using vesicles to study the effect of sterols on the … · 2007-03-30 · using vesicles to study...

USING VESICLES TO STUDY THE EFFECT OF

STEROLS ON THE MECHANICAL STRENGTH OF

LIPID MEMBRANES AND THE PROTEIN-MEMBRANE

LIPID INTERACTIONS

by

Philipus Josepus Patty

B.Sc., Institute of Teacher Training and Education of Ujung Pandang Indonesia, 1991.MSc., Simon Fraser University, 2000.

a thesis submitted in partial fulfillment

of the requirements for the degree of

Doctor of Philosophy

in the Departmentof

Physics

c© Philipus Josepus Patty 2006SIMON FRASER UNIVERSITY

2006

All rights reserved. This work may not be

reproduced in whole or in part, by photocopy

or other means, without the permission of the author.

APPROVAL

Name: Philipus Josepus Patty

Degree: Doctor of Philosophy

Title of thesis: Using Vesicles to Study the Effect of Sterols on the Mechani-cal Strength of Lipid Membranes and the Protein-MembraneLipid Interactions

Examining Committee: Dr. Patricia Mooney, ProfessorChair

Dr. Barbara Frisken, ProfessorSenior Supervisor

Dr. Michael Wortis, ProfessorDepartment of Physics

Dr. Jenifer Thewalt, Associate ProfessorDepartment of PhysicsDepartment of Molecular Biology and Biochemistry

Dr. Martin Zuckermann, Adjunct ProfessorDepartment of PhysicsInternal Examiner

Dr. Bela Joos, ProfessorOttawa Charleton Institute for PhysicsUniversity of OttawaExternal Examiner

Date Approved:

ii

Abstract

In this thesis, studies of two biophysical topics will be discussed: the effect of sterols

on the mechanical strength of lipid membranes, and the interaction between Cytidine 5’-

triphosphate(CTP):phosphocholine cytidylyltransferase (CCT) and membrane lipids.

The mechanical strength of lipid membranes was probed by measuring the lysis tension

of vesicles, as determined from the minimum pressure required to extrude vesicles through

small pores. The vesicles used in these experiments were made from mixtures of 1-palmitoyl-

2-oleoyl-sn-glycero-3-phosphatidylcholine (POPC) and various sterols, including cholesterol,

lanosterol and ergosterol. The effect of the sterol concentration on the lysis tension was

determined. The results show that all sterols increase the lysis tension of POPC membranes,

where cholesterol shows the largest effect followed by lanosterol and ergosterol. The increase

in the lysis tension of POPC membranes by sterols is correlated to the increase in the chain

order of the lipids by sterols. The increase in the strength of lipid membranes by sterols

indicates their contribution to cell viability, which depends on maintaining an intact plasma

membrane.

The interaction between CCT and membrane lipids was studied by observing vesicle

aggregation induced by CCT. This was conducted by measuring the size and polydispersity

of vesicles before and after the introduction of CCT using dynamic light scattering. Vesicles

for these studies were made both from lipids that activate CCT (activating lipids) and lipids

that do not activate CCT (non-activating lipids). The activating lipids investigated include

anionic lipids (class I lipids) and lipids that produce negative curvature in membrane (class

II lipids). Aggregation occurs when CCT is introduced to samples of class I lipid vesicles.

In contrast, there is no indication of aggregation when CCT is introduced to samples of

vesicles made from both non-activating and class II lipids. The occurrence of aggregation

depends on the binding strength of CCT to the membrane. The results suggest that CCT

cross-bridges two vesicles, which is a new aspect of CCT interaction with lipid membranes.

iii

To : my wife, Cynthia, and our precious daughter, Willingda

iv

Acknowledgments

I would like to thank Dr. Barbara Frisken for her support and guidance and for giving me

opportunities to study many things in Physics. I would like to thank Dr. Michael Wortis

and Dr. Jenifer Thewalt as my committee members for their input and time. I would also

like to thank Dr. Rosemary Cornell and Dr. Svetla Taneva for the collaboration in CCT

protein-membrane interaction studies as well as Dr. Martin Suckermann, Dr. Ya-Wei and

Dr. Jenifer Thewalt for the collaboration in sterols-membrane studies. I am grateful for

Dr. li Yang for helpful discussions in SEM measurements. I really appreciate the help I

received from all my friends who share the lab. with me : Dr. Art. Bailey, Dr. Jun Gao,

Dr. Laurent Rubatat, Dave Lee, Yong Sun and other friends in the Physics Department for

their help and friendship. I am really thankful to The Rector of Pattimura University for

allowing me to pursue this degree. I would also like to thank my family, the Patty’s and

the Wattimena’s for their support. Finally I am really grateful to my wife, Cynthia and our

daughter, Willingda, for their loves, supports and prayers.

v

Contents

Approval ii

Abstract iii

Acknowledgments v

Contents vi

List of Figures ix

1 Introduction 1

2 Scattering Angle Dependence of the Size of Vesicles Measured by Dynamic

Light Scattering 4

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2.2 Radius distribution by DLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2.1 Types of Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2.2 Intensity-weighted Distributions . . . . . . . . . . . . . . . . . . . . . 8

2.2.3 Number-weighted Distribution . . . . . . . . . . . . . . . . . . . . . . 11

2.2.4 Relation between Gi(R) and Gn(R) . . . . . . . . . . . . . . . . . . . 12

2.2.5 Relation between Rh and the number-weighted distribution . . . . . . 12

2.3 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.3.1 Extruded Vesicle Preparation . . . . . . . . . . . . . . . . . . . . . . . 13

2.3.2 DLS Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.4 Results and Discussions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

vi

3 Vesicle Size and Lamellarity 27

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27


3.2.1 PCTE Pore Size Measurements . . . . . . . . . . . . . . . . . . . . . . 33

3.2.2 Vesicle Size Measurements . . . . . . . . . . . . . . . . . . . . . . . . . 34

3.2.3 Vesicle Lamellarity Determination . . . . . . . . . . . . . . . . . . . . 34


3.3.1 Pore Size Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

3.3.2 Vesicle Size Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3.3.3 Vesicle Lamellarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

4 Effect of Sterols on Lipid Membranes 54

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54


4.2.1 Vesicle Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

4.2.2 Lysis Tension Determination . . . . . . . . . . . . . . . . . . . . . . . 67


4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

5 Vesicles Aggregation Induced by CCT Protein 81

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81


5.2.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

5.2.2 Vesicle Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

5.2.3 Vesicle Size Measurements . . . . . . . . . . . . . . . . . . . . . . . . . 87

5.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

5.3.1 Effect of CCT on Activating and Non-Activating Lipid Vesicles . . . . 89

5.3.2 Effect of CCT on Class I and Class II Lipid Vesicles . . . . . . . . . . 98

5.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

5.4.1 CCT-Membrane Binding Strength and Vesicle Aggregation . . . . . . 100

5.4.2 Cross-Bridging Mechanism for Aggregation . . . . . . . . . . . . . . . 101

5.4.3 Electrostatic Repulsion of Charged Membranes versus Aggregation . . 103

5.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

vii

6 Conclusion 108

viii

List of Figures

2.1 Gaussian and Schulz Probability distribution function . . . . . . . . . . . . . 9

2.2 Rh and σRhas a function of q . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2.3 Ri and σRiof 50 nm vesicles as a function of q . . . . . . . . . . . . . . . . . 17

