans (PCDFs),
biphenyls (PCBs), hexachlorobenzene (HCB),
2,3,7,8-tetrachlorodibenzo-p-dioxinand (TCDD)
and others (Mutter et al., 1988; Geusau et al.,
1999; Moser and McLachlan, 1999). This removal
occurs through a process described as the fat-flush
hypothesis, which provides a theoretical basis for
a two-step model of organic pollutant transfer in
the gastrointestinal tract (Schlummer et al., 1998).
Another recent application of SPE involves its use
Of further interest is the potential of SPE to
interact with and insert into biological mem-
branes. Previous studies have shown that the tem-
peratures of the gel (Lb) to liquid crystalline (La)
and bilayer to inverted hexagonal (HII) phase
transitions, as well as the formation of non-bi-
layer cubic phases can be altered by the presence
of hydrophobic and amphipathic molecules.
These include lipidated carbohydrates, such as the
endotoxic lipopolysaccharides (LPS) composed of
a polysaccharide and a lipid component, termed
lipid A (Zahringer et al., 1994). LPS are the major
constituents of the outer leaflet of the outer mem-
brane of Gram-negative bacteria. Lipid A, the
lipid anchor constituent of LPS (Rietschel et al.,
1994), is known to express a large spectrum of
biological activities such as modulation of the
immune response. These biological activities may
be a direct consequence of their interaction with
cell membranes (Nikaido et al., 1977; Branden-
burg et al., 1993, 1998; Seydel et al., 1994; Mun-
stermann et al., 1999), or with membrane proteins
(Choi et al., 1998; Cock de et al., 1999; Nakamura
et al., 1999). Because of these biological functions,
glycopolyester analogues of lipid A that function
to antagonize the immunostimulatory effects of
lipid A and LPS from pathogenic bacteria (Christ
et al., 1995; Asai et al., 1998; Kawata et al., 1999)
have received interest.
Membrane fusion (Chernomordik, 1996) and
the activity of the membrane-bound enzymes such
as protein kinase C (Mosior and Epand, 1999)
were shown to be increased by the presence of
non-lamellar forming lipids. The modulation of
these activities by the physical properties of the
membrane may be a consequence of dehydration

G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
187
of the membrane interface (Epand and Bottega,
2.2. Differential scanning calorimetry (DSC)
1988; Giorgione et al., 1998a,b; Epand, 1998;
Jimenez-Monreal et al., 1999). We wished to de-
termine how SPEs affect membrane physical
properties as an indication of the effects they have
in biological systems.
Lipid films were made by dissolving DEPE and
small amounts of sucrose polyesters in chloro-
form:methanol (2:1 v/v). The solvent was evapo-
rated with nitrogen and then dried under vacuum
for 2 h. The films were stored under argon until
use. Samples were hydrated with 20 mM Pipes
buffer pH 7.4, containing 0.15 M NaCl, 1 mM
EDTA, 0.002% NaN3. The DSC of the suspen-
sions were measured with a Nanocal DSC from
Calorimetry Sciences Corp., Provo, UT, USA us-
ing a heating scan rate of 0.75°/min. The scans
were analysed using the program Origin (version
5.0).
2. Materials and methods
Phospholipids as obtained from Avanti Polar
Lipids, Inc. were pure as determined by thin-layer
chromatography (TLC) and were used without
further purification except for DEPE which was
recrystallized from warm methanol. All chemicals
were of reagent quality and were purchased from
Aldrich Chemical. TLC were performed on
Merck 60F254 silica gel coated aluminum plates
and components were detected by heating plates
after spraying with a phosphomolybdic acid in
ethanol solution. For column chromatography,
2.3. X-ray diffraction
,
Nickel-filtered CuK (u=1.54 A) X-rays were
obtained from a Rigaku-Denki rotating anode.
X-rays were focused using a Frank’s type camera
and recorded using a position-sensitive propor-
tional counter TEC Model 205 (Sen et al., 1990).
Unoriented lipid dispersions, prepared as de-
scribed above, were measured in 1.5 mm glass
capillaries in all diffraction experiments. The spec-
imen temperature (90.5°C) during X-ray diffrac-
tion was controlled by a thermoelectric device.
The calculated d-spacing is given as the average of
two determinations. The first sharp peaks from
the centre of the diffraction patterns are due to
the beam stop. In the diffractograms shown, the
intensities are given in arbitrary units, and are
plotted versus the parameter Q, which is the
reciprocal of the d-spacing, d, and related to the
later by:
,
silica gel 70–230 mesh, 60 A, was used. The
melting point measurement for the pure crys-
talline SPE was determined by differential scan-
ning calorimetry (DSC) on a Seiko 220C at a
heating rate of 1°C/min with a nitrogen flow of
200 ml/min in an oscillating mode.
2.1. Sample preparation
2
For 31P and H NMR, multilamellar lipid dis-
persions were prepared by evaporating a chloro-
form solution containing 50 mg of phospholipids
and the appropriate amount SPE under a stream
of dry nitrogen followed by 5 h under vacuum.
The resulting dry lipid film was hydrated with
excess deuterium depleted water (Aldrich) or dis-
tilled water for the 2H and 31P NMR samples,
respectively. The samples were repeatedly (five
times) vortexed at temperatures above the gel to
liquid crystal phase transition of the phospholipid
followed by freeze–thawing and then lyophilized.
The dry powdered mixtures were rehydrated with
200 ml of the appropriate water and vortex–
freeze–thawed five times to obtained a uniform
lipid suspension.
Q=2y/d.
Typical absolute lattice spacing, calibrated
,
against nonadecane (d=26.2 A), are accurate to
,
,
90.5 A for lattices up to ꢀ80 A.
2
2.4. 31P and H NMR
2
31P and H NMR spectra were obtained at 161
and 61.1 MHz respectively, on a VARIAN IN-
2
OVA 400 NMR spectrometer. H NMR spectra
were acquired using the quadrupolar echo se-

