PREPARATION AND CHARACTERIZATION OF MICROEMULSIONS CONTAINING SPHINGOMYELIN AND CHOLESTERYL OLEATE: A 3lP NMR STUDY Lisa Yajie Zhao B. Sc., Hunan University, 1984 THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in the department of chemistry O Lisa Yajie Zhao 1993 SIMON FRASER UNIVERSITY June 1993 All rights reserved. This work may not be reproduced in whole or in part, by photocopy or other means, without permission of the author
127
Embed
Preparation and characterization of microemulsions ...summit.sfu.ca/system/files/iritems1/7926/b15206634.pdfpreparation and characterization of microemulsions containing sphingomyelin
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
PREPARATION AND CHARACTERIZATION OF MICROEMULSIONS
CONTAINING SPHINGOMYELIN AND CHOLESTERYL OLEATE: A 3lP NMR STUDY
Lisa Yajie Zhao
B. Sc., Hunan University, 1984
THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
in the department
of
chemistry
O Lisa Yajie Zhao 1993
SIMON FRASER UNIVERSITY
June 1993
All rights reserved. This work may not be
reproduced in whole or in part, by photocopy
or other means, without permission of the author
Approval
Name: Lisa Yajie Zhao
Degree: Master of Science
Title of Thesis: PREPARATION AND CHARACTERIZATION OF MICROEMULSIONS CONTAINING SPHINGOMYELIN AND
CHOLESTERYL OLEATE: A 3lP NMR STUDY
Examining Committee:
Chairman: Dr.
Dr. R. J. Cushley Senior Supervisor
Dr. Y. L. Chow Supervisory committee
Dr. W. R. Richards Supervi~qry committee,
Dr. N. Haunerland Internal examiner
Date approved: '$$, 1993
PARTIAL COPYRIGHT LICENSE --
I hereby grant to Simon Fraser University the right to lend my
thesis, project or extended essay (the title of which is shown below) to
users of the Simon Fraser University Library, and to make partial or
single copies only for such users or in response to a request from the
library of any other university, or other educational institution, on its own
behalf or for one of its users. I further agree that permission for multiple
copying of this work for scholarly purposes may be granted by me or the
Dean of Graduate Studies. It is understood that copying or publication
of this work for financial gain shall not be allowed without my written
permission.
Title of Thesis/Project/Extended Essay:
Author: (signature)
" (date)
ABSTRACT
As a model for the lipid organization of lipoproteins, sphingomyelinl
cholesteryl oleate (SPMICO) microemulsions about 23-26 nm in diameter were
prepared, which is in the size range of LDL. A temperature study using
phosphorus-31 NMR showed that the linewidth for SPMICO microemulsions is
slightly larger than that of SPM vesicles. This suggests that the neutral core
(CO) of the microemulsions may modify the motion of the phospholipid (SPM)
monolayer.
In order to study the lipid-protein interaction, microemulsions in the
presence of protein (apo HDL3) were prepared. A temperature dependence
study indicated that the 31P NMR spectra could not be simulated using a single
Lorentzian lineshape function; instead, a superposition of two Lorentzians was
needed to get a reasonably good fit at low temperatures (below 25•‹C).
Lineshape analysis suggested the presence of two magnetically inequivalent
domains within the reconstituted lipoproteins.
The lateral diffusion coefficient (D) of sphingomyelin in SPM/CO
microemulsions was determined from the viscosity-dependence of 3lP NMR
linewidths. At 25 "C, D was 1 + 0.3 x 10-9 cm2 s-1 for SPM / CO
microemulsions. The value of D in SPM / CO microemulsions is approximately
1.4 times smaller than in native LDL. A possible explanation is that the core of
the microemulsions may play a role in slowing SPM diffusion, or the relatively
rigid sphingomyelin phospholipid monolayer may be responsible for the slower
diffusion.
The residual chemical shift anisotropy of sphingomyelin in SPh
microemulsions at 25•‹C was 33 ppm, measured by field-dependence of the 31 P
NMR linewidth. This value is similar to that found for egg PC vesicles and egg-
PCitriolein microemulsions, indicating that the three systems may have similar
headgroup orientations.
The lateral diffusion constant of sphingomyelin in SPMICOlapo HDL3
reconstituted particles was determined from the viscosity-dependence of 3lP
NMR linewidths. NMR spectra were fitted by a superposition of two Lorentzians.
The lateral diffusion constants for the two environmentally different domains
were determined to be 1 + 0.4 x 10-9 cm2 s-1 (broad component ) and 2.1 + 0.8
x 10-8 cm2 s-1 (narrow component), respectively. A possible explanation for two
quite different diffusion coefficients is that the diffusion constant of the broad
component is due to the "protein-deficient" phospholipids (or phospholipids not
close to the protein), whereas the diffusion constant of the narrow component is
due to the "protein-rich" phospholipids (or phospholipids adjacent to protein).
DEDICATION
To Daiqing
and
Jennifer
ACKNOWLEDGMENTS
I would like to thank my supervisor Dr. R. J. Cushley for his guidance,
encouragement and support throughout the course of this research project. I
also wish to extend my thanks to Dr. Richards and Dr. Chow for their time as
members of my supervisory committee.
To Dr. W. D. Treleaven, for his advice, expertise, encouragement,
friendship and warmth for the past two years, I am grateful.
To the members of the research group for their assistance.
To Fred Chin, for his computer assistance.
To Jennifer, for making my life much brighter, and bringing me all the
happiness I needed. Thank you, my dear baby.
To my parents and my sister for their love, support and the source of my
strength.
Finally, I would like to thank my husband, Daiqing Liao for everything. His
knowledge and full understanding make this degree possible.
TABLE OF CONTENTS
Title ...................................................................................................
Figure 2.4: proton decoupled 3lP NMR spectra of polymorphic phases
available to liquid crystalline phospholipids (Gennis, 1989).
A. Bilayer: egg phosphatidylcholine
B. Hexagonal (H It): Soya bean phosphatidylethanolamine C. Isotropic Motions: Small Unilamellar Vesicles,
Microemulsions
CHAPTER 3
MATERIALS AND METHODS
3.1 Materials
Bovine brain sphingomyelin and cholesteryl oleate were purchased from
Sigma Chemical Company, St. Louis, Missouri. No impurities were detected by
thin layer chromatography in chloroform/methanol/water (65:25:4). Lipids were
used without further purification. Aquacide I I (sodium salt of
carboxymethylcellulose, MW = 500,000) was purchased from Calbiochem
Corporation. Glycerol was purchased from Fisher Scientific Company. The lipid
analysis kits for determining total cholesterol were purchased from Boehringer
Mannheim.