2.4 Ri and σRiof 100 nm vesicles as a function of q . . . . . . . . . . . . . . . . . 18

2.5 Ratios of Rh to Ri and σRhto σRi

for vesicles with Ri of 60 and 90 nm as a

function of σRi. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.6 Rn and σRnof 50 nm vesicles as a function of q . . . . . . . . . . . . . . . . . 22

2.7 Rn and σRnof 100 nm vesicles as a function of q . . . . . . . . . . . . . . . . 23

2.8 Calculated Rn and σRnas a function of σi . . . . . . . . . . . . . . . . . . . . 24

3.1 Radius and polydispersity of extruded vesicles as a function of nominal pore

radius . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

3.2 Emission and excitation scans for POPC:NBD-PC vesicles . . . . . . . . . . . 36

3.3 Fraction of inner lipids for unilamellar vesicles as a function of vesicle radius . 39

3.4 Scanning electron micrograph of surface of PCTE membranes . . . . . . . . . 40

3.5 Histogram of pore radius data for a) 25 nm, b) 50 nm, c) 100 nm and d)

200 nm pores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

3.6 The mean radii and the polydispersities of the extruded vesicles and the pores

as a function of nominal pore radius . . . . . . . . . . . . . . . . . . . . . . . 44

3.7 Comparison of pores and vesicles radius distribution . . . . . . . . . . . . . . 45

3.8 Radius distribution of vesicles produced by extruding sample of lipid disper-

sion through 100 nm pores with and without a freeze-thaw-vortex procedure . 48

3.9 Intensiy of POPC:NBD-PC and fraction of inner lipids . . . . . . . . . . . . . 50

3.10 Fraction of inner lipids as a function of number of passes for vesicles extruded

through 50 nm and 100 nm pores . . . . . . . . . . . . . . . . . . . . . . . . . 51

ix

4.1 Shapes of lipid molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

4.2 The molecular structure of POPC . . . . . . . . . . . . . . . . . . . . . . . . 58

4.3 The molecular structure of POPC . . . . . . . . . . . . . . . . . . . . . . . . 59

4.4 Membrane shape changes as a response to mechanical forces . . . . . . . . . . 62

4.5 The molecular structure of cholesterol, lanosterol and ergosterol . . . . . . . . 63

4.6 Schematic diagram of a vesicle at the entrance of a pore . . . . . . . . . . . . 68

4.7 Histogram of pore radius data for PCTE with nominal pore radius of 50 nm . 71

4.8 Viscosity corrected flowrate as a function of pressure . . . . . . . . . . . . . . 72

4.9 The lysis tension of POPC vesicles as a function of sterol concentration . . . 74

4.10 The lysis tension of membrane as a function of M1, κ and Ka . . . . . . . . . 75

4.11 The lysis tension of SOPC membrane as a function of area expansion modulus 78

5.1 Two modes of CCT binding to membranes . . . . . . . . . . . . . . . . . . . . 85

5.2 Intensity-intensity correlation function data from DLS measurements of vesi-

cles made from a) Egg PC and b) Egg PG before and after introduction of

CCT protein. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

5.3 The mean radius and the polydispersity of vesicles made from Egg PC and

Egg PG. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

5.4 A typical TEM micrographs of pure Egg PG vesicles and Egg PG vesicles

plus CCT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

5.5 Histogram of vesicle radius data for pure Egg PG and Egg PG plus CCT. . . 94

5.6 The mean radius and the polydispersity of Egg PC:Egg PG mixture. . . . . . 96

5.7 The mean radius and the polydispersity of Egg PG vesicles before and after

introduction of CCT or CCT-236 or peptide . . . . . . . . . . . . . . . . . . . 97

5.8 The mean radius and the polydispersity of vesicles made from Egg PG and

POPS (class I CCT activator), Egg PC:DAG (85:15) and Egg PC:DOPE

(40:60) (class II CCT activator), Egg PC:OA (a special CCT activator that

features both class I and II), as well as Egg PC (a weak CCT activator). . . . 99

5.9 Electrostatic and van der Waals energy between vesicles as a function of

surface separation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

x

Chapter 1

Introduction

This thesis contains studies of the effect of sterols on the mechanical strength of lipid

membranes and the interaction between Cytidine 5’-triphosphate (CTP):phosphocholine

cytidylyltransferase (CCT) and membrane lipids. It consists of six chapters, where four

chapters discuss four different experiments. Chapter 2 discusses the angle dependence of

the size of the vesicles measured by DLS. The detail studies on vesicles size measured by

DLS is conducted since DLS is used in experiments discussed in Chapters 3 to 5. Chapter 3

discusses the effect of the size of the pores in polycarbonate track-etched membrane (PCTE)

on the size and lamellarity of vesicles produced by extrusion. In chapter 4, experiments

on the effect of sterols including cholesterol, lanosterol and ergosterol on the mechanical

strength of lipid membranes is described. Chapter 5 discusses the interaction between CCT

and membrane lipids which is studied by observing vesicle aggregation induced by CCT.

1

CHAPTER 1. INTRODUCTION 2

In Chapter 2, the intensity weighted and number weighted mean radii and polydisper-

sities of the vesicles as a function of scattering angle are compared. Two different sizes of

vesicles were used in these studies in order to show the effect of vesicle size on the results.

Two functional form of the distribution were investigated, Gaussian and Schulz, in order

to show the effect of symmetry of the distribution on the results. There is a scattering an-

gle dependence of the intensity-weighted mean radius when both the Schulz and Gaussian

distributions are used. The scattering angle dependence of number weighted mean radius

was found to be insignificant when a Schulz distribution was used, but only acceptable for

small vesicles when the Gaussian distribution was used. The results also suggest that an

asymmetric distribution is more appropriate to represent the size distribution of extruded

vesicles.

In Chapter 3, the size and lamellarity of extruded vesicles as a function of size of the

pores are discussed. The size and lamellarity of vesicles were characterized using DLS and

fluorescence quenching assay, respectively, while the size of the pores was characterized

using Scanning Electron Microscopy (SEM). The size and the degree of multilamellarity of

extruded vesicle increase as the size of the pores increases. The polydispersity of the vesicles

produced using large pores is large, although the polydispersity of the pores is small. The

polydispersity of the vesicles is most likely due to the presence of small vesicles that are

produced spontaneously when the lipid was dispersed in water, or a consequence of the

process of extrusion.