188
G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
quence as described previously (Fenske et al.,
1991) using a y/2 pulse length of 3.5 ms (5 mm
(i.d.) solenoid coil) an interpulse delay of 35 ms, a
recycling delay of 1.0 s, and a spectral width of
500 kHz. Scans (500–2000) were accumulated.
The free induction decay (FID) was shifted to the
top of the echo and the resulting FID was
apodized by multiplying with a Lorentzian line
broadening of 300 Hz. Proton decoupled 31P spec-
tra were acquired using a Hahn echo pulse se-
quence (Rance and Byrd, 1983) a y/2 pulse length
of 3.5 ms (5 mm (i.d.) solenoid coil), an interpulse
delay of 1.5 ms, a recycle delay of 2.5 s, and a
spectral width of 125 kHz. Samples were enclosed
in a glass jacket in which the temperature was
regulated to 91°C.
precipitate was dissolved in a minimum amount
of hexane and washed with methanol until the
methanol layer was colourless. After drying over
anhydrous sodium sulphate and filtration, the
hexane was removed from the crude product by
rotary evaporation. Subsequent column chro-
matography on silica gel with hexane:ether:acetic
acid (70:30:1) as eluate gave the sucrose octaester
as the first eluant which was crystallized from
ethyl acetate. SOP was obtained. Using methyl
palmitate, SOP was obtained, as a white powder,
in a single crystalline phase, m.p. 48.0°C from
DSC (literature value: crystalline a-phase 50.5°C
and crystalline b-phase 55.0°C). SOS was ob-
tained; using methyl sterate obtained, as a white
powder, m.p. 54.0 and 57.0°C from DSC (litera-
ture value: crystalline a-phase 61.0°C and crys-
talline b-phase 64.5°C ; Jan dacek and Webb,
1978). SOO and SOL were obtained, using methyl
oleate, and methyl linoleate, as a viscous yellow
oil.
2.5. Synthesis of the sucrose octaester
This synthesis was based on a procedure devel-
oped by Akoh and Swanson (1990). The different
sucrose octaesters were synthesized by substitut-
ing the appropriate fatty acid. A 100 ml three-
necked round bottom flask equipped with a
magnetic stirrer, thermometer, fractional distilla-
tion column connected to a condenser with a
vacuum take-off line leading to a liquid nitrogen
cold trap, manometer, and a vacuum pump. The
methyl ester, of the appropriate fatty acid, was
added in an eightfold mole ratio to the reaction
flask and flushed with dry N2 gas for 30 min.
Sucrose octaacetate (0.55 g, 0.81 mmol) was
added and the system was flushed for an addi-
tional 15 min. Sodium metal (0.05 g, 2% of the
reactants w/w) was added, the reaction flask was
then immersed in a preheated silicon oil bath at
80°C. The reaction mixture was gradually heated
to 110°C and vacuum applied to obtain a pressure
of about 5 mm Hg. N2 gas flow was input every
20 min for a 1 min. period. The reaction was
carried out under these conditions for 2 h at
which point heat and pressure were removed.
After cooling to room temperature, 1 ml of acetic
acid in 100 ml hexane was added to the solution.
Subsequent addition of 1 l of methanol caused a
precipitate to form and the resulting heteroge-
neous mixture was stirred overnight. The
methanol was removed by gravity filtration. The
2.6. Synthesis of
6-6%-di-O-hydroxy-hexa-O-palmityl-sucrose
(SHxP)
Sucrose (12 g, 35.1 mmol) was dissolved in dry
pyridine at 90°C for 10 min. The solution was
cooled at room temperature and triphenylmethyl
chloride (10 g, 36.0 mmol) was added. The mix-
ture was stirred at room temperature for 90 h
under dry N2 gas. The solution was poured into
ice water and stirred for 20 min. The aqueous
phase was extracted with dichloromethane (3×30
ml). The combined organic phase was dried with
anhydrous sodium sulphate, and evaporated to a
yellow syrup. Removal of residual solvent in
vacuo gave a white crystalline powder. TLC
(chloroform:methanol:acetic acid (85:15:1)) indi-
cated the presence of four compounds. The mix-
ture was subjected to column chromatography
(silica gel with chloroform:methanol:acetic acid,
85:15:1) to give four fractions. Each fraction was
acetylated with dry pyridine and an excess of
acetic anhydride at room temperature under dry
N2 gas for 12 h. The reaction was quenched with
ice water and the aqueous phase was extracted
with dichloromethane (3×30 ml). The organic

G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
189
Fig. 1. DSC heating scans of DEPE containing increasing mole fractions of SOO or SHxP. Scan rate 0.75°/min. (A) 1. No additive
(pure DEPE) and with SOO added at a mole fraction of: 2. 0.00076; 3. 0.00151; 4. 0.00302; 5. 0.00459; 6. 0.0120. (B) 1. No additive
(pure DEPE) and with SHxP added at a mole fraction of: 2. 0.0040; 3. 0.0083; 4. 0.0125; 5. 0.0165.
the filtrate was rotary evaporated to give 6,6%-di-
O-hydroxy-hexa-O-palmityl-sucrose. Crystalliza-
tion from ethyl acetate gave 114 mg (0.064 mmol)
of the desired product having m.p. 45.5°C from
DSC.
phase was dried with anhydrous sodium sulphate
and evaporated to dryness. The resulting acety-
lated fractions were analysed and fraction 3 was
identified as 6,6%-di-O-trityl-hexa-O-acetyl-sucrose
1
by comparison of its m.p. (103–104°C) and its H
NMR spectrum with those of previously reported
6,6%-di-O-trityl-hexa-O-acetyl-sucrose
(Hough,
1972; Otake, 1972). 6,6%-di-O-trityl-hexa-O-acetyl-
sucrose (1.0 g, 1.2 mmol) was reacted with methyl
palmitate (2.0 g 7.4 mmol) as described above for
the synthesis of the sucrose octaester to give the
3. Results
3.1. DSC
6,6%-di-O-trityl-hexa-O-palmitoyl-sucrose
(200
DEPE containing unsaturated acyl chains read-
ily converts into the inverted hexagonal (HII)
phase with increasing temperature. The tempera-
ture at which the La is converted to the HII phase
is sensitive to the presence and type of additives in
the membrane (Epand, 1985). DSC thermograms
of the Lb/La and La/HII phase transitions of
DEPE in the absence and presence of an increas-
ing amount of an SPE additive are shown in Fig.
1 using SOO and SHxP as examples. SHxP is the
only SPE from this study that has an effect on the
Lb/La phase transition at the mole fractions stud-
ied. Mole fractions of 0.004 up to 0.017 of SHxP
gradually lowered the Lb/La phase transition en-
thalpy (DH) of DEPE and progressively broad-
ened the peak width without shifting the
transition temperature. SHxP also generated a
mg, 0.08 mmol). Hydrogen bromide (48% in
glacial acetic acid; 3 ml) was added with stirring
to a solution of the sucrose hexapalmitate deriva-
tive in 50% (v/v) acetic acid:dichloromethane so-
lution (10 ml) at 5.0°C. After 15 min, the organic
layer was washed with sodium hydrogen carbon-
ate (10 ml) and then with water (10 ml). TLC
shows the presence of three compounds. Column
chromatography on silica gel with hex-
ane:ether:acetic acid (70:30:1) followed by chloro-
form:methanol:acetic acid (85:15:1) as elutate,
gave 6,6%-di-O-hydroxy-hexa-O-palmitoyl-sucrose
containing triphenylmethanol. Solvent was re-
moved under vacuum and light petroleum ether
(low boiling) was added to the crude product. The
insoluble triphenylmethanol was filtered out, and