Sonication buffer
The sonication buffer contains 100 mM KCI, 10 mM Tris-HCI. Na2C03
was used to adjust the pH of the buffer to 8.6. Solid KBr was used to raise the
buffer density to 1.21 g/ml.
Dialysis buffer
The dialysis buffer contains 100 mM KCI, 10 mM Tris-HCI. Na2C03 was
used to adjust the pH of the buffer to 8.6.
3.2 Methods
3.2.1 Isolation of High Density Lipoproteins (HDL)
HDL3 was isolated by sequential isopycnic floatation between the
densities p=1.125 and p=1.210 glml, at 4 "C using a Beckman L5-75
ultracentrifuge, and a Ti 50.2 rotor. Potassium bromide was used to raise the
density. Samples were centrifuged at 45,000 rpm for 24 hours at p = 1.1 25 glml,
and 48 hours at p=1.210 glml.
3.2.2 Isolation of HDL3 Apolipoproteins
HDL3 ( 0.45 ml, 10 to 15 mg protein per ml ) was added to one of the 30
ml glass centrifuge tubes. Approximately 25 ml of cold (-20 "C)
chloroform/methanol (CHC13:MeOH 2:1 vlv) was added to each tube. The
solution was incubated at -20 "C for 1 hour. The tubes were centrifuged for
approximately 5 minutes using a desktop centrifuge, and the solvent was
decanted from the precipitate. Apolipoproteins were incubated four times with
cold anhydrous diethyl ether (15 minutes each at -20•‹C ). Following the final
wash, the apolipoproteins were air dried, and stored at -20 "C. Analysis of the
apolipoproteins by SDS polyacrylamide gel electrophoresis revealed that no
human serum albumin was present.
3.2.3 Preparation of Microemulsions
Homogeneous microemulsions were prepared by the following
procedures (Darfler, 1990; Ginsburg et al., 1982). About 150 mg sphingomyelin
and 60 mg cholesteryl oleate were codissolved in chloroform / methanol (2: l)
mixture. The solvent was evaporated under a stream of nitrogen, and the lipid
mixture was further dried under high vacuum overnight. A lipid-buffer emulsion
was formed by mixing lipid with approximately 10 ml of sonication buffer. The
mixture was sonicated for 5.5 hours (9O0lO duty cycle) at 60 "C under nitrogen
using a Heat Systems W-375 sonicator operating at 30% output power. The
circulation temperature was monitored by a thermocouple system around the
sonication mixture. The sonication mixture was centrifuged for 15 minutes to
remove titanium particles, then was dialyzed against hypotonic saline buffer with
3 - 4 changes of 1 -liter each, over 14 hours. The sample was centrifuged at
50,000 rpm (1 65,0009) for 30 minutes(l5"C). The very top layer was removed
and saved for the light scattering measurement. The second top layers were
collected and recentrifuged again at 50,000 rpm from 15-30 hours. The bottom
vesicle-containing layer was discarded. The microemulsions were removed from
the top of the second spin. The microemulsions can be prepared reproducibly
with about 23-27 nm in diameter as shown by the light scattering measurement,
and these samples are stable for several weeks. The microemulsions were
further concentrated using either ultracentrifugation or aquacide at room
temperature. The microemulsions were then transferred to an NMR tube.
Temperature dependent 3 1 ~ NMR experiments were conducted immediately.
3.2.4 Sphingomyelin Vesicles
Sphingomyelin vesicles were prepared according to the procedure of
Barenholz (Barenholz et al., 1976). In short, sphingomyelin was suspended in
50 mM KCI; the lipid suspensions were sonicated using a Heat Systems
Sonicator (W-375) under nitrogen for 30 minutes. The sonication temperature
(0•‹C) was controlled by an ice-water mixture. Following sonication, the vesicles
were separated by ultracentrifugation. The vesicles were prepared reproducibly
with approximately 25 - 27 nm in diameter. Vesicles were stable for several
days to two weeks. The temperature dependence of 31 P NMR spectra was
recorded.
3.2.5 Quasi-Elastic Light Scattering
Quasi-elastic light scattering measurements were performed using a
Nicomp model 270 or 370 submicron particle sizer.
3.2.6 Preparation of Reconstituted Lipoproteins
The appropriate apoproteins were dialysed against buffer (1 00 mM KCI,
10 mM Tris HCI, Na2C03, density = 1.005 glml, pH = 8.6) for several hours. The
apoprotein solution was transferred to a vial, and the microemulsion solutions
were added to the vial very gently. The apolipoprotein-microemulsion mixture
was stirred and incubated at 50 + 1 "C for 18 hours. 31 P NMR experiments at
various temperatures were conducted using these samples. After NMR
measurements, protein concentration of the reconstituted lipoprotein was
measured by the Lowry method (Lowry et al., 1951). The phospholipid
concentration in recombinant lipoprotein was measured as described (Ames,
1966). Cholesteryl oleate concentration in recombinant lipoprotein was
determined with the specific assay kits from Boehringer Mannheim.
3.2.7 NMR Measurements
The 3lP NMR experiments were performed at 102.2 MHz without proton
decoupling using a home-built spectrometer and a 5.9 T Nalorac
superconducting magnet. Data collection and the Fourier transformation were
performed on a Vax station I. Temperature was controlled by a solid state
temperature controller built by the Simon Fraser University electronics shop with
an accuracy of + 0.25 "C. The signal to noise ratio was enhanced by using
exponential multiplication. The samples were allowed to equilibrate for 30 min.
at a given temperature before data were acquired. The spectral parameters are
given in the figure legends.
For the field dependence studies, 3lP NMR spectra were collected at 40
MHz (SY-100),102.2 MHz (home-built NMR), 160.2 MHz (Bruker AMX-400), and
202.46 MHz (Bruker AMX-500) .
3.2.8 Diffusion Measurements
Glycerol was added to the freshly prepared samples (SPMICO
microemulsions and SPMICOlapoHDL3 reconstituted particles) to increase the
viscosity. The viscosity of SPMICO microemulsions and reconstituted
measured using an Ostwald viscometer. 31 P NMR spectra were collected at
each viscosity at 25 "C.