CHAPTER 1. INTRODUCTION 3

In Chapter 4, the mechanical strength of lipid membranes as a function of sterol con-

centration is discussed. The membrane strength was probed by measuring the lysis ten-

sion of vesicles made from mixtures of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylcholine

(POPC) and various sterols. The lysis tension was determined from the minimum pressure

required to extrude vesicles through small pores. All sterols increase the lysis tension of

POPC membranes, where cholesterol shows the largest effect followed by lanosterol and

ergosterol.

In Chapter 5, the interaction between CCT and membrane lipids are discussed. The

CCT-membrane lipid interaction was studied by observing vesicles aggregation induced by

CCT. DLS was used to explore the aggregation of vesicles made both from lipids that ac-

tivate CCT (activating lipids) and lipids that do not activate CCT (non-activating lipids).

The activating lipids investigated include anionic lipids (class I lipids) and lipids that pro-

duce negative curvature in membrane (class II lipids). Aggregation occurs when CCT is

introduced to samples of class I lipid vesicles. In contrast, there is no indication of aggre-

gation when CCT is introduced to samples of vesicles made from both non-activating and

class II lipids.

Chapter 2

Scattering Angle Dependence of

the Size of Vesicles Measured by

Dynamic Light Scattering

2.1 Introduction

Vesicles produced by extrusion have been used widely as model biological membranes and

as carriers in drug delivery technology [1]. Extruded vesicles range in size from tens to

hundreds of nanometers depending on the size of the polycarbonate membrane filters and

the pressure used during extrusion and have polydispersities of 20-30% [2, 3]. If they ae

made in water, they are also spherical [4]. The size distribution of the extruded vesicles has

been characterized by different methods such as : freeze-fracture-electron microscopy [2, 4],

4

CHAPTER 2. SCATTERING ANGLE DEPENDENCE OF THE SIZE OF VESICLES MEASURED BY

cryogenic transmission electron microscopy [5, 6], field flow fractionation [7], static light

scattering [8] and dynamic light scattering (DLS) [9, 10, 11, 12, 13]. In particular DLS has

been used extensively because sample preparation is simple, the measurement is noninvasive

and the measurement time is relatively short compared to other methods, including static

light scattering.

DLS measures the intensity-intensity autocorrelation function of light scattered by a

sample, which can be related to the field-field autocorrelation function. The field-field au-

tocorrelation function decays exponentially with a decay rate proportional to the diffusion

coefficient for monodisperse samples, or a distribution of decay rates if the sample is poly-

disperse. If the sample is polydispserse, correlation function data are usually analyzed in

terms of the moments or cumulants of the decay rate distribution [14, 15]. Alternatively,

the decay rate distribution is determined by direct numerical inversion of DLS data using

regularized Laplace inversion (Contin) [16]. The mean and standard deviation of the decay

rate distribution can be determined by either method, from which the hydrodynamic radius

and relative standard deviation are calculated.

Because the scattered intensity is weighted by both the mass and the form factor of the

scattering objects, the hydrodynamic radius determined by these techniques can depend on

the angle at which the scattered intensity is measured. This effect is significant for larger,

more polydisperse particles such as vesicles. To overcome this problem, the number-weighted

distribution, which is scattering angle dependent, must be determined.

There are a few groups that directly determine number-weighted distributions. A


number-weighted distribution of arbitrary functionality can be obtained using discrete Laplace

inversion algorithms [17]. Alternatively, number-weighted distributions have been deter-

mined by multiangle DLS measurements [18]. This requires long measurement times, espe-

cially when large vesicles and small angles are involved.

More common is to attempt to convert the results of either Cumulants analysis or an

intensity-weighted distribution to a number-weighted distribution. Results from Cumulants

analysis have been converted to number distributions that have the form of a log-normal

distribution [19, 20], a Gaussian or normal distribution [21], a Schulz distribution [19, 22, 23]

or even an arbitrary distribution function [24]. These methods are based on the assumption

that particles are small such that the form factor is equal to 1. In fact, this assumption is

not valid in many measurements involving large vesicles. The form factor can be included

in conversion of intensity-weighted to number-weighted distributions [25].

This chapter describes the comparison of the intensity-weighted and number-weighted

mean radii and polydispersities of vesicles. Two different sizes of vesicles were used in order

to see the effect of vesicle size on the results. DLS measurements were made at a wide range

of scattering angles in order to show the scattering angle dependence of the results. The

non-linear least squares fitting methods were used to fit the intensity-intensity autocorre-

lation function to the data, where the the intensity or number-weighted distribution was

included in the fitting procedure. Two functional forms for the distribution were investi-

gated, Gaussian and Schulz, in order to show the effect of the symmetry of the distribution


on the results. There is a scattering angle dependence of the intensity-weighted mean ra-

dius for both the Schulz and Gaussian distributions. The scattering angle dependence of

number weighted mean radius was found to be insignificant when a Schulz distribution was

used, but only acceptable for small vesicles when the Gaussian distribution was used. The

results also suggest that an asymmetric distribution is more appropriate to represent the

size distribution of extruded vesicles.

2.2 Radius distribution by DLS

2.2.1 Types of Distribution

Gaussian and Schulz distributions are used in these studies to represent symmetric and

asymmetric distributions, respectively. The Gaussian distribution is written as

G(R) =

(

1√2πs

)

exp

[

−(R − R)2

2s2

]

, (2.1)

where R and s are the mean radius and standard deviation, respectively. We define polydis-

persity as the relative standard deviation, σ = s/R. The Schulz distribution, on the other

hand, is written as,

G(R) =

(

m + 1

R

)m+1 Rm

m!exp

[

−R(m + 1)

R

]

, (2.2)

where σ2 = 1/(m + 1).

Figure 2.2.1 show Gaussian and Schulz probability distribution function for vesicles with

mean radius of 60 nm and polydispersities of a) 0.14, b) 0.33, and c) 0.45. Both distributions


coincide when polydispersity is small and deviates from each other when polydispersity is

large.

2.2.2 Intensity-weighted Distributions

The quantity measured in DLS is the intensity-intensity autocorrelation function g(2)(τ).

In most cases, this function can be written in terms of field-field autocorrelation function

g(1)(τ) through the Siegert relation [26, 27],

g(2)(τ) = 1 + β[g(1)(τ)]2 , (2.3)

where τ is the delay time and β is a factor that depends on the experimental geometry.

For monodisperse particles undergoing Brownian motion, the field-field autocorrelation

function decays exponentially

g(1)(τ) = exp[−Γτ ] , (2.4)

where the decay rate Γ = Dq2 depends on the diffusion coefficient of the particles D and

the magnitude of the scattering wave vector q = 4πnλ

sin θ2 , where λ is the wavelength of the

light source in vacuum, n is the refractive index of the medium and θ is the scattering angle.