190
G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
Fig. 2. Plot of transition temperatures versus SPE mole fractions. (A) DSC heating scans of DEPE containing increasing mole
fractions of SOP or SOS. Scan rate 0.75°/min. (B) 1. No additive (pure DEPE) and with SOP added at a mole fraction of: 2. 0.0020;
3. 0.0051; 4. 0.0100; 5. 0.020. (C) 1. No additive (pure DEPE) and with SOS added at a mole fraction of: 2. 0.0020; 3. 0.005; 4. 0.010;
5. 0.020.
new transition peak with low DH at 44.8°C, which
corresponds to the melting of the crystalline state
of SHxP. The attribution of this peak to the
melting of SHxP crystals is based on the absence
of any change in the DEPE bilayer 31P NMR
powder patterns (see below) indicating that the
44.8°C peak must come from a phase change in
the SHxP such as melting of crystals.
All SPE additives used in this study reduce the
bilayer to hexagonal phase transition temperature
(TH) of DEPE at exceptionally low mole frac-
tions. Inspection of Fig. 2A reveals that at similar
SPE/DEPE ratios the effectiveness in reducing TH
is dependent on the degree of esterification of the
sucrose moiety as well as on the length and on the
degree of unsaturation of the acyl chains. For
instance, the hexa-saturated SPE additives reduce
TH over a wide range of additive mole fraction
(Fig. 2A). With the unsaturated and saturated
sucrose octaesters, a rapid decrease in the transi-
tion temperature is observed below a mole frac-
tion of 0.0008, with smaller shifts occurring above
a mole fraction of 0.005, as seen with the SOP
and SOS (Fig. 2B and C). However, in contrast to
the low melting unsaturated SPEs, the sucrose
octaester with saturated acyl chains exhibits com-
plex endotherms having multiple components.
The identities of the multiple components at 53°C
for SOP and at 56°C and 59°C for SOS (Fig. 2B
and C), were attributed to the melting of crys-
talline state of SOP and to the two crystalline
polymorphs of SOS (Jandacek and Webb, 1978),
respectively, based on the melting temperature
similarity and 31P NMR data. At a mole fraction
ca 0.012 the two unsaturated forms of SPE reduce
the DEPE TH by 10.5 and 7.2°C for SOO and
SOL, respectively (Fig. 2A). This is greater than
the shift in TH observed at a similar mole fraction
of the saturated octaesters, SOP and SOS, which
shift equally the TH of DEPE by ca 4.6°C. Con-
versely, SHxP, which structurally differs from
SOP only by having two free hydroxyl groups on
the sucrose moiety, lowered TH of DEPE by
8.2°C, which is similar to SOL and intermediate
between those of saturated octaesters and SOO.
These results demonstrate that SPE can have sub-
stantial effects on the phase transition properties
of DEPE and are therefore partitioning into the
membrane at low concentrations of additive.

G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
191
identity of the phases present in the samples (Fig.
3). 31P NMR spectra of pure DEPE at 45°C
exhibited a lineshape with a residual chemical
shielding anisotropy (CSA) of 42 ppm, indicative
of rapid axially symmetric motion (Cullis and de
Kruijff, 1978b) in a fluid La phase. Upon heating
above TH the spectral lineshape was only repre-
sentative of the non-bilayer HII phase. 31P NMR
spectra obtained from DEPE mixtures containing
0.01 mole fraction of, SOO and SHxP as a func-
tion of temperature are also shown in Fig. 3 as
examples. Comparison of the temperature depen-
dence of the 31P lineshapes for pure lipid and lipid
mixtures reveals that two common features are
noticeable for all DEPE/SPEs mixtures studied:
(i) all SPEs promote the formation of the HII
phase at lower temperatures; (ii) all additives in-
duce the formation of an isotropic signal. Table 1
summarizes the temperatures at which the lipid
phase transitions occur, as determined from the
DSC and 31P experiments. The identification of
the DSC endotherm peaks as the La/HII phase
transitions can be established through comparison
with the 31P NMR spectral changes at the corre-
sponding temperature. This is evident in Fig. 3 in
the 31P NMR lineshapes at 45°C where all SPE/
DEPE samples are in the La phase. This supports
the assignment of the DSC peak at 45°C to the
melting of the crystalline form of SHxP. Further-
more, data in Table 1 indicate that the extent of
the DEPE bilayer disturbance is closely related to
the molecular structural variations among SPEs.
The second interesting aspect of the effect of
SPEs on DEPE is the occurrence in the 31P NMR
Fig. 3. 31P NMR spectra of DEPE with 1 mol% of SPEs as a
function of temperature. First row shows the SPE/DEPE
samples in the lamellar phase. Second row indicates the tem-
perature at which the isotropic components started from.
Third row indicates the temperature at which the HII phase
started to form. Fourth row shows the complete transition of
the DEPE into the HII phase with the superimposition of the
isotropic component. Fifth row represent the SPE/DEPE sam-
ple in the HII phase only. All samples were gradually heated in
increments of 2°C with 15 min of temperature equilibration
before accumulation of the FID.
3.2. DEPE 31P NMR spectroscopy
31P NMR is a useful spectroscopic method to
probe the molecular organization and motional
behaviour of lipid headgroups and therefore can
consequently monitor the morphology of lipid
structures (Cullis and de Kruijff, 1978a). The 31P
NMR powder patterns were used to verify the
Table 1
Temperature values of first observable isotropic signal and of the La–HII phase transition observed by DSC and by 31P NMR of 1
mol% SPE/DEPE mixtures
Samples composition
31P isotropic signal (°C)a
31P La–HII phase transition (°C)a
DSC La–HII phase transition (°C)b
Pure DEPE
SOP
SOS
SOL
SHxP
SOO
–
60
56
52
50
50
59
55
60
55
53
53
65.5
60.4
60.4
57.8
56.8
54.5
a NMR temperature error91°C.
b DSC temperature error90.2°C.

192
G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
Fig. 4. 31P NMR spectra of the five heating/cooling cycles of DEPE with 1 mol% SHxP at 45°C. The heating/cooling cycles were
performed by gradually heating the samples to 65°C. After 15 min of equilibration the temperature was lowered in increments of
10°C to a temperature of 45°C, at which point the FID was accumulated. This was performed five times for all samples.
spectra of a sharp isotropic spectral component
superposed on the axially symmetric spectrum of
the La phase as well as in the mixed phase region
(Fig. 3). Upon cooling from the HII phase, the
isotropic peak was regenerated and remained to
low temperatures in coexistence with the La
phase. The isotropic-La phase persisted until the
sample was cooled to temperatures below 25°C at
which point the spectrum reverted to that of a
typical Lb phase.
tion of isotropic signal was examined more quan-
titatively. The normalized integrated area of the
31P NMR isotropic signal was plotted as a func-
tion of the number of heating/cooling cycles (Fig.
5). Fig. 5 shows that SOP and SOL have less
ability to induce the formation of the isotropic
component compared with SOS and SOO. SHxP
and SOO caused rapid and almost complete con-
version of DEPE into an isotropic phase. The
thermal hysteresis of the conversion to the
isotropic components and their ability to form at
temperatures that ranged from ꢀ2 to 5°C below
the La/HII phase transition (Table 2) and to re-
main stable at ꢀ5–10°C above the La/HII phase
transition reflect the characteristic thermodynami-
cally stability of inverted cubic phase over the
An isotropic 31P NMR signal observed with
large lipid aggregates is indicative of the arrange-
ment of lipids in structures with high curvature,
such as the inverted cubic phases where the 31P
chemical shift anisotropy is totally averaged by
the rapid reorientation of the lipids. Pure DEPE
has not been observed to form cubic phases, but
other PEs can generate the cubic phase when the
lipid system is repeatedly cycled through the La/
HII phase transition temperature (Shyamsunder et
al., 1988; Veiro et al., 1990). To further analyse
the origin of the isotropic components, a sequence
of five temperature cycles, heating/cooling,
through the La/HII phase transition were per-
formed on pure DEPE and on the SPE/DEPE
mixtures. The pure DEPE spectrum did not show
an isotropic spectral component after five cycles,
whereas the temperature cycling of all SPE/DEPE
mixtures resulted in a progressive increase of the
isotropic signal. Fig. 4 shows the 31P spectra of
SHxP as a representation of the temperature cy-
cling effect on the amplitude of the isotropic
signal. Interestingly, the degree of increase in the
isotropic signal was not the same for all SPE
studied, suggesting that differences in molecular
structure may influence the ability of SPE to
generate isotropic phases. For this reason, the
ability of the various SPEs to induce the forma-
Fig. 5. Comparative representation of the SPE molecular
structural dependent potency to induce the formation of
isotropic component in DEPE as a function of heating/cooling
cycles.