3.2.9 NMR Lineshape Analysis
Phosphorus NMR spectra of SPMICO microemulsions, sphingomyelin
vesicles and reconstituted lipoproteins were analyzed using a four parameter
(root-mean-square base line, signal amplitude, and estimated linewidth for each
signal plus the chemical shift) iterative least-squares fit of the resonance to a
single Lorentzian function. The 3IP spectra of reconstituted lipoprotein at 15 "C
and 25 "C were analyzed using a seven parameter (baseline, signal amplitude of
the broad component, chemical shift of broad component, linewidth at half-height
of broad component, signal amplitude of narrow component, chemical shift of
narrow component, linewidth at half-height of the narrow component) iterative
least-squares fit of the phosphorus resonance to a superposition of two
Lorentzian functions.
P
CHAPTER 4
RESULTS AND DISCUSSION
4.1 Unilamellar Vesicles
The unilamellar vesicles were made successfully by using the method of
Barenholz (Barenholz et al., 1976). As described in Chapter 3, sphingomyelin
was dissolved in chloroform and dried under vacuum for about 2 hours, and then
allowed to interact with buffer at pH 8.6 and specific ionic strength (100 mM KCI).
They then swelled to form multilamellar liposomes. Multilamellar liposomes were
subsequently subjected to ultrasonic irradiation to produce small unilamellar
vesicles. The vesicles were 25 nm in diameter and contained a single bilayer.
Figure 4.1 shows the size distribution of sphingomyelin vesicles determined by
light scattering (the particles in the range of 60-100 nm are large unilamellar
vesicles). The inner and outer leaflet of the bilayers were shown not to be
equivalent as demonstrated by 3lP nuclear magnetic resonance studies using
the chemical shift reagent Pr 3+ to demonstrate the difference between the inner
and outer layer (Yeagle, 1990). The homogeneous vesicles are also thought to
provide a surface that has curvature similar to low density lipoprotein (Diameter
= - 25 nm). Therefore, we may extrapolate from the vesicle system to the
biomembrane or lipoprotein system. But caution should be exercised when one
attempts to make some extrapolation to the native lipoprotein because of the
simplicity of the small unilamellar vesicles.
Figure 4.1 : The size distribution of sphingomyelin vesicles by Quasi-Elastic
Light Scattering.
i representative 3' P NMR spectra of sphingomyelin vesicles at selected
temperatures. The solid line represents an iterative least-square fit of single
Lorentzian function to the data (crosses). The linewidths of sphingomyelin
vesicles at different temperatures over the range 15 "C to 50 "C are shown in
Table 4.1. The trend of decreasing linewidth as temperature was raised is
reversible. From Table 4.1, we can see that below 45 "C, the linewidth of 31 P
NMR spectra shows great temperature dependence. Above 45 "C, the linewidth
of 31P NMR spectra shows slight temperature dependence.
4.2 Microemulsions: Models for Smaller Cholesterol-Rich Lipoproteins
As models to study the lipid organization in lipoproteins, protein free
homogeneous microemulsions have been prepared using specific phospholipid
and cholesteryl esters, which constitute the important components of
lipoproteins. The microemulsion system was formed with a single species of
phospholipid, therefore, fundamental details of the lipid-lipid and lipid-protein
interaction could be studied without the complexity of heterogeneous molecular
compositions found in the native lipoproteins.
Typically microemulsions are formed by extensive sonication of both
phospholipid and cholesteryl ester in an aqueous solution at a temperature
above the main order-disorder transition temperature of both lipid components.
Fractionation was carried out by ultracentrifugation. Ultracentrifugation is the
most widely used technique for lipoprotein separation because particles can be
separated from each other on the basis of difference in the buoyant densities or
Figure 4.2: 31P NMR spectra of sphingomyelin vesicles a). 15 "C, b). 20 "C, c). 25 "C, d). 50 "C The spectra were simulated by an iterative least-squared fit to a Lorentzian lineshape function (solid line) to the spectral data
points (crosses). Spectral parameters for all cases are as following: pulse width = 6.5 ps (90" flip angle), sweep width = 10,000 Hz,
delay between pulses = 2.5 s, dataset = 2K zero filled to 4K, dwell time = 50 ps, line broadening = 5 Hz, Number of scans =
10,000.
Figure 4.2 (a): The P NMR spectrum of sphingomyelin vesicles at 15 "C
Figure 4.2 (b): The P NMR spectrum of sphingomyelin vesicles at 20 "C
42
Figure 4.2 (c): The 31P NMR spectrum of sphingomyelin vesicles at 25 "C
Figure 4.2 (d): The 31P NMR spectrum of sphingomyelin vesicles at 50 "C
Table 4.1: The linewidth of sphingomyelin vesicles at selective
temperature
I Temperature ( "C ) Linewidth ( Av 1/, ) (Hz)*
I * Numbers in parentheses are about 10O/0 of the linewidths
on the basis of their sedimentation rate. The density of the sonication buffer was
adjusted to 1.21 glml before sonication. Following sonication, the sample was
dialyzed against the dialysis buffer (1 00 mM KCI, 10 mM Tris-HCI, density=
1.005 glml, pH = 8.6). Na2C03 was used to adjust the pH of the dialysis buffer
to a desired value. The reason for doing this is as follows: After sonication for
five and half hours, the products of the sphingomyelin and cholesteryl oleate
mixture were small-sized microemulsions, large-sized microemulsions,
multilamellar liposomes and small unilamellar vesicles. When the density of the
sonication buffer was adjusted to 1.21 glml, the multilamellar liposomes and
small unilamellar vesicles that are produced during the process of sonication are
much heavier than microemulsions. Following sonication, the background
density was adjusted to 1.005 glml. Under the untracentrifugation condition, all
the particles can be separated from each other according to buoyant density,
with microemulsions floating at the top of the centrifuge tube, the background
buffer in the middle layer, and multilamellar liposomes and vesicles sinking to the
bottom of the centrifuge tube. The schematic representation of the separation is
shown in Figure 4.3. Therefore, adjusting the sonication buffer density to 1.21
glml is the critical step to have a good separation of homogeneous
microemulsions.
Preparation of LDL-sized microemulsions also requires other critical
steps. Lipids in the appropriate ratio were thoroughly mixed in solvent, and then
the solvent was completely removed prior to sonication. The temperature of the
mixture was maintained at 60•‹C throughout the sonication procedure. Failure to
maintain this temperature would result in the formation of undesired particles.
The volume of sonication buffer added to the sonication vessel should be in the
+-)Bigger-sized microemulsion
Desired-sized microemulsion
+-I Background buffer
-Vesicles or multilamellar liposomes
Figure 4.3 : Schematic representation of ultracentrifugation separation of microemulsion.
sonication. The distance between the microtip and the bottom of the sonication
vessel should be in the range of 0.8 - 1 cm. The sonication would be inefficient if
the microtip is too far away from the bottom of the vessel, while the sonication
vessel would break easily if the microtip is too close to the bottom of the vessel.