The hydrodynamic radius can then be determined using the Stokes-Einstein relation:

Rh =kBT

6πD=

kBTq2

6πΓ. (2.5)

For polydisperse particles, there will be a distribution of decay rates instead of a single

decay rate. In this case, g(1)(τ) is given by

g(1)(τ) =

∞∫

0

G(Γ) exp[−Γτ ]dΓ , (2.6)


0.00

0.01

0.02

0.03

0.04

Gauss distribution Schulz distribution

G(R

)a. = 0.14

0.000

0.005

0.010

0.015

0.020 b. = 0.33

G(R

)

0 20 40 60 80 100 120 1400.000

0.005

0.010

0.015

0.020 c. = 0.45

Radius (nm)

G(R

)

Figure 2.1: Gaussian and Schulz probability distribution function for vesicles with meanradius of 60 nm and polydispersities of a) 0.14, b) 0.33, and c) 0.45.


where G(Γ) describes the distribution of decay rates and is normalized. G(Γ) is characterized

by a mean decay rate

Γ =

∞∫

0

Γ G(Γ)dΓ (2.7)

and relative variance

sΓ2

Γ2 =

∞∫

0

(

Γ − Γ)2

Γ2 G(Γ)dΓ . (2.8)

After assuming a functional form for G(Γ), the mean decay rate Γ and associated polydisper-

sity σΓ can be determined directly by fitting Eq. 2.3 to the intensity-intensity autocorrelation

function data using Eq. 2.6 for the field-field autocorrelation function.

Alternatively, in a moments-based analysis, g(1)(τ) is expanded in terms of the moments

of the distribution [15]

g(1)(τ) = exp[−Γτ ](

1 +µ2

2!τ2 −

µ3

3!τ3 + ...

)

, (2.9)

where µn are the moments of the distribution and, in particular, µ2 = sΓ2 and µ3 represent

the variance and skewness of the distribution, respectively. Γ and σΓ can be determined

directly by fitting Eq. 2.3 to the intensity-intensity autocorrelation function data using

Eq. 2.9 for the field-field autocorrelation function.

After the decay rate distribution has been determined, the particle size and size distri-

bution can be estimated. The hydrodynamic radius Rh is calculated using Γ and Eq. 2.5. In

these methods, the relative variance of Rh is generally assumed to be equal to the relative

variance of Γ.

Instead of working with a decay rate distribution, the analysis can be reformulated in

terms of a radius distribution. Equation 2.6 can be written in terms of an intensity-weighted


radius distribution Gi(R) using Eq. 2.5

g(1)(τ) =

∞∫

0

Gi(R) exp

[

−kBTq2

6πηRτ

]

dR, (2.10)

where Gi(R)dR = G(Γ)dΓ and Gi(R) is normalized. After assuming a functional form for

Gi(R), the intensity-weighted mean radius Ri and polydispersity σRican be determined

directly by fitting Eq. 2.3 to the intensity-intensity autocorrelation function data using

Eq. 2.10 for the field-field autocorrelation function.

2.2.3 Number-weighted Distribution

In general, the scattered intensity is proportional to the square of the mass M of the particle

so that each particle’s contribution to the scattered intensity depends strongly on its size.

The scattered intensity also has an angular dependence due to interference effects that is

represented by the form factor of the particles. This angular dependence becomes more

pronounced for larger particles, with larger particles contributing more to the scattered

intensity at small angles than at large angles. Consequently, Rh and Ri determined from

G(Γ) and Gi(R) become angle-dependent. In order to determine a size distribution that is

independent of angle, M2 and the form factor should be included in the analysis.

For vesicles, the form factor can be written as [10]

F (R) =

[

sin(qR)

qR

]2

, (2.11)

and M2 is proportional to R4. The expression for g(1)(τ) in terms of the number-weighted


distribution Gn(R) is then written as,

g(1)(τ) =

∞∫

0

Gn(R)R4F (R) exp[

−kBTq2

6πηRτ]

dR

∞∫

0

Gn(R)R4F (R) dR

. (2.12)

The number-weighted distribution Gn(R) can be determined directly by fitting Eq. 2.3

to the intensity-intensity autocorrelation function data using Eq. 2.12 for g(1)(τ) with an

appropriate functional form of Gn(R). The mean number-weighted radius Rn and polydis-

persity σRncan then be determined from this distribution.

2.2.4 Relation between Gi(R) and Gn(R)

Number-weighted distribution Gn(R) can be written in terms of intensity-weighted distri-

bution Gi(R) as [17]:

Gn(R) = AGi(R)

R4F (R), (2.13)

where

A =

∞∫

0

Gi(R)

R4F (R)dR

−1

(2.14)

has been introduced so that Gn(R) is normalized.

The mean radius and polydispersity of the number-weighted distribution Gn(R) can

then be calculated from the intensity-weighted distributions using Eq. 2.13.

2.2.5 Relation between Rh and the number-weighted distribution

From the Stokes-Einstein relation, Eq. 2.5, the hydrodynamic radius can be written as,

Rh =kBTq2

6π

1

Γ=

kBTq2

6π

1

R−1. (2.15)


The hydrodynamic radius and associated polydispersity σRhcan then be calculated from

the number distribution Gn(R),

Rh =

∞∫

0

Gn(R)R4F (R) dR

∞∫

0

R−1Gn(R)R4F (R) dR

=

∞∫

0

Gn(R)R4F (R) dR

∞∫

0

Gn(R)R3F (R) dR

(2.16)

and

σ2Rh

=

[

∞∫

0

Gn(R)R4F (R) dR

] [

∞∫

0

R2Gn(R)F (R) dR

]

[

∞∫

0

R3Gn(R)F (R) dR

]2 − 1 . (2.17)

2.3 Materials and Methods

2.3.1 Extruded Vesicle Preparation

The method that we use to prepare vesicles has been described previously [28]. The lipid

used in these studies was 1-stearoyl-2-oleoyl-sn-glycero-3-phosphatidylcholine (SOPC); it

was purchased from Avanti Polar Lipids (Birmingham, AL). The vesicles were prepared by

hydrating SOPC in purified water from a Milli-Q plus water purification system (Millipore,

Bedford MA). The use of purified water ensures the production of spherical vesicles [4]. The

hydrated sample was taken through a freeze-thaw-vortex process and pre-extruded through

2 x 400 nm diameter PCTEs. The pre-extruded samples were then extruded through PCTEs

with nominal pore radii of 50 and 100 nm. For the rest of the paper, vesicles produced using

these pore sizes are defined as 50 nm and 100 nm vesicles, respectively.


2.3.2 DLS Measurements

DLS measurements were performed on samples consisting of a ratio of 0.1 mg lipid to 1

ml water. Five measurements were taken for each sample at scattering angles ranging from

20◦ to 150◦. An ALV DLS/SLS-5000 spectrometer/goniometer manufactured by ALV-Laser

GmbH (Langen, Germany) and a HeNe laser (λ = 633 nm) were used for these experiments.