G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
193
Table 2
made because of the small number of observed
reflections.
The average order parameter BSCD\, and the average length
of the acyl chain BL\ derived from 2H NMR
,
Sample composition
BSCD\
BL\ (A)
3.4. DPPC 31P NMR spectroscopy
DPPC-d62 only
SHxP
SOP
SOS
SOO
0.163
0.167
0.169
0.154
0.165
0.165
11.87
11.94
12.01
11.73
11.91
11.91
In order to further examine the SPE effect on
lipid phase behaviour, the 31P NMR spectra of
DPPC, DMPC, egg PC and DEPC multilayers in
the absence and presence of 1 mol% of various
SPE were measured at different temperatures. The
31P NMR spectra of 1 mol% of SOS, as a repre-
sentative of the sucrose octaester and 1 mol% of
SHxP with DPPC are shown in Fig. 7. At 35°C all
spectra are typical of a bilayer lipid arrangement
SOL
other thermotropic phases (Gagne et al., 1985;
Gruner et al., 1988; Ellens et al., 1989; Siegel,
1999). It has also been shown that addition of
only 1–10% alamethicin to DEPE can stabilize a
cubic phase over a wide range of temperature
(Keller et al., 1996). Furthermore, visual inspec-
tion of the samples immediately after the heating/
cooling cycles showed no apparent change in
sample turbidity as may be expected if the
isotropic spectral component was due to the ex-
tensive formation of micelles or small vesicles. It
should be emphasized that while the 31P data are
consistent with the formation of cubic phase lipid,
the presence of a 31P NMR isotropic signal in
phospholipid mixture is not a unique indicator of
cubic phase (Thayer and Kohler, 1981; Tilcock et
al., 1986).
3.3. X-ray diffraction
To further characterize the isotropic component
observed in SPE/DEPE mixtures, 31P NMR spec-
troscopy (Fig. 3) and small angle X-ray diffrac-
tion of a 1 mol% SOO/DEPE mixture and pure
DEPE were acquired at 55°C. The pure DEPE
diffraction pattern was characteristic of the La
,
phase with unit cell spacing of 53.7 A and showed
no change during the time period observed (20
min; Fig. 6A). The sample containing 1 mol%
,
SOO also gave a spacing of 53.7 A, but an
,
additional spacing of 100 A was also detected
,
(Fig. 6B). The magnitude of the 100 A spacing
increased with time at constant temperature. Ob-
Fig. 6. X-ray diffraction patterns of unoriented DEPE disper-
sions measured as a function of time measured at 55°C. (A)
DEPE with 1 mol% SOO. (B) DEPE only. The first sharp
peaks from the centre of the diffraction patterns are due to the
beam stop.
,
servation of this large 100 A spacing is consistent
with the presence of a cubic phase structure but
definitive identification of the phase cannot be

194
G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
incubated at 70°C for 15 min and then centrifuged
for 1 min in a microcentrifuge (at room tempera-
ture). The resulting supernatants exhibited no
isotropic 31P NMR spectrum confirming that SPE
is intercalated into DPPC and forms regions of
high curvature. Furthermore, the degree of po-
tency of the individual SPE in inducing the forma-
tion of isotropic components in DEPE correlated
well with the DPPC measurements. It is notewor-
thy that SOP is less potent in initiating the forma-
tion of the isotropic signal for both
phospholipids. 31P NMR of SPEs admixed with
DMPC, egg PC and DEPC were measured, over a
range of temperatures covering the Lb and La
phases of each of these lipids. At 45°C an
isotropic component was observed for the egg
PC/SPE mixtures and grew in intensity with in-
creased temperature, as did the isotropic compo-
Fig. 7. 31P NMR spectra of DPPC containing (a) 1 mol% SOS,
(b) 1 mol% SHxP, measured as a function of temperature.
which exhibit lineshapes characteristic of axially
symmetric motion, having a residual CSA of ca 72
ppm characteristic of the Lb phase (Cullis et al.,
1976). No indication of the formation of nonbi-
layer phases was found at temperatures below
42°C for the SPE/DPPC mixtures. At 42°C all
sucrose octaester samples show the presence of an
isotropic signal superimposed on a La phase
lineshape having a residual CSA of ca 48 ppm
(Cullis et al., 1976) with the exception of the
SOP/DPPC mixture which had an isotropic com-
ponent at 44°C. The quantity of lipid in the
isotropic phase increased with temperature. The
SHxP sample showed the presence of a tempera-
ture induced isotropic signal. However, the
lineshape is no longer characteristic of a simple La
phase bilayer with CSA of ca 48 ppm. Instead the
SHxP 31P NMR spectrum at 45 and at 50°C
displays a composite of Lb and La phase bilayers.
Upon cooling, the isotropic signal reverted back
to the same intensity as observed during the heat-
ing and vanished at the same temperature that it
was first observed. The presence of lipids in an
isotropic component can be rationalized in two
ways. Either the presence of SPE induced a frac-
tion of the lipids to form small SPE/lipid particles
of high curvature that accumulate outside the
bilayer and grow as the temperature is increased,
or the SPE induced the formation of an area of
high curvature within the bilayer structure. To
test these hypotheses, samples of SPE/DPPC were
nent of the SPE/DPPC mixtures.
A small
isotropic component was observed in the 31P
NMR spectra of SPE/DMPC mixtures at 60°C
which increases in intensity, up to highest temper-
ature measured 70°C. Upon cooling the isotropic
component vanished at the same temperature that
it was first observed. 31P NMR spectra of SPE/
DEPC mixtures did not show the presence of
non-bilayer phase.
3.5. DPPC-d62 2H NMR spectroscopy
In the present work we have determined how
SPEs influence lipid acyl chain dynamics and/or
conformational properties. This may be related to
the mechanism by which they induce nonlamellar
structure in aqueous dispersions of SPE-phospho-
lipid. In order to characterize the influence of SPE
on the order parameter profile of lipid acyl chains,
2H NMR spectra were obtained for aqueous dis-
persions of SPE/DPPC-d62 mixtures (all DPPC-
d62 measurements were made from a mixture of
10 mol% of DPPC-d62 in DPPC) in the La phase
as shown in Fig. 8. Pure DPPC-d62 gives rise to
the expected typical La phase spectrum (Fig. 8)
consisting of a superposition of axially symmetric
line shapes having quadrupolar splittings charac-
teristic of lipids undergoing fast reorientation
around their long molecular axis (Seelig, 1977;
Davis, 1979, 1983). In agreement with the DSC