Sonication for 5.5 hours gave optimum yields of the desired particles.
In order to make sure that the sonication process did not damage
sphingomyelin or cholesteryl oleate, thin layer chromatography was used to
check the chemical compositions of lipids after the sonication. The experimental
results showed that no degradation products of sphingomyelin or cholesteryl
oleate could be detected. That the SPMICO microemulsion can be successfully
prepared by the above-mentioned protocol demonstrates that sonication is an
efficient method to prepare stable, homogeneous SPMICO microemulsions of
LDL size. The homogeneous size distribution was confirmed by quasi-elastic
light scattering.
Figure 4.4 shows the size distribution of SPMICO microemulsions
determined by the light scattering technique. Greater than 95% of the
microemulsions fell within the measured size range (23-27 nm). Occasionally
trace amounts of large microemulsions (about 80 nm) contaminated the sample
if ultracentrifugal separation was incomplete. It can easily be demonstrated that
these particles make no contribution to the NMR spectra within the spectral
window in these experiments. The big particles (80 nm) are about 3 times
bigger than the small ones (about 25 nm in diameter). If the linewidth
contribution from diffusion is neglected, and we consider only the isotropic
Mean Diameter of Major Particles (98 O/O) = 25 nm
Figure 4.4: The size distribution of SPMICO microemulsions measured by Nicomp 370 particle sizer.
rotation of the particle, from the Stokes-Einstein equation (equation 2.1 7 in
chapter 2), zt of the larger particles is calculated to be 33 times larger than that
of the small ones. Since linewidth varies directly with correlation time, the
linewidth of the big particle will be about 33 times larger than the small particles,
therefore, the linewidth of the analyzed big particles can only contribute to the
baseline of the NMR spectra.
The sphingomyelin/cholesteryl oleate microemulsions were studied by 31 P
NMR. Figure 4.5 shows representatives of the 3lP NMR spectra of SPMICO
microemulsions at different temperatures. The spectra at all the temperatures
can be represented by a single Lorentzian function as demonstrated by Figure
4.5. This suggests that all of the phospholipids are magnetically equivalent.
The temperature dependence of 31 P linewidth of SPMICO
microemulsions is shown in Table 4.2. Linewidths are calculated by the
computer simulation to the s1 P spectra to a single Lorentzian lineshape function.
By comparing Table 4.1 with Table 4.2, it is clear that the linewidth of the spectra
of sphingomyelin vesicles and SPM /CO microemulsions are similar within the
experimental error at temperatures below 40 "C. This indicates that the motions
of the headgroup at temperature below 40•‹C are similar. The gel to liquid
crystalline phase transition temperature of around 37 "C is in good agreement
with the data presented here. At temperatures above 40 "C, the SPMICO
microemulsions have slightly broader linewidths compared to SPM vesicles.
This suggests that the neutral core may alter the motion or orientation of the
headgroup in the microemulsion monolayer.
Figure 4.5: 3' P NMR spectra of sphingomyelin/cholesteryl oleate microemulsions at a). 15 "C, b). 25 "C, c). 35 "C, d). 45 "C The spectra were simulated by an iterative least-square fit of a Lorentzian lineshape function (solid line) to the spectral data
points (crosses). Spectral parameters for those cases are as following: pulse width = 6.5 ps (90" flip angle), sweep width = 25,000 Hz,
delay between pulses = 2.5 s, dataset = 2k zero filled to 4K, dwell time = 20 1s. line broadening = 5 Hz at 15 "C and 25 "C, line
broadening = 0 Hz at 35 "C and 45 "C.
Figure 4.5 (a): 31P NMR spectra of SPMICO microemulsion at 15 OC
Figure 4.5 (b): 31P NMR spectra of SPMICO microemulsion at 25 OC
5 1
Figure 4.5 (c): 3lP NMR spectra of SPM/CO microemulsion at 35 'C
Figure 4.5 (d): 3lP NMR spectra of SPM/CO microemulsion at 45 "C
Linewidth (Av 1/, ) (Hz) *
4.3 SPM/CO/apo-HDL3 Reconstituted Lipoprotein System (LDL size)
Recombinant lipoproteins were prepared by incubating the SPMICO
microemulsion with the purified apo-HDL3 above the order-disorder transition
temperature. Two steps were found to be critical for the successful preparation
of recombinant lipoprotein:
Exhaustive dialysis of the apo-HDL3 against dialysis buffer (1 00 mM KCI, 10
mM Tris-HCI, Na2C03, density= 1.005 glml, pH = 8.6) shoud be carried out in
order to get the same background buffer as microemulsions. After the
dialysis, the appearance of apo-HDL solution should be clear (no precipitate
appears in the solution).
The mixing procedure should be performed very gently.
In addition, the size distribution of the SPMICO microemulsion should be
very uniform and in the range of 23-26 nm. To date, we have reproducibly
prepared the recombinant lipoprotein (SPMICOlapo-HDL3 ) in low density
lipoprotein size. In order to study the effect of temperature on the structure of
reconstituted particles, the 3lP NMR spectra were collected over the
temperature range 10-50 "C. It is interesting to note that at lower temperature
when apo-HDL3 was added to the protein-free SPMICO microemulsions, the 31 P
linewidth of the reconstituted lipoprotein increased by approximately 20-40 Hz
compared to SPM vesicles and SPMICO microemulsions.
After the reconstituted lipoprotein was made, the size of the reconstituted
lipoprotein was checked immediately by quasi-elastic light scattering as shown in
Figure 4.6. The size was the same as that before addition within experimental
error. From this observation, we can conclud e that the linewidth increase after
apo-HDL, was incorporated was not due to any increase of particle size. One
possible explanation for this observation (the linewidth increase at lower
temperatures was not due to increase in particle size) is that bulky apo-HDL,
may influence the motion of sphingomyelin headgroup after the incorporation.
Another possible explanation is that the bulky apo-HDL3 changes the orientation
of phospholipid headgroups, thus resulting in the change in chemical shift
anisotropy of the phospholipid headgroup. From equations 2.1 5 and 2.19 in
Chapter 2, we can see that this could lead to a change in linewidth. The spectra
of reconstituted lipoproteins were presented in Figure 4.7 to Figure 4.9. Each
spectrum can be simulated by a superposition of two Lorentzian functions whose
linewidth at half-height (AVt/2) differ significantly.