Fits to the data were made using weighted non-linear least squares fitting routines. In

general, no significant difference in the goodness of fit parameter was observed for use of

Schulz as compared to Gaussian distributions in fits of the same data.

2.4 Results and Discussions

Equation 2.3 was fit to the intensity correlation function data using the four different forms

of g(1)(τ), Eqs. 2.6, 2.9, 2.10 and 2.12, derived above. The hydrodynamic radius Rh and

polydispersity σRhwere determined from Γ and σΓ obtained both from the results of fitting

for the decay rate distribution and from moments analysis. A Gaussian distribution was

used for G(Γ) in Eq. 2.6 and it was found to be necessary to fit Eq. 2.9 only up to second

order. The results for the hydrodynamic radius Rh and the polydispersity σRhas a function

of q for 50 nm and 100 nm vesicles are shown in Figs. 2.2a and 2.2b, respectively. Rh

has been normalized to the value of Rh at 20◦ to allow comparison of the q-dependence

of the hydrodynamic radius calculated for 50 nm and 100 nm vesicles. The results from

both approaches are consistent. Rh decreases with increasing q, and the q-dependence is

more pronounced for the larger vesicles. Results for σRhdo not indicate any particular


q-dependence.

The intensity-weighted mean radius Ri and polydispersity σRiwere determined from the

results of fitting with Eq. 2.10, while the number-weighted mean radius Rn and polydis-

persity σRnwere determined from the results of fitting with Eq. 2.12. Figures 2.3 and 2.4

show a) the intensity-weighted mean radius Ri and b) the polydispersity σRias a function

of q for 50 nm and 100 nm vesicles, respectively. The results for Rh and σRhdetermined by

using g(1)(τ) expressed in terms of the Gaussian decay-rate distribution are also shown for

comparison. As expected, the intensity-weighted mean radius also decreases as q increases,

with the q-dependence more pronounced for larger vesicles. The mean radius of the larger

vesicles decreases by 20% over the angular range measured in these experiments.

There is a significant difference between the mean radius and polydispersity found using

the decay rate and radius distributions in g(1)(τ). This reflects the fact that these are

actually different averages; Ri is obtained by averaging over R while Rh is obtained by

averaging over 1/R−1. Ri is larger than Rh for both Gaussian and Schulz distributions

where the Gaussian Ri is slightly smaller than the Schulz Ri. On the other hand, the Schulz

σRiis similar to σRh

while and the Gaussian σRiis systematically smaller than both the

Schulz σRiand σRh

.

Relation between Ri and Rh as well as σRiand σRh

can be observed by calculating the

values of Rh and σRhfrom a value of Ri as a function of σRi

. Calculation was done using

Eqs.2.16 and 2.17 without R4 and F (R) terms and Gi(R) was used instead of Gn(R). Results

of the calculation for vesicles with Ri of 60 and 90 nm where Gaussian and Schulz distribution


0.80

0.84

0.88

0.92

0.96

1.00

0.005 0.010 0.015 0.020 0.0250.21

0.24

0.27

0.30

0.33

0.36

Rh/R

h(20

o )

a.

b.

Moment based-analysis ( 50 nm) Moment based-analysis (100 nm) Decay rate distribution ( 50 nm) Decay rate distribution (100 nm)

Rh

q (nm-1)

Figure 2.2: a) The hydrodynamic radius normalized by the radius measured at 20◦ andb) associated polydispersity of 50 nm and 100 nm vesicles as a function of q. The resultswere determined using g(1)(τ) consisting of either a decay rate distribution (Eq. 2.6) ormoments-based analysis (Eq. 2.9).


63

66

69

72

75

78

0.005 0.010 0.015 0.020 0.025

0.20

0.22

0.24

0.26

0.28

0.30

G(Ri) Gaussian G(Ri) Schulz G( ) Gaussian

Ri (n

m)

b.

Ri

q (nm-1)

a.

Figure 2.3: a) The intensity-weighted mean radius Ri and b) the polydispersity σRiof

50 nm vesicles as a function of q. The results were determined using g(1)(τ) expressed interms of the intensity-weighted radius distribution Gi(R) (Eq. 2.10), where Gaussian andSchulz distributions were used for Gi(R). For comparison, the results for 50 nm vesiclesfrom Fig. 2.2 are also shown.


90

100

110

120

0.005 0.010 0.015 0.020 0.025

0.24

0.26

0.28

0.30

0.32

0.34

G(Ri) Gaussian G(Ri) Schulz G( ) Gaussian

a.

Ri (n

m)

b.

Ri

q (nm-1)

Figure 2.4: a) The intensity-weighted mean radius Ri and b) the polydispersity σRiof

100 nm vesicles as a function of q. The results were determined using g(1)(τ) expressedin terms of intensity-weighted radius distribution Gi(R) (Eq. 2.10), where Gaussian andSchulz distributions were used for Gi(R). For comparison, the results for 100 nm vesiclesfrom Fig. 2.2 are also shown.


were used for Gi(R) are shown in Fig. 2.5. In Fig. 2.5 Rh and σRhare presented in terms of

their ratio to Ri and σRi, respectively, for clarity. The figure shows that for polydispersity up

to 25% Rh is 7% smaller than Ri and the effect of the distribution symmetry is not significant.

On the other hand, σRhis 5% larger than σRi

when Gaussian distribution was used for Gi(R)

but there is no significant difference between these two values when Schulz distribution was

used. The figure shows that these two averages coincide when the polydispersity is small and

start to deviate from each other when polydispersity is large. When Gaussian distribution

was used, however, these two averages starts to deviate at smaller value of polydispersity

than when Schulz distribution was used. These results suggest that Ri and Rh as well as

σRiand σRi

can be used interchangeably when the polydispersity is relatively small but

need to be distinguished when the polydispersity is large.

Figures 2.6 and 2.7 show a) the number weighted mean radius Rn and b) the poly-

dispersity σRnas a function of q for 50 nm and 100 nm vesicles, respectively. With the

exception of the data taken at smallest q, there is no significant q-dependence for Rn and

σRndetermined using the Schulz distribution and values obtained for the polydispersity

are reasonable. However, when the Gaussian is used, more q-dependence is observed in the

results for Rn, especially for the larger vesicles. There is a significant difference in the values

found for Rn and σRndepending on whether the Gaussian or Schulz distribution is used.

The mean radius found using the Schulz distribution is larger than that found using the

Gaussian distribution. The value of σRnfrom the Gaussian distribution, however, is very

large particularly at small q. The fact that the Schulz distribution does a better job fitting


0.84

0.87

0.90

0.93

0.96

0.99

0.05 0.10 0.15 0.20 0.25 0.30 0.35

1.0

1.5

2.0

2.5

3.0

60 nm Gaussian 90 nm Gaussian 60 nm Schulz 90 nm Schulz

Rh/R

ia.

b.