G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
195
2
Fig. 8. H NMR spectra of DPPC-d62 and DPPC-d62 in the presence of 1 mol% of the different types of SPEs and 10 mol% of SOP
at 45°C.
and 31P NMR results, above, the spectral line
shapes indicate that all DPPC-d62/SPE mixtures
at 45°C were present in the La phase. In order to
obtain a more detailed acyl chain order parameter
profile, the segmental order parameters, SCD, were
extracted from the spectra by the procedure of
Bloom and co-workers (Bloom et al., 1981) and
are plotted in Fig. 9 as a function of carbon
number. It should be noted that some of the
spectra in Fig. 8, in particular those containing
SOO, SHxP and SOP, exhibit lineshapes that may
be reflective of a non-random distribution bilayer
orientations. In these cases the use of the proce-
dure of Bloom and co-workers (Bloom et al.,
second spectrum was simulated using an ellip-
soidal distribution of bilayer orientations (Pott
and Dufourc, 1995). The order profiles extracted
from the two spectra were found to be identical to
within digital resolution of the spectra (data not
shown) and corresponded to within a few percent
of the values used to simulate the spectra. For the
present study it is expected that application of the
procedure of Bloom and co-workers (Bloom et
al., 1981) to the spectra in Fig. 8 allow the order
profile along the acyl chain to be extracted reli-
ably. It is clear that at the low SPE contents that
trigger non-lamellar phase transitions in DPPC,
the acyl chain order profiles show no dramatic
differences in overall shape or in magnitude. To
characterize the overall effect of the different
SPEs on the molecular ordering of the lipid acyl
chain, the average order parameter BSCD\, and
the average length of the acyl chain BL\
(Schindler and Seelig, 1975; Seelig, 1975), were
calculated (Table 2). The values for the DPPC-d62
BSCD\ and BL\ at 45°C are consistent with
2
1981) to extract order parameters from H NMR
powder spectra is of concern. To establish that
applying this analysis to these spectra does not
lead to gross errors in the extracted order parame-
2
ters, the analysis was applied to two simulated H
powder spectra in which the order parameter
profile was identical. For one spectrum, a random
distribution of bilayer orientations was used. The

196
G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
the SPE molecules intercalate into the bilayer and
have effects on the lipid bilayer that are depen-
dent on details of the SPE molecular structure.
Thus far, our experiments have been concerned
primarily with the effects of low concentrations of
SPEs on phospholipid phase behaviour. One may
anticipate that at higher concentrations SPE
might have a greater effect on DEPE and DPPC
polymorphism.
the values reported previously for DPPC-d62
(Seelig, 1975; Davis, 1979). Although, the SPE
content in the lipid mixtures is only ca 1 mol%,
general trends for the SPE effect on DPPC-d62
SCD profile are discernable. The shapes and values
of the SCD profiles are essentially unaffected by
the intercalation of unsaturated SPE. Inspection
of Fig. 9 and Table 2 indicates that the SPE
containing the unsaturated acyl chains, SOO and
SOL, do not greatly influence acyl chain organiza-
tion and monolayer thickness. Both saturated
SHxP and SOP induced a small, but noticeable,
increase in the orientational ordering in the acyl
chain region of carbon 2 to 14 which corresponds
to an increase of monolayer thickness. This is in
contrast to SOS which induced a decrease in the
same acyl chain region causing the monolayer to
compress. These results indicate that at 1 mol%
3.6. Concentration effect (10 mol%)
The effect of a larger concentration, 10 mol%,
of SOP-d248 and 10 mol% of SHxP on the DEPE
2
polymorphism was probed with DSC, H and 31P
NMR measurements. The 31P NMR powder pat-
tern (Fig. 10A) at 35 and 45°C indicated that the
DEPE was in the lamellar Lb and La phase,
respectively. At 50°C the majority of the DEPE
was converted into the HII phase. Above 50°C up
to 65°C the powder patterns were representative
of the non-bilayer HII phase. The 31P NMR spec-
tra of 10 mol% of SHxP showed a similar effect
on DEPE. From the DSC thermogram (Fig. 10B)
it can be seen that the main phase transition (Tm)
is present at 38°C, but that the La/HII phase
transition is no longer visible. In addition a new
peak is seen at 44°C. This peak is attributed to the
melting of the crystalline of SOP-d248 based on the
31P NMR and SOP-d248 melting point.
Samples containing 1 and 10 mol% of SOP-
d248, in DPPC, were probed with 2H NMR to
monitor changes in the physical state of SOP-d248
as a function of concentration at 35, 43 and 47°C.
2
The temperature dependent H NMR of SOP-d248
mixed with DPPC at 1 and 10 mol% as a function
of temperature are shown in Fig. 11. The 2H
spectra of the sample containing 10 mol% SOP-
d248 at 35 and 43°C are representative of axially
asymmetric powder patterns. These lineshapes
were previously characterized as a two site trans–
gauche isomerization of the SOP-d248 acyl chains
(Buchanan et al., 2000). Upon heating to 47°C,
two degrees above SOP-d248 melting temperature
(Buchanan et al., 2000). The 2H lineshape col-
lapses into a motionally averaged isotropic signal.
Conversely, the 1 mol% SOP-d248 2H spectrum at
Fig. 9. Comparison of order parameter profiles of DPPC-d62
with different types of SPEs at 45°C.