However, we were able to represent the data by a single Lorentzian when
the temperature reaches above 35•‹C. Indeed, as demonstrated in Figure 4.10,
the spectrum recorded at 45 "C could be fitted to a single Lorentzian with very
good accuracy. It is possible that a superposition of Lorentzians also exists
above 35"C, but the linewidth of the respective spectral components is not
sufficiently different so that we can resolve them. Table 4.3 shows the linewidths
of 3' P NMR spectra SPMICOlapo-HDL3 reconstituted particles as a function of
the temperature. The linewidths (both broad and narrow components) were
obtained from simulation of the superposition of two Lorentzian functions fitting
to the experimental data points.
The ability to resolve two superimposed spectral components
unambiguously depends on the magnitude of the difference in linewidth between
Mean Diameter of Major Particles (97 %) = 24.0 nm
Figure 4.6 (a) Light scattering results from the SPMICO microemulsion samples before addition of apo-HDL3.
Mean Diameter of Major Particles (98 'lo) = 24.0 nm
Size (nrn)
Figure 4.6 (b) Light scattering results from the SPMICO microemulsion
samples after addition of apo-HDL3.
r
the broad and the narrow component. When the two Lorentzian functions differ
in width by a factor of three or more, they can be easily resolved. On the other
hand, if the difference in linewidth between the broad and narrow component is a
factor of two or less, they may be very difficult to resolve.
The possible explanation for this phenomenon (spectra were comprised of
two components) is that protein alters the motion or the orientation of the
headgroup of phospholipids differently in its different domains. It is possible that
the broad component of the spectra is due to the influence of the protein. We
can not rule out the possibility that the broad component is not influenced by the
protein, but that the narrow component is influenced. But no matter what is the
case, we did observe two different domains in the particle. As the temperature
increased, the rate of the exchange between the "protein - rich" domain and
"protein - deficient" domain increases, thus averaging the observed linewidth.
For instance, at temperatures above 35 "C, the motion of the molecule was so
fast that we were unable to detect a spectral difference between the two
components. A similar phenomenon was also observed for the spectrum at
45•‹C.
Phospholipid - protein interaction is one of the most important topics in the
study of the structure of lipoproteins and biological membranes. The interaction
between the two major components of the membrane (protein and lipid) is very
complex, and the influence of one component on the other cannot be expected
to be the same because the structures of the two components are greatly
different from each other. Therefore, an elucidation of the effects of each
component on the other is crucial to an understanding of the relationship
Figure 4.7: 31 P NMR spectra of sphingomyelin/cholesteryl oleatelapo-HDL3
particles at 1 5 "C. The spectra were simulated by an iterative least-squared fit to (a) a single Lorentzian lineshape function (four parameters) or (b) a superposition of two Lorentzian functions (seven parameters) of the spectral data points (crosses). The solid line represents the calculated fit.
Spectral parameters are as following: pulse width = 6.5 ps (90" flip angle), sweep width = 50,000 Hz,
delay between pulses = 2.5 s, dataset= 2k zero filled to 4K, dwell time = 10 ps, number of scans = 20,000, line broadening = 5 Hz.
Figure 4.7 (a): 3'P NMR spectrum of SPMICOlapo-HDL3 reconstituted particles at 15 "C (single Lorentzian).
Figure 4.7 (b): 31P NMR spectrum of SPMICOlapo-HDL3 reconstituted
particles at 15 "C (sum of two Lorentzians).
Figure 4.8: 31 P NMR spectra of sphingomyelin/cholesteryl oleatelapo-HDL3 reconstituted lipoprotein particles at 20 OC. The spectra were simulated by an iterative least-squared fit to (a) a single Lorentzian lineshape function (four parameters) or (b) a superposition of two Lorentzian functions (seven parameters)
of the spectral data points (crosses). The solid line represents the calculated fit. Spectral parameters are as following: pulse width = 6.5 ps (90" flip angle), sweep width = 50,000 Hz,
delay between pulses = 2.5 s, dataset = 2k zero filled to 4K, dwell time = 10 ps, number of scans = 20,000, line broadening = 5 Hz.
Figure 4.8 (a): 31 P NMR spectrum of SPMICOlapo-HDL3 reconstituted
particles at 20 "C (single Lorentzian).
Figure 4.8 (b): 31P NMR spectrum of SPMICOlapo-HDL3 reconstituted particles at 20 "C (sum of two Lorentzians).
Figure 4.9: 3lP NMR spectra of sphingomyelin I cholesteryl oleate / apo-HDL3 reconstituted lipoprotein particles at 25 "C. The spectra were simulated by an iterative least-squared fit to (a) a single Lorentzian lineshape function (four parameters) or
(b) a superposition of two Lorentzian functions (seven parameters) of the spectral data points (crosses). The solid line represents the
calculated fit. Spectral parameters are as following: pulse width = 6.5 ps ( 90•‹ flip angle ), sweep width = 50,000 Hz,
delay between pulses = 2.5 s, dataset = 2k zero filled to 4K, dwell time = 10 ps, number of scans = 20,000, line broadening = 5 Hz.
Figure 4.9 (a): 31 P NMR spectrum of SPMICOlapo-HDL3 reconstituted particles at 25 "C (single Lorentzian).
Figure 4.9 (b): 31P NMR spectrum of SPMICOlapo-HDL3 reconstituted
particles at 25 "C (sum of two Lorentzians).
Figure 4.10: 31 P NMR spectra of sphingomyelin/cholesteryl oleatelapo-HDL3 reconstituted lipoprotein particles at 30 "C. The spectra were simulated by an iterative least-squared fit to (a) a single Lorentzian lineshape function (four parameters) or (b) a superposition of two Lorentzian functions (seven parameters) of the spectral data points (crosses). The solid line represents the
calculated fit.
Spectral parameters are as following: pulse width = 6.5 ps (90" flip angle), sweep width = 50,000 Hz,
delay between pulses = 2.5 s, dataset = 2k zero filled to 4K, dwell time = 10 ps, number of scans = 20,000, line broadening = 5 Hz.
Figure 4.10 (a): 31P NMR spectrum of SPMICOlapo-HDL3 reconstituted
particles at 30•‹C (single Lorentzian).
Figure 4.10 (b): 31P NMR spectrum of SPMICOlapo-HDL3 reconstituted
particles at 30 "C (sum of two Lorentzians).
Figure 4.1 1 : 3lP NMR spectra of sphingomyelin/cholesteryl oleatelapo-HDL3
reconstituted lipoprotein particles at 45 "C.