Rh/

Ri

Ri

Figure 2.5: a) Ratios of Rh to Ri and b) σRhto σRi

for vesicles with Ri of 60 and 90 nm asa function of σRi

.


data for extruded vesicles is consistent with the work of other authors [17].

The exception for q-independence results of Rn at small q is caused by the fact that at

this q-range, the measurement is very sensitive to dust or aggregates in the sample. Neither

would be accounted for in these distributions, which are monomodal.

The number-weighted mean radius Rn and the associated polydispersity σRncan also be

determined from the intensity-weighted mean radius Ri and the associated polydispersity

σRiusing Eq. 2.13. Using Ri and σRi

from the results of the fits as shown in Figs. 2.3 and

2.4, Rn and σRnwere calculated. The results of the calculation are shown in Figs. 2.6 and

2.7 for 50 nm and 100 nm vesicles, respectively. Only the results for the Schulz distribution

are shown in Figs. 2.6 and 2.7. For this case, calculated Rn and σRnagree well with the

results from directly fitting the number-weighted distribution to the data. However, the

results for the Gaussian distribution are not shown because the results calculated for Rn

and σRnare significantly different from the results of the direct fit of the number distribution

to the data; results for Rn are much too small and for σRnare much too large.

One possible source for the failure of the conversion from an intensity-weighted Gaussian

distribution to a number-weighted distributions is that the conversion for Gaussian is very

sensitive to the polydispersity. This can be confirmed by calculating Rn and σRnfrom Ri

and σRiusing Eq. 2.13 and a range of σRi

. Calculated results for Rn and σRnfor vesicles

with Ri of 60 and 90 nm as a function σi are shown in Figs. 2.8a and 2.8b, respectively.

The values chosen for Ri are close to those measured for 50 and 100 nm vesicles, respec-

tively. In the graph, the number-weighted values are normalized by the intensity-weighted


30

40

50

60

0.005 0.010 0.015 0.020 0.0250.2

0.3

0.4

0.5

0.6

Gn(R) Gaussian Gn(R) Schulz Gn(R) Schulz (Calculation)

Rn (n

m)

a.

b.

Rn

q (nm-1)

Figure 2.6: a) The number-weighted mean radius Rn and b) the polydispersity σRnof

50 nm vesicles as a function of q. The results were determined using g(1)(τ) expressed interms of number-weighted radius distribution Gn(R) (Eq. 2.12), where Gaussian and Schulzdistributions were used for Gn(R). For comparison, Rn and σn, calculated using Eq. 2.13and fit results for Ri and σRi

from Fig. 2.3, are also shown. The error bars represent thestandard deviation of the mean value from 5 measurements.


30

45

60

75

90

105

0.005 0.010 0.015 0.020 0.0250.2

0.3

0.4

0.5

0.6

Gn(R) Gaussian Gn(R) Schulz Gn(R) Schulz (Calculation)

Rn (n

m)

a.

b.

Rn

q (nm-1)

Figure 2.7: a) The number-weighted mean radius Rn and b) the polydispersity σRnof

100 nm vesicles as a function of q. The results were determined using g(1)(τ) expressed interms of number-weighted radius distribution Gn(R) (Eq. 2.12), where Gaussian and Schulzdistributions were used for Gn(R). For comparison, Rn and σn, calculated using Eq. 2.13and fit results for Ri and σRi

from Fig.2.4, are also shown. The error bars represent thestandard deviation of the mean value from 5 measurements.


0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.05 0.10 0.15 0.20 0.25 0.30 0.35

1

2

3

4

5

6

50 nm vesicles (Gaussian) 100 nm vesicles (Gaussian) 50 nm vesicles (Schulz) 100 nm vesicles (Schulz)

Rn/R

ia.

b.

Rn/

Ri

Ri

Figure 2.8: Values for a) Rn and b) σRncalculated as a function of σi of vesicles with Ri

values of 60 and 90 nm using Eq. 2.13 for both Gaussian and Schulz distributions of vesicles.


values to make it easy to compare the results for 50 and 100 nm vesicles. When σi is small,

Rn and σn are almost the same as Ri and σi, respectively. As σi increases, Rn becomes

smaller than Ri while σRnbecomes larger than σi. For σi larger than a certain threshold

value, Rn becomes very small while σRnvery large. In this case, the threshold value above

which the difference between the two results is unacceptable is much smaller for the Gaus-

sian distribution, approximately 0.15, than it is for the Schultz distribution, approximately

0.30. As the polydispersity observed in these samples is larger than 0.15, the conversion

from an intensity-weighted Gaussian distribution to a number-weighted distribution yields

unreasonable values.

The inconsistency of the results for Rn and σRnbetween directly fitting DLS data and

calculation from Ri and σRiwhen Gaussian distribution were used suggests that conversion

must be done by considering what type of distribution used and range of polydispersity

of the samples. In recent studies on vesicle size distribution, for example, Maulucci et al.

[29] calculated number-weighted mean radius for sonicated vesicles of approximately 6 to

7 nm from intensity-weighted mean radius varying from 30 to 100 nm for different q using

both symmetry (Gaussian) and asymmetry (Log normal) distributions. Radius of 7 nm is

too small for a vesicle that is predicted to have a smallest value of approximately 11 nm

[30]. There is no data on number-weighted polydispersity from these studies, but by looking

at much different between intensity and number-weighted mean radii, it can be predicted

that the polydispersty is very large. The unreasonable results of the number-weighted mean

radius and polydispersity are most likely caused by large polydispersity of sonicated vesicles.


It has been observed that the polydispersity of sonicated vesicles is larger than 50 % [31],

larger than the threshold value for both symmetry and asymmetry distribution shown in

Fig. 2.8.

2.5 Summary

These studies demonstrate an approach that can be used to directly determine number-

weighted radius distributions from DLS data. By fitting number-weighted radius distri-

butions, and including the mass and form factor in the fitting, the apparent problem of

q-dependence of DLS-measured particle size can be overcome. The studies also show that

results obtained using an intensity-weighted distribution should be converted to number-

weighted results with care, especially if the polydispersity is large. In particular, Gaussian

and Schulz distributions were used to represent symmetric and asymmetric distributions,

respectively, to determine size distributions for vesicles extruded through 50 and 100 nm

pores. The q-dependence of the radius determined using a number-weighted distribution

was found to be insignificant for 50 and 100 nm vesicles when a Schultz distribution was

used, but only acceptable for 50 nm vesicles when the Gaussian distribution was used. The

results also suggest that an asymmetric distribution is more appropriate to represent the

size distribution of extruded vesicles.