G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
197

198
G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
Fig. 11. 2H NMR spectra of 1 and 10 mol% of SOP-d248 in DPPC as a function of temperature.
35°C shows a superposition of an axially asym-
metric and a symmetric powder pattern. The
former powder pattern is attributed to non-inter-
calated crystalline SOP-d248. The latter is at-
tributed to the fraction of the SOP-d248 that is
intercalated in DPPC. At 43°C, one degree above
the Tm of DPPC, the SOP-d248 spectra is mainly
characterized by an isotropic signal superimposed
on an axially asymmetric powder pattern. The
collapse of the axially symmetric into a motion-
ally averaged isotropic signal, at a temperature
below SOP-d248 melting point and above DPPC
Tm temperature, indicates that the intercalated
the presence of an isotropic signal (data not
shown).
Finally, to examine the effect of excess, non-in-
tercalated SPE on the structural and dynamic
properties of lipid acyl chains, 10 mol% SOP/
DPPC-d62 and 10 mol% SHxP/DPPC-d62 samples
were examined with 2H NMR. The resulting
powder patterns and order parameters were simi-
lar to the pure DPPC-d62 sample (Figs. 8 and 9
respectively).
These results further confirmed that at concen-
trations of B1 mol% SPE can integrate into
phospholipid membranes, and that the excess
SPE accumulates outside the membrane in either
the crystalline, when below the SPE melting
point, or melted state when above. Thus, at low
concentration the SPE will preferably intercalate
into the phospholipids to avoid contact with the
aqueous phase, as much as possible, until it
reaches an intercalation saturation point of B1
mol%. At higher concentrations the SPE most
likely aggregates outside the membrane generat-
ing a SPE rich phase. This pool of SPE being
more hydrophobic than the phospholipids would
then favour the transfer of the intercalated SPE
into a SPE rich phase outside the membrane.
2
SOP-d248 are highly mobile, on the H NMR time
scale, in the lipid La phase. As mentioned above
it was shown from 31P NMR that all 1 mol%
SPE/lipid mixtures induced the formation of cu-
bic phases. The rapid diffusion of the SOP-d248
over the cubic phases high curvature structure
would offer such a rapid reorientational averag-
ing so as to generate an isotropic signal. For this
same sample and for a 10 mol% SHxP/DPPC
sample the 31P NMR powder pattern indicated
that the DPPC at 35°C was in the Lb phase with
a residual CSA of ca 72 ppm. Above 42°C both
samples spectra where characteristic of the La
phase with a residual CSA of ca 48 ppm, without

G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
199
4. Discussion
effect on the DEPE HII transition temperature.
The later suggests that SOP and SOS are soluble
in the DEPE bilayer only to a very small extent
and then with further addition result in the for-
mation of crystalline SOP- or SOS-rich domains.
Evidence of such crystalline domains, rich in SOP
or SOS is demonstrated by the presence of the
DSC endothermic transition occurring at temper-
atures similar to the saturated sucrose ester crys-
talline melting points. Such a transition is not
accompanied by any DEPE bilayer conforma-
tional or motional changes as monitored by 31P
NMR. Comparable low bilayer incorporation was
also reported for TS in DEPE (Epand et al., 1988)
and for TP in egg PC multilamellar dispersion
(Gorrissen et al., 1982; Hamilton et al., 1991).
However, even if the saturated and unsaturated
SPE parallel the TAG effect on the lipid bilayer,
the SPEs have the ability to stabilize the HII phase
transitions at concentrations 10-fold lower than of
TAG. A possible explanation may be attributed
to SPEs location and conformation in the lipid
bilayer. 13C NMR studies have demonstrated that
saturated and unsaturated TAG and cholesteryl
ester (CE) in egg PC bilayer vesicles and multil-
amellar dispersions are localized at the bilayer
surface. Such studies have concluded that the
carbonyl groups of these molecules interact with
the aqueous–lipid interface region of the bilayer
and the acyl chains and the steroid rings aligned
with the acyl chains of the PC (Hamilton and
Small, 1982; Hamilton, 1989; Hamilton et al.,
To date investigations regarding SPE have been
focused on studies of physiological effects of SPE
on animals and humans (Crouse and Grundy,
1979; Jandacek, 1982; Jandacek et al., 1990;
Mattson and Jandacek, 1991; Jandacek et al.,
1999). There is a lack of information concerning
the potential effects of SPE on cell membranes. A
better understanding of the factors which influ-
ence SPE effects on biological membranes is of
importance in defining the range of current and
potential in vivo uses of SPE and may provide a
suitable model for physiologically important gly-
colipids and bacterial endotoxin such as lipid A.
The combined use of DSC, X-ray diffraction,
2
31P, and H solid state NMR methods provided a
detailed description of organizational changes in
lipid bilayers induced by the interaction of addi-
tives with the bilayer. The results presented here
show that all SPEs investigated have the ability to
insert into a lipid bilayer, to act as HII phase
stabilizers and as an initiator of formation of a
cubic phase at very low SPEs content. SPEs inter-
actions with a lipid bilayer appear to be strongly
dependent on the level of esterification of the
sucrose moiety, on the degree of acyl chain satu-
ration, on the length of the acyl chains and on the
concentration of SPE.
The DSC data (Fig. 2) indicate that all SPE
studied have the ability to intercalate into a lipid
bilayer where the level of interaction with the lipid
bilayer was shown to be dependent on the SPE
structural features such as the sugar hydroxyla-
tion level and the type of acyl chains. The sucrose
octaesters’ potent ability to reduce the TH of
DEPE clearly contrasts with the effect of SHxP at
very low mole fractions. Thus the ability of SPEs
to promote the HII phase transition is dependent
on the degree of sucrose acylation. Furthermore,
the unsaturated SPE have a greater ability to
reduce the TH of DEPE than the saturated SPE.
The mono-unsaturated SOO has a greater ability
to depress TH than the di-unsaturated SOL at 1
mol%. The saturated SOP and SOS effects on the
thermotropic phase behaviour of DEPE are simi-
lar to saturated TP and TS (Epand et al., 1988).
Above 7 mol%, SOP and SOS have no further
2
1991). H solid state NMR of selectively deuter-
ated crystalline SOP showed that the SOP confor-
mation can be depicted as having all eight acyl
chains extend both above and below a rigid disac-
charide unit and that the molecular conformation
was shown to be analogous to the crystalline
TAG which has its acyl chains extending on both
sides of the rigid glycerol backbone (Buchanan et
al., 2000). Therefore, it is reasonable to infer that
once in a aqueous–lipid milieu, the SPEs are
probably localized in a similar manner to that of
TAG, where the carbonyl groups of the rigid
disaccharide moiety are oriented at the aqueous-
lipid interface and where all acyl chains are ex-
tended along the lipid acyl chains, spanning the
thickness of only one monolayer. Molecular mod-