The spectra were simulated by an iterative least-squared fit to (a) a single Lorentzian lineshape function (four parameters) or (b) a superposition of two Lorentzian functions (seven parameters) of the spectral data points (crosses). The solid line represents the calculated fit. Spectral parameters are as following: pulse width = 6.5 ps (90" flip angle), sweep width = 50,000 Hz,
delay between pulses = 2.5 s, dataset = 2k zero filled to 4K, dwell time = 10 ps, number of scans = 20,000, line broadening = 5 Hz.
Figure 4.11 (a): 3 1 ~ NMR spectrum of SPMICOlapo-HDL3 reconstituted particles at 45 "C (single Lorentzian).
Figure 4.11 (b): 31P NMR spectrum of SPMICOlapo-HDL3 reconstituted
particles at 45 "C (sum of two Lorentzians).
Table 4.3 : 3' P NMR linewidths of sphingomyelinlcholesteryl oleatelapo-
HDLR recombinant lipoproteins at selective temperature (about
Temperature
("C)
Broad component
(Hz)
narrow component
(Hz)
% broad
component
* Uncertainties of the linewidths are 10 %.
between the major components of biomembranes or lipoproteins.
31P NMR is one of the most powerful tools for exploring lipid-protein
interactions. This technique is only sensitive to the behavior of the phospholipid
headgroups, and it is the phospholipid headgroup with its ionic charges that is
best suited for interacting with the surface of the protein that also contains ionic
charges. Significant attractive force could be imagined between phospholipid
headgroups and protein. In contrast, for hydrocarbon chains, the only possible
interaction with protein are between two hydrophobic surfaces in a hydrophobic
medium which are governed by hydrophobic forces and weak attractive (van der
Waals) forces.
Another distinct advantage of 31P NMR is that it is non-perturbing and
does not require labeling. This is important since changes in the structure of
phospholipid headgroups by introduction of labels may significantly alter the
behavior of the labeled phospholipid (Yeagle, 1990).
4.4 A Lateral Diffusion Study of SPMICO Microemulsions
The representative 31P NMR spectra of sphingomyelin/cholesteryl oleate
microemulsions at several different viscosities at 25 "C are shown in Figure 4.1 2.
The linewidths, given in Table 4.4, are obtained from computer fits of the spectra
to a single Lorentzian functions.
A plot of (Av - C) versus q is shown in Figure 4.13. At low viscosity or
intermediate viscosity, the linewidths increase with the increases in viscosity.
Figure 4.12: 3lP NMR spectra of sphingomyelinlcholesteryl oleate microemulsions at several different viscosities at 25 "C. The spectra were simulated by an iterative least-squared fit to a single Lorentzian lineshape function (four parameters) of the spectral data points (crosses). The solid line represents the calculated fit. Spectral parameters are as following: pulse width = 6.5 ps (90" flip angle), sweep width =
50,000 Hz, delay between pulses = 1.5 s, dataset= 2k zero filled to 4K, dwell time = 10 ps, number of scans 2 30,000.
Figure 4.12(a): The 31P NMR spectrum of SPMICO microemulsion at a
viscosity of 0.98 centipoise.
Figure 4.12(b): The 3'P NMR spectrum of SPMICO microemulsion at a viscosity of 1.29 centipoise.
Figure 4.12(c): The 3 l P NMR spectrum of SPMICO microemulsion at a viscosity of 2.16 centipoise.
Figure 4.12(d): The 3 1 ~ NMR spectrum of SPMICO microemulsion at a
viscosity of 4.45 centipoise.
Figure 4.1 2(e): The 31 P NMR spectrum of SPMICO microemulsion at a viscosity of 6.01 centipoise.
Figure 4.1 2(f): The 3lP NMR spectrum of SPMICO microemulsion at a viscosity of 8.46 centipoise.
- - - -
Table 4.4: The viscosity dependence of linewidths of sphingomyelinl
cholesteryl oleate microemulsions
However, at high viscosity, the increase in the linewidth gradually levels off. This
is important since it indicates that as q increases beyond r 35 centipoise,
particle tumbling (rotation) does not significantly contribute to the linewidth.
A plot of (Av 1 ~ 2 - C) versus q-1 is shown in Figure 4.14. The solid line
represents the least square fit of the reciprocal of linewidth to the experimental
data points. In order to obtain D (lateral diffusion coefficient), the radius of the
SPMICO microemulsion particles must be known. The radius of the particle was
determined by quasi-elastic light scattering to be about 12.5 nm.
We were specifically interested in diffusion of sphingomyelin in the
microemulsions. From equation 2.21, the lateral diffusion coefficient (D) of
sphingomyelin in SPMICO microemulsions was determined to be 1.0k0.3 x 10-9
(cm 2 s-1). A comparison of diffusion coefficients for serum lipoproteins,
microemulsions, reconstituted lipoproteins and phospholipid bilayers is shown in
Table 4.5. The lateral diffusion coefficient of SPMICO microemulsions is
approximately 1.4 times smaller than that of phospholipids in low density
lipoproteins (the value for LDL at 25 "C is 1.4 + 0.5 x 10-9 cm s-1). The reason
to compare this to that of LDL is that low density lipoproteins are cholesteryl
ester rich lipoproteins. The microemulsions were prepared using cholesteryl
oleate as a neutral core, and the size of microemulsions were about the same as
that of low density lipoproteins.
The core of SPMICO microemulsions is composed of only cholesteryl
esters, which undergo a thermal liquid-crystalline to liquid phase transition over
b the range of 20 - 40 "C (Deckelbaum et al., 1977). At 25 "C, the core lipid of
I I 1 I I
1500- -
1000- -
-
0 10 X) 30 40 60
q ( centipoise )
Figure 4.13: The linewidths of 3' P NMR spectra of SPMICO microemulsions versus viscosity.
Figure 4.14: Plot of (Av - C) -1 versus q - 1 of sphingomyelin
/cholesteryl oleate microemulsions.
microemulsions is just into the phase transition. If the core-monolayer
interactions are co-operative in the microemulsions, then the solid like
cholesteryl oleate may be responsible for the slower diffusion of the
phospholipids. Another possible explanation is that the phenomenon could
happen through interdigitation of the acyl chain of sphingomyelin with cholesteryl
oleate.
SPMICO microemulsions appear to behave just like LDL. The diffusion
coefficient constant for low density lipoproteins (core lipids are mainly cholesteryl
ester) at 45•‹C is 10 times greater than that at 25 "C (Cushley et al., 1987;
Fenske et al., 1 990). Whereas in VLDL (core lipids are triglyceride-rich
lipoproteins; triglyceride over the range of 25 to 45 "C is a liquid-like phase), D
remains essentially constant with temperature going from 25 "C to 45 "C as
shown in Table 4.5 (Cushley et al., 1987; Chana et al., 1990).