Chapter 3

The Influence of the Pore Size

Distribution on the Size

Distribution and Lamellarity of the

Extruded Vesicles

3.1 Introduction

Vesicles are quasi spherical shells made from lipids that enclose an aqueous medium and

separate it from an external aqueous medium. This characteristic makes vesicles a powerful

tool in a wide range of application [1]. For example vesicles can compartmentalize certain

active agents, thus, can be used as vehicles or carriers. They can be used to deliver drugs

27

CHAPTER 3. VESICLE SIZE AND LAMELLARITY 28

to certain site of the body [32] or to deliver biomaterials such as : DNA to various cell

and microorganism in order to alter their genetic code [33] . In addition, because of their

similarities with cells, vesicles can also be used as model cells in molecular biology research

[34]. From all of the possible applications of vesicles in many disciplines, drug delivery seems

to be the most widely investigated area [35].

Vesicles can be produced in a wide range of sizes with different number of bilayers

depending on the techniques applied. However, for most applications, unilamellar vesicles

with the size between 25 nm to 250 nm and narrow distribution are of interest [1, 36]. In

drug delivery for example, narrow size distribution of vesicles averaging less than 50 nm

radius has been proved to be important in the clinical success [40]. The narrow distribution

of vesicles size is required because the circulation lifetime of the drug containing vesicles is

size dependent [41] . Litzinger et al [42] showed that vesicle size and polydispersity have a

strong impact on dosage, targeting, and the rate of clearance from the body when vesicles

are used to deliver drugs to specific targets. The unilamellarity of the vesicles, on the other

hand, is required such that the vesicles can have high efficiency in the content of drug or

other materials captured. Size and lamellarity, therefore, are 2 of the important physical

properties of the vesicles.

Detergent dialysis [43], extrusion through small pores [44], and reverse-phase evaporation

[45] are three of the techniques that can be used to produce vesicles with the range of radius

between 25 and 250 nm [1]. The lower end size of 25 nm radius can be produced using

detergent dialysis or extrusion techniques and the upper end size of 250 nm radius can be


generated using a reverse-phase evaporation technique. The size in between can be achieved

by extrusion methods where the pore size and the pressure can be varied depending on

the size required for the vesicles. These techniques are required to break up multilamellar

large vesicles (MLVs) formed when lipids are dispersed in aqueous environment. In detergent

dialysis, the MLVs are broken up by the presence of particular detergent. The detergent than

is removed from the mixture of detergent-lipid which results in the formation of unilamellar

vesicles with mean radius of 20 nm and can be increased when cholesterol is introduced

in the samples [43]. In reverse-phase evaporation, the organic solvent is used to dissolve

lipids in aqueous environment generating aqueous-organic emulsions. Unilamellar vesicles

with mean radius of 240 nm will be formed when the organic solvent is evaporated [45].

In extrusion, MLVs are pushed through pores in PCTE membranes repetitively at least

10 times. The size and lamellarity of the vesicles are varied from 25 nm to 200 nm and

unilamellar to multilamellar, respectively, depending on the pore size used and the pressure

applied [2, 3].

The use of detergents and organic solvent in detergent dialysis and reverse phase evapo-

ration techniques, respectively, limits the two techniques only to certain detergent or organic

solvent that is suitable for lipids used. There is also an issue on incomplete-removal of de-

tergent and organic solvent from the samples. In addition, the time course for both methods

in producing LUVs is in the order of several hours. Since extrusion requires no agents such

as detergent or organic solvent, this technique advantages detergent dialysis and reverse-

phase evaporation techniques. Moreover time course for extrusion is only in the order of


few minutes.

Number of studies have been conducted to investigate the parameters influencing the size

and lamellarity of the extruded vesicles. These includes studies on the influence of PCTE

pore size on the size and lamellarity of the vesicles [2, 37], the effect of lipid composition,

extrusion pressure and temperature on the size of vesicles [28], the influence of the pore size

and extrusion pressure on the size of the vesicles [3] and the effect of a polymer lipid and

freeze-thawing process on the size and lamellarity of vesicles [38].

Using pore size varying from 15-200 nm, Mayer et al., showed that small pores produces

LUVs, while large pores generates MLVs or combination of LUVs and MLVs. They also

observe the dependence of the vesicle size on the pore size and the results are summarized in

Fig.3.1 [2]. The results show that radius of vesicles increases with the radius of pores. When

the small pores were used, however, the radius of the vesicles is larger than the pores, while

it is opposite when large pores were used. The polydispersity of the vesicles also increases

with the size of the pores. The fact that size of the vesicles represents size of the pores is

only true when small pores were used in extrusion. When larger pores were used, vesicle

size becomes smaller than pore size and deviates more when pore size increases. Although

larger pores can be used in producing larger vesicles, the vesicles are very polydisperse and

multilamellar. This is probably one of the reasons that nominal pore radius of 50 nm is most

frequently used in extrusion, thus limiting the the application of the technique in terms of

size of vesicle produced.


0

50

100

150

200

0 50 100 150 200

0.20

0.25

0.30

0.35

Rv (n

m)

Freeze-Fracture EM DLS Rv = Rp line

a.

b.

Rp (nm)

Figure 3.1: a) Radius and b) polydispersity of extruded vesicles as a function of nominalpore radius. The graphs are based on the data provided in Tab.1 from reference [2].


Knowledge of the factors influencing the polydispersity and the multilamellarity of ex-

truded vesicles produced using large pores may extend the usefulness of extrusion tech-

niques to larger vesicle size. For example applying smaller pressure during extrusion results

in vesicles 20% larger than applying higher pressure [3]. Another factor that can affect

the polydispersity of the vesicles is the polydispersity of the pores. There is a possibility

that large polydispersity of extruded vesicles is contributed by large polydispersity of the

pores. However, most previous studies mentioned used the nominal value of the pores and

no polydispersity of the pores available, thus no information on this relation. Detail char-

acterization on size distributions of the pores and associated extruded vesicles can reveal

the relation between the polydispersities of both pores and vesicles. This will also reveal

the actual pore size instead of the nominal one, so that the relation between vesicle size and

pore size is more reliable.

These studies investigate the influence of the pore size distribution on the size distribu-

tion of associated extruded vesicles and vesicle lamellarity. Scanning electron microscopy

and dynamic light scattering were used to characterize the size of the pores and the size of

extruded vesicles, respectively, while a fluorescence quenching assay was used to characterize

the vesicle lamellarity. The nominal pore radii varying from 25 - 200 nm were used in these

observation.



Lipids used, 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and 1-palmitoyl-2-

(6-((7-nitro2-1,3-benzoxadiazol-4-yl)amino)caproyl)-sn-glycero-3-phosphocholine (NBD-PC)

were purchased from Avanti Polar Lipids Inc. (Alabaster, AL). Polycarbonate track-etched

(PCTE) membranes with nominal radii of 25 and 200 nm were purchased from Osmonics

Inc. (Livermore, CA) and PCTEs with nominal radii of 50 and 100 nm manufactured by

Whatman Nucleopore Inc. (Clifton, NJ) were purchased from VWR Canlab (Missisauga,

ON).