200
G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
elling was utilized to evaluate the possibility of
such an SPE conformation. The models were
made by using a sucrose back-bone from which
the eight or six hydroxyl were esterified with the
C16 acyl chains to represent the SOP and SHxP,
respectively. Gas-phase geometry optimization us-
ing the MM+ (Allinger, 1977) force-field, as
implemented in the Hyperchem 5.11 package, pre-
dicts the structure illustrated in Fig. 12 as a local
minimum. From these models, the disaccharide
moieties of SOP and SHxP have maximum cross
of non-lamellar organization characterized by an
isotropic 31P NMR spectrum. This transition ex-
hibits a thermal hysteresis where the transition
from non-lamellar to lamellar phase occurs only
when the samples are cooled to temperature be-
low the gel phase. Furthermore, the ability of
these non-bilayer components to form at tempera-
tures below the La/HII phase transition and to
remain present after repeated cycling through the
HII phase transition is indicative of the formation
of a three-dimensional lipid aggregate structure
such as cubic phase that is more thermodynami-
cally stable then the La or HII phase over a wide
temperature range. X-ray diffraction was used to
further characterize the isotropic component in
,
section diameters of 13.6 and 12.2 A, respectively.
The depicted SPE conformation and intercalation
into a bilayer can also be related to that of the
Gram-negative bacterial disaccharide glycolipid
endotoxin, lipid A (Nikaido et al., 1977). More-
over, it was shown that lipid A tends to segregate
into domains away from phospholipids. Consider-
ing the higher hydrophobic potential of SPE rela-
tive to the amphipathic phospholipid, it seems
quite plausible that the SPE will tend to aggregate
into patches of polymers, rather then be homoge-
neously distributed as individual monomers
throughout the phospholipids matrix. Further
confirmation of the SPE location in the bilayer is
evident from the ability of SPE to induce the
formation of non-lamellar structures.
,
SPE/DEPE mixtures. A spacing of 100 A was
detected, indicating the presence of a structure
with long periodicity. These results suggest the
presence of cubic phase structures in SPE/DEPE
,
mixtures. The 100 A spacing formed at 55°C in a
time-dependent manner. This temperature is be-
low TH, indicating that the structure formed is
more stable at this temperature than either the La
or HII phase.
The capability of SPE to induce the formation
of non-lamellar phases can be rationalized by
considering the physical factors that affect lipid
polymorphism. The ability of lipids to form non-
lamellar phase has been proposed to be dependent
on the intrinsic lipid monolayer curvature as well
as interstitial packing and hydration (Gruner,
In the case of aqueous dispersions of DEPE,
31P NMR reveals that SPEs induce the formation
2
1985). Based on the H NMR of DPPC-d62, the
SCD profile shows that the SPEs only induce
changes in the plateau region of the lipid acyl
chains, leaving the last methylene and the methyl
groups unchanged. This provides evidence that
SPE does not act by perturbing the hydrophobic
region significantly. Thus, it is unlikely that the
bulky disaccharide moiety is found intercalated
deeply in the hydrophobic regions of the lipid.
Therefore, instead of relaxing the intersticial
packing constraint between monolayer tubes, the
capability of SPEs to induce non-lamellar phase
formation must be associated with a decrease in
spontaneous radius of curvature, caused by dehy-
dration of the lipid, due to the intercalation of the
disaccharide moiety in-between the lipid head-
groups at the aqueous lipid–water interface.
Fig. 12. Geometry optimized representation of the suggested,
bilayer intercalated, SOP and SHxP conformation. Note the
symmetric barrel and cone shape of SOP (right) and SHxP
(left) respectively.

G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
201
Although SPE effects on lipid polymorphism
are largely a consequence of interactions near the
membrane–water interface, the difference in the
effects caused by different SPEs can be rational-
ized in terms of the nature of the acyl chains. For
instance, it is well known that the presence of one
or two cis double bonds in the acyl chains has the
effect of reducing the length of the acyl chain
relative to corresponding saturated acyl chain and
leads to an increased acyl chain splay. Thus, SOO
and SOL would have a larger effect in promoting
non-lamellar phases than the saturated SPEs.
recently, it has been demonstrated that (Thurn-
hofer et al., 1991) the lipid absorption mechanism
involves protein mediated bilayer perturbation of
the external surface of the small intestine brush
border membrane. The exact role of these lipid
phases during fat digestion and absorption are
not yet completely understood.
References
Akoh, C.C., 1995. Crit. Rev. Food Sci. Nutr. 35, 405–430.
Akoh, C.C., Swanson, B.G., 1990. J. Food Sci. 55, 236–243.
Allinger, N.L., 1977. J. Am. Chem. Soc. 99, 8127.
Asai, Y., Iwamoto, K., Watanabe, S., 1998. FEBS Lett. 438,
15–20.
4.1. Physiological rele6ance
That SPEs can destabilize bilayer structures of
simple multilamellar liposomes composed of
phosphatidylethanolamine or phosphatidylcholine
at physiologically relevant temperatures is signifi-
cant. Of particular importance is that their effect
is induced when SPEs are inserted into mem-
branes at concentrations as low as 1 mol% (ap-
prox. 4 wt.%). The present study suggests that
SPEs insert into membranes so that their carbonyl
residues are proximal to the aqueous interface and
that they may exhibit a tendency to aggregate. As
a result, the local bilayer structure is perturbed
sufficiently to favour formation of non-lamellar
structures.
That SPE has the potential to disrupt mem-
brane bilayer structure raises questions as to its
use in vivo. To date, no evidence has been re-
ported that SPE have deleterious effects in vivo
that might be associated with disruption of cell
membranes. However, an understanding of the
potential effects SPEs have on cell membranes
holds promise for designing better SPE molecules
or SPE mixtures that are tailored for specific
physiological use such as decreasing the levels of
low density LDL cholesterol or of lipophilic envi-
ronmental contaminants, and as an oral contrast
agent for human abdominal MRI studies. That
SPEs induce non-lamellar phases may be useful in
their role as a dietary fat substitute. It has been
suggested that the fat digestion and absorption
process involves a number of bilayer and non-bi-
layer lipid phases such as cubic phases (Patton
and Carey, 1979; Lindstrom et al., 1981). More
Ballinger, R., Magin, R.L., Webb, A.G., 1991. Magn. Reson.
Med. 19, 199–202.
Bloom, M., Davis, J.H., MacKay, A.L., 1981. Chem. Phys.
Lett. 80, 198–202.
Brandenburg, K., Mayer, H., Koch, M.H., Weckesser, J.,
Rietschel, E.T., Seydel, U., 1993. Eur. J. Biochem. 218,
555–563.
Brandenburg, K., Koch, M.H., Seydel, U., 1998. Eur. J.
Biochem. 258, 686–695.
Buchanan, G.W., McManus, G., Jarrell, H.C., 2000. Chem.
Phys. Lipids 104, 23–34.
Chernomordik, L., 1996. Chem. Phys. Lipids 81, 203–213.
Choi, K.B., Wong, F., Harlan, J.M., Chaudhary, P.M., Hood,
L., Karsan, A., 1998. J. Biol. Chem. 273, 20185–20188.
Christ, W.J., Asano, O., Robidoux, A.L., Perez, M., Wang,
Y., Dubuc, G.R., et al., 1995. Science 268, 80–83.
Cock de, H., Brandenburg, K., Wiese, A., Holst, O., Seydel,
U., 1999. J. Biol.Chem. 274, 5114–5119.
Crouse, J.R., Grundy, S.M., 1979. Metabolism 28, 994–1000.
Cullis, P.R., de Kruijff, B., 1978a. Biochim. Biophys. Acta
513, 31–42.
Cullis, P.R., de Kruijff, B., 1978b. Biochim. Biophys. Acta
507, 207–218.
Cullis, P.R., De Kruyff, B., Richards, R.E., 1976. Biochim.
Biophys. Acta 426, 433–446.
Davis, J.H., 1979. Biophys. J. 27, 339–358.
Davis, J.H., 1983. Biochim. Biophys .Acta 737, 117–171.
Ellens, H., Siegel, D.P., Alford, D., Yeagle, P.L., Boni, L., Lis,
L.J., et al., 1989. Biochemistry 28, 3692–3703.
Epand, R.M., 1985. Biochemistry 24, 7092–7095.
Epand, R.M., 1998. Biochim. Biophys. Acta 1376, 353–368.
Epand, R.M., Bottega, R., 1988. Biochim. Biophys. Acta 944,
144–154.
Epand, R.M., Epand, R.F., Lancaster, C.R., 1988. Biochim.
Biophys. Acta 945, 161–166.
Fenske, D.B., Hamilton, K., Jarrell, H.C., Florio, E., Barber,
K.R., Grant, C.W., 1991. Biochemistry 30, 4503–4509.
Gagne, J., Stamatatos, L., Diacovo, T., Hui, S.W., Yeagle,
P.L., Silvius, J.R., 1985. Biochemistry 24, 4400–4408.