4.5 A Field Dependence Study of SPMICO Microemulsion
The field dependence of a SPMICO microemulsion was conducted at four
different frequencies (four magnetic fields), specifically:
40 MHz (Bruker SY-100)
102 MHz (our "home built" Machine )
160.2 MHz ( Bruker AMX 400)
202.46 MHz ( Bruker AMX 500 )
The 31P NMR linewidths measured at different frequencies are given in Table
4.6. The linewidth was broader at higher field strength. From equation 2.22, we
can obtain the chemical shift anisotropy.
Table 4.5: Comparison of lateral diffusion coefficients for serum
The chemical shift anisotropy (Ao) calculated from the slope in Figure
4.1 5 (where Te is defined in equation 2.1 6) is 33 ppm. Table 4.7 shows the
comparison of the chemical shift anisotropy of different native lipoproteins as
well as reconstituted biomembrane. From Table 4.7, we can see that the
chemical shift anisotropy (Ao) value for SPMICO microemulsions is very close to
the value for egg PCfrO microemulsions at 25•‹C (or the value for egg PC
vesicles at 50•‹C). This suggests that the headgroup orientations for these three
systems are similar.
4.6 A Lateral Diffusion Study of SPMICOlapo-HDL3 Reconstituted
Lipoproteins
Representative 3lP NMR spectra of sphingomyelin/cholesteryI oleatelapo-
HDL3 reconstituted lipoproteins at different viscosities at 25 "C are shown in
Figure 4.1 6 to Figure 4.21. During the course of fitting the 31P linewidths for the
reconstituted lipoproteins, we found that at higher viscosity we were unable to
represent our data by a single Lorentzian function. A superposition of two
Lorentzians was necessary to get a reasonable fit. At lower viscosities,
especially with the sample without glycerol or with small amounts of glycerol (low
viscosity), the experimental data fit a single Lorentzian function. The two
spectral domains at lower viscosities still exist but we cannot resolve two
components at 25 "C in these spectra. This is most likely due to the fact that
the linewidths of these two spectral components are not sufficiently different at
25•‹C. This argument is supported by observations in our previous temperature
study in this thesis. At temperature below 25 "C (without glycerol present) the
data could not be represented as a single Lorentzian function. However, as
Table 4.6: Variation of linewidths of 3IP NMR spectra of SPMICO microemulsions at different NMR field
strengths.
vo2 ( Hz2 )
1.643 x 10
10.44 x 10 '5
26.24 x 10 '5
40.99 x 10 '5
Linewidth (Av 1,2 ) (HZ)
2 5 ( 6 )
4 9 ( 8 )
8 9 ( 1 2 )
1 2 8 ( 1 6 )
Table 4.7: Comparison of chemical shift anisotropy for serum
lipoproteins, microemulsions, and phospholipid bilayers.
VLDL
VLDL
LDL
LDL
HDL2
H DL3
egg P c r r o
SPMJCO microemulsions
Egg PC vesicles
Egg PC vesicles
DPPC liposomes
DPPC liposomes
a Chana, et al., (1 990);
c Parmar et al., ( 1 985 );
* This study
Viscosity
b Fenske et al. (1 990)
d Mclaughlin et. al. (1 975).
Figure 4.15: (Av - C) as a function of vO2 for SPMICO microemulsion at
25•‹C. The straight line is a least square fit to the data points.
Figure 4.16: 31P NMR spectra of sphingomyelin/cholesteryl oleatelapo-HDL3
reconstituted lipoprotein particles at 25 "C (viscosity = 0.99
centipoise). The spectra were simulated by an iterative least-squared fit to
(a) a single Lorentzian lineshape function (four parameters) or (b) a superposition of two Lorentzian functions (seven parameters) of the spectral data points. The solid line represents the
calculated fit. Spectral parameters are as following: pulse width = 6.5 ps (90' flip angle), sweep width =
50,000 Hz, delay between pulses = 1.5 s, dataset = 2k zero filled to 4K, dwell time = 10 ps, number of scans 2 50,000.
Figure 4.16 (a): SPMICOlapo-HDL3 reconstituted lipoproteins at viscosity 0.99 centipoise (single Lorentzian), no LB.
Figure 4.1 6 (b): SPMICOlapo-HDL3 reconstituted lipoproteins at viscosity 0.99 centipoise (sum of two Lorentzians), no LB.
Figure 4.1 7 : 3' P NMR spectra of sphingomyelin/cholesteryl oleatelapo-HDL3
reconstituted lipoprotein particles at 25 "C (viscosity = 1.53
centipoise). The spectra were simulated by an iterative least-squared fit to (a) a single Lorenttian lineshape function (four parameters) or (b) a superposition of two Lorentzian functions (seven parameters) of the spectral data points. The solid line represents the
calculated fit. Spectral parameters are as following: pulse width = 6.5 ps (90' flip angle), sweep width =
50,000 Hz, delay between pulses = 1.5 s, dataset = 2k zero filled to 4K, dwell time = 10 ps, number of scans 2 50,000.
Figure 4.17 (a): SPMICOlapo-HDL3 reconstituted lipoproteins at viscosity 1.53 centipoise (single Lorentzian), no LB.
Figure 4.17 (b): SPMICOlapo-HDL3 reconstituted lipoproteins at viscosity 1.53 centipoise (sum of two Lorentzians), no LB.
Figure 4.18: 31P NMR spectra of sphingomyelin/cholesteryl oleate/apo-HDL3 reconstituted lipoprotein particles at 25 "C (viscosity = 2.71
centipoise). The spectra were simulated by an iterative least-squared fit to (a) a single Lorentzian lineshape function (four parameters) or
(b) a superposition of two Lorentzian functions (seven parameters) of the spectral data points. The solid line represents the calculated fit. Spectral parameters are as following: pulse width = 6.5 ps (90" flip angle), sweep width =
50,000 Hz, delay between pulses = 1.5 s, dataset = 2k zero filled to 4K, dwell time = 10 ps, number of scans 2 50,000.
Figure 4.18 (b): SPMICOlapo-HDL3 reconstituted lipoproteins at viscosity
2.71 centipoise (sum of two Lorentzians) (LB= 5 Hz).