202
G.G. McManus et al. / Chemistry and Physics of Lipids 109 (2001) 185–202
Mellies, M.J., Vitale, C., Jandacek, R.J., Lamkin, G.E.,
Glueck, C.J., 1985. Am. J. Clin. Nutr. 41, 1–12.
Moser, G.A., McLachlan, M.S., 1999. Chemosphere 39,
1513–1521.
Geusau, A., Tschachler, E., Meixner, M., Sandermann, S.,
Papke, O., Wolf, C., et al., 1999. Lancet 354, 1266–1270.
Giorgione, J.R., Huang, Z., Epand, R.M., 1998a. Biochem-
istry 37, 2384–2392.
Mosior, M., Epand, R.M., 1999. J. Lip. Res. 9, 21–41.
Munstermann, M., Wiese, A., Brandenburg, K., Zahringer,
U., Brade, L., Kawahara, K., et al., 1999. J. Membr. Biol.
167, 223–232.
Giorgione, J.R., Huang, Z., Epand, R.M., 1998b. Biochem-
istry 37, 2384–2392.
Gorrissen, H., Tulloch, A.P., Cushley, R.J., 1982. Chem. Phys.
Lipids 31, 245–255.
Mutter, L.C., Blanke, R.V., Jandacek, R.J., Guzelian, P.S.,
1988. Toxicol. Appl. Pharmacol. 92, 428–435.
Nakamura, A., Johns, E.J., Imaizumi, A., Yanagawa, Y.,
Kohsaka, T., 1999. Kidney Int. 56, 839–849.
Nikaido, H., Takeuchi, Y., Ohnishi, S.I., Nakae, T., 1977.
Biochim. Biophys. Acta. 465, 152–164.
Otake, T., 1972. Carbohydr. Res. 45, 2895–2898.
Patton, J.S., Carey, M.C., 1979. Science 204, 145–148.
Pott, T., Dufourc, E., 1995. Biophys. J. 68, 965–977.
Prince, D.M., Welschenbach, M.A., 1998. J. Am. Diet Assoc.
98, 565–569.
Rance, M., Byrd, A., 1983. J. Magn. Reson. 52, 221.
Rietschel, E.T., Kirikae, T., Schade, F.U., Mamat, U.,
Schmidt, G., Loppnow, H., et al., 1994. Faseb. J. 8,
217–225.
Schindler, H., Seelig, J., 1975. Biochemistry 14, 2283–2287.
Schlummer, M., Moser, G.A., McLachlan, M.S., 1998. Toxi-
col. Appl. Pharmacol. 152, 128–137.
Seelig, J., 1977. Q. Rev. Biophys. 10, 353–418.
Seelig, J., 1975. Z. Klin. Chem. Klin. Biochem. 13, 368.
Sen, A., Hui, S.W., Mannock, D.A., Lewis, R.N., McElhaney,
R.N., 1990. Biochemistry 29, 7799–7804.
Seydel, U., Brandenburg, K., Rietschel, E.T., 1994. Prog. Clin.
Biol. Res. 388, 17–30.
Shyamsunder, E., Gruner, S.M., Tate, M.W., Turner, D.C.,
So, P.T., Tilcock, C.P., 1988. Biochemistry 27, 2332–2336.
Siegel, D.P., 1999. Biophys. J. 76, 291–313.
Thayer, A.M., Kohler, S.J., 1981. Biochemistry 20, 6831–
6834.
Gruner, S.M., 1985. Proc. Natl. Acad. Sci. USA 82, 3665–
3669.
Gruner, S.M., Tate, M.W., Kirk, G.L., So, P.T., Turner, D.C.,
Keane, D.T., et al., 1988. Biochemistry 27, 2853–2866.
Hamilton, J.A., 1989. Biochemistry 28, 2514–2520.
Hamilton, J.A., Small, D.M., 1982. J. Biol. Chem. 257, 7318–
7321.
Hamilton, J.A., Fujito, D.T., Hammer, C.F., 1991. Biochem-
istry 30, 2894–2902.
Hough, L., 1972. Carbohydr. Res. 21, 144–147.
Jandacek, R.J., 1982. Drug Metab. Rev. 13, 695–714.
Jandacek, R.J., Webb, M.R., 1978. Chem. Phys. Lipids 22,
163–176.
Jandacek, R.J., Ramirez, M.M., Crouse, J.R.D., 1990.
Metabolism 39, 848–852.
Jandacek, R.J., Kester, J.J., Papa, A.J., Wehmeier, T.J., Lin,
P.Y., 1999. Lipids 34, 771–783.
Jimenez-Monreal, A.M., Aranda, F.J., Micol, V., Sanchez-
Pinera, P., de Godos, A., Gomez-Fernandez, J.C., 1999.
Biochemistry 38, 7747–7754.
Kawata, T., Bristol, J.R., Rossignol, D.P., Rose, J.R.,
Kobayashi, S., Yokohama, H., et al., 1999. Br. J. Pharma-
col. 127, 853–862.
Keller, S.L., Gruner, S.M., Gawrisch, K., 1996. Biochim.
Biophys. Acta 1278, 241–246.
Kelly, S.M., Shorthouse, M., Cotterell, J.C., Riordan, A.M.,
Lee, A.J., Thurnham, D.I., et al., 1998. Br. J. Nutr. 80,
41–49.
Koonsvitsky, B.P., Berry, D.A., Jones, M.B., Lin, P.Y.,
Cooper, D.A., Jones, D.Y., et al., 1997. J. Nutr. 127,
1636S–1645S.
Lindstrom, M., Ljusberg-Wahren, H., Larsson, K.,
Borgstrom, B., 1981. Lipids 16, 749–754.
Matteson, F.H., Volpenhein, R.A., 1972. J. Lipid. Res. 13,
325–328.
Thurnhofer, H., Lipka, G., Hauser, H., 1991. Eur. J. Biochem.
201, 273–282.
Tilcock, C.P., Cullis, P.R., Hope, M.J., Gruner, S.M., 1986.
Biochemistry 25, 816–822.
Veiro, J.A., Khalifah, R.G., Rowe, E.S., 1990. Biophys. J. 57,
637–641.
Zahringer, U., Lindner, B., Rietschel, E.T., 1994. Adv. Carbo-
hydr. Chem. Biochem. 50, 211–276.
Mattson, F.H., Jandacek, R.J., 1991. Lipids 26, 750–753.
.