Figure 4.1 9: 3' P NMR spectra of sphingomyelin/cholesteryl oleatelapo-HDL3 reconstituted lipoprotein particles at 25 "C (viscosity = 5.30 centipoise). The spectra were simulated by an iterative least-squared fit to (a) a single Lorenttian lineshape function (four parameters) or (b) a superposition of two Lorentzian functions (seven parameters)
of the spectral data points. The solid line represents the calculated fit. Spectral parameters are as following: pulse width = 6.5 ps (90" flip angle), sweep width = 50,000 Hz,
delay between pulses = 1.5 s, dataset = 2k zero filled to 4K, dwell time = 10 p , number of scans 2 50,000.
Figure 4.1 9 (b): SPMICOlapo-HDL3 reconstituted lipoproteins at viscosity 5.3 centipoise (sum of two Lorentzians) (LEI= 10 Hz).
Figure 4.20: 3lP NMR spectra of sphingomyelin/cholesteryl oleatelapo-HDL3 reconstituted lipoprotein particles at 25 "C (viscosity = 9.89 centipoise). The spectra were simulated by an iterative least-squared fit to (a) a single Lorentzian lineshape function (four parameters) or
(b) a superposition of two Lorentzian functions (seven parameters) of the spectral data points. The solid line represents the calculated fit. Spectral parameters are as following: pulse width = 6.5 pi (90" flip angle), sweep width = 50,000 Hz,
delay between pulses = 1.5 s, dataset = 2k zero filled to 4K, dwell time = 10 ps, number of scans 2 50,000.
Figure 4.20 (a): SPMICOlapo-HDL3 reconstituted lipoproteins at viscosity
9.89 centipoise (single Lorentzian) (LB = 25 Hz).
Figure 4.20 (b): SPMICOlapo-HDL3 reconstituted lipoproteins at viscosity
9.89 centipoise (sum of two Lorentzians) (LB= 25 Hz).
Figure 4.21 : 31 P NMR spectra of sphingomyelin/cholesteryl oleate/apo-HDL3 reconstituted lipoprotein particles at 25 "C (viscosity = 17.24
centipoise). The spectra were simulated by an iterative least-squared fit to (a) a single Lorentzian lineshape function (four parameters) or
(b) a superposition of two Lorentzian functions (seven parameters) of the spectral data points. The solid line represents the
calculated fit. Spectral parameters are as following: pulse width = 6.5 ps (90' flip angle), sweep width = 50,000 Hz,
delay between pulses = 1.5 s, dataset = 2k zero filled to 4K, dwell time = 10 ps, number of scans 2 50,000.
Figure 4.21 (a): SPMICOlapo-HDL3 reconstituted lipoproteins at viscosity
Fielding, C. J. (1 978) In Disturbances in Lipid and Lipoprotein Metabolism (Dietschy, J. M., Gotto, A. M., Jr., and Ontko, J. A., eds.), Amer. Physiol. Soc., Bethesda, MD. pp. 83-98.
Forte, G. M., Nichols, A. V. and Glaeser, R. M. (1968) Chem. Phys. Lipids 2:
396-398.
Gennis, R.B. (1 989) Biomembranes. Springer-Verlag, New York.
Ginsburg, G.S., Small, D.M. and Atkinson, D. (1982) J. Biol. Chem. 257: 8216 - 8227.
Ginsburg, G. S., Walsh, M. T., Small, D. M. and Atkinson, D. (1984) J. Biol.
Chem. 259: 6667-6673.
Gotto, A.M., Jr. (1 987) Plasma lipoproteins. Elsevier, New York, NY.
Hamilton J. A. and Morrisett, J. D. (1 986) Methods Enzymol. 128: 472-51 5.
Hamosh, M. and Hamosh, P. (1 983) Mol. Aspects Med. 6: 199-289.
Hara, H. and Yokoyama, S. (1 991) J. Biol. Chem. 266: 3080-3086.
Hara, H. and Yokoyama, S. (1 992) Biochemistry 31 : 2040-2046.
Harris, R.K. (1 983) Nuclear Magnetic Resonance Spectroscopy: A Physicochemical View. Pitman Publishing Inc., Marshfield, MA.
Hauser, H., Radloff, R. R., Sundell, S. and Pascher, 1. (1988) J. Am. Chem. Soc,
Reisinger, R. E. and Atkinson, D. (1990) J. Lipid Res. 31 : 849 - 858.
Reichl, D., and Sterchi, J. M. (1 992) Biochim. Biophys. Acta 11 27: 28-32.
Scanu, A.M. (1 979) In The Biochemistry of Atherosclerosis (Scanu, A.M., Wissler, R.W. and Getz, G.S., eds.). Marcel Dekker, New York, pp. 3-8.
Schmitz, G., Beuck, M., Fisher, H., Nowicka, G. and Robenek, H. (1990) J. Lipid
Res. 31 : 1741 -1 752.
Seelig, J. (1 978) Biochim. Biophys. Acta 515: 105-140.
Seelig, J. and Seelig, A. (1 980) Quat. Rev. Biophys. 13: 19-61
Seelig, J. (1 982) In Physical Methods on Biological Membranes and Their Model Systems (Conti, F., Blumberg, W.E., de Gier, J., Pochiari, F. eds.). Plenum Press, New York. pp. 27-37.
Singer, S.J. and Nicolson G.L. (1 972) Science 175: 720-731
Small, D.M. (1981) in Membrane Molecules, Toxin and Cells (Block,K.; Bolis, L.
and Tosteson, D.C. eds.) John Weight, PGS. Inc, pp. 11 -34.
Stein, O., Oette, K., Haratz., Halperin, G. and Stein, Y. (1988) Biochim. Biophys. Acta 960: 322-333.
Stein, O., Dabach, Y., Hollander, G., Halperin, G., Oette, K. and Stein, Y (1 989a) Biochim. Biophys. Acta 1006: 144-1 46.
Stein, O., Oette, K., Hollander, G., Dabach, Y., Ben-Naim, Y. and Stein, Y. (1 989b) Biochim. Biophys. Acta 1003:175-82.
Stryer, L. (1988) Biochemistry, Third edition, W. H. Freeman and Company, New York, NY.
Swaney, J. B. (1983) J. Biol. Chem. 258: 1254-1 259.
Swaney, J. B. and Chang, B. C. (1 980) Biochemistry 19: 5637-5644.
Treleaven, W.D., Cushley, R.J. and Zhao, L. (1 991 ) Biophys. J. 59: 395.
Treleaven, W.D., Parmar, Y. I., Gorrissen, H. and Cushley, R. J. (1 986)
Biochim. Biophys. Acta. 877: 198-21 0.
Vauhkonen, M., Somerharju, P. (1 990). Chem. Phys. Lipids 52: 207-21 6.