3.2.1 PCTE Pore Size Measurements

PCTE membrane pores were characterized using Scanning Electron Microscopy. The mi-

croscope used was FEI DualBeam Strata 235 Field-Emission Scanning Electron Microscope

(SEM). Prior to the measurements, the PCTE membranes were carbon coated to prevent

electron accumulation on the membrane surface due to the non-conductivity of the mem-

branes. The images were analyzed to determine pore radius using Adobe Photoshop Soft-

ware Version 7 the results were put in the histogram. The histogram data was fit using the

expression of the expectation value for each interval [39],

N = n

Rmax∫

Rmin

G(R)dR, (3.1)

where, n is a constant containing a normalization factor and number of the pores analyzed,

Rmin and Rmax are the minimum and maximum values of the interval in the histogram,

respectively and size distribution G(R) has been assumed to follow Gaussian distribution


from which the mean radius and the polydispersity were determined.

3.2.2 Vesicle Size Measurements

The size distribution of the extruded vesicles was characterized using DLS. Equation 2.3

was fit to the autocorrelation function data where Eq. 2.12 was used as the field-field au-

tocorrelation function and the number-weighted radius distribution was assumed to follow

Schulz distribution.

For these measurements, the samples of vesicles were made from POPC in pure water.

The preparation of the samples of vesicles is similar to that explained in subsection 2.3.1

of Chapter 2. The apparatus used in these measurements have been explained in subsec-

tion 2.3.2 of Chapter 2.

3.2.3 Vesicle Lamellarity Determination

Vesicle lamellarity was determined using a fluorescence quenching assay [46, 47]. The assay

was based on the facts that NBD-lipid derivatives have a yellowish-green fluorescence but

can be quenched by a solution of sodium dithionite. When a solution of sodium dithionite

is introduced to the sample of vesicle made from the mixture of lipid and NBD-lipid, the

exposed NBD-lipid will be quenched resulting in the reduction of the fluorescence intensity.

The number of bilayer can be predicted from the difference in intensity before and after the

introduction of sodium dithionite.

Samples of vesicles were prepared from mixture of POPC and fluorescently labeled lipid,

an NBD-PC, with the amount of NBD-PC less than 1 mol percent. POPC and NBD-PC


were dissolved separately in chloroform and their aliquot of required concentration were

taken and put in the round bottom glass before drying by rotary evaporation. The drying

lipids, then, were placed under vacuum overnight to remove trace residual solvent. The

lipid films were suspended in the solvent of 10 mM Tris with pH of 7.4. The suspension was

then taken through freeze-thaw-vortex process and was extruded once through 2 PCTEs

with nominal pore radius of 200 nm (pre-extrusion). The pre-extruded vesicles were then

extruded through required pore size of PCTEs.

The assay was done using a PTI Quanta Master Luminescence Spectrofluorimeter. 2 ml

of samples was put in the quart cell and was excited using light with a wavelength of

490 nm and the intensity was detected at a wavelength of 532 nm. The values of excitation

and emission wavelengths were chosen after doing excitation and emission scans before the

measurements. One of the scans results is shown in Fig. 3.2.

The intensity was monitored for 300 s where sodium dithionite solution was added after

60 s and Triton X 100 was added after 250 s. Solution of sodium dithionite quenches the

NBD-PC at the outer leaflet for unilamellar vesicles or outer leaflet of the outermost bilayer

for multilamellar vesicles. For simplicity, lipids located at these leaflets are defined as the

outer lipids in the text. Solution of Triton X 100 was used to lyse vesicles to detect the

rest of unquenched NBD-PC located in the inner leaflet of bilayer for unilamellar vesicles or

inner leaflet of the outermost bilayer plus inner bilayers for the multilamellar vesicles. For

simplicity, lipids located at these leaflets are defined as the inner lipids in the text.


350 400 450 500 550 600 650

15

30

45

60

Excitation Emission

Inte

nsity

(104 H

z)

Wavelength (nm)

Figure 3.2: Emission and excitation scans for POPC:NBD-PC vesicles.


The fraction of inner NBD-PC can be calculated through the relation,

Ni =

(

ID − IT

1 − IT

)

100%, (3.2)

where ID and IT is the normalized intensity after introduction of sodium dithionite and

Triton X-100, respectively. In these studies, the NBD-PC molecules are assumed to be

located randomly among POPC molecules, hence, the fraction of inner NBD-PC lipids

represents the fraction of POPC lipids.

For small vesicles, the effect of the radius of curvature on the amount of inner and outer

lipids in the vesicle bilayer/s may need to be taken into account. Number of lipids in two

leaflets of a bilayer is proportional to the area of the sphere formed by those leaflets. The

radii of those spheres are different by the thickness of the bilayer. For unilamellar vesicles,

the fraction of inner lipids as a function of vesicle radius can be estimated using the relation,

Ni(1 bilayer) =

(

(R − t)2

R2 + (R − t)2

)

, (3.3)

where R and t are the vesicle radius and thickness of the bilayer, respectively. For bilamellar

vesicles, the fraction of the inner lipids can be estimated using the relation,

Ni(2 bilayers) =

(

(R − t)2 + (R − 10)2 + (R − 10 − t)2

R2 + (R − t)2 + (R − 10)2 + (R − 10 − t)2

)

, (3.4)

where the separation between the 2 bilayers is assumed to be 10 nm [2].

Figure 3.3 shows the fraction of the inner lipids for a unilamellar vesicle as a function of

the vesicle radius assuming bilayer thickness of 3.9 nm [48]. For simplicity, the line in Fig. 3.3

is called unilamellar line. If the fraction of the inner lipids from vesicles with a particular

radius estimated from Eq. 3.2 is within the unilamellar line, the vesicles are unilamellar. If


the fraction is above the unilamellar line, vesicles can be multilamellar or both unilamellar

and multilamellar. For the purpose of these studies, it is assumed that vesicles only have up

to 2 bilayers (bilamellar vesicles), therefore bilamellar and multilamellar vesicles will be used

interchangeably in the text of this chapter. The fraction of unilamellar (N1) and bilamellar

(N2) vesicles, then, can be determined from the value of Ni from the fluorescence quenching

assay using the relation,

Ni = N1 ∗ Ni(1 bilayer) + N2 ∗ Ni(2 bilayers), (3.5)

where N1 + N2 = 1.


3.3.1 Pore Size Distribution

The SEM images of each PCTE membrane were taken at at least 30 random areas. A

typical SEM micrograph of surfaces of PCTE membranes was shown in Fig. 3.4. The mean

radius and the polydispersity of the pores in these images were determined by analyzing

556, 365, 596 and 1028 pores for 25 nm, 50 nm, 100 nm and 200 nm pores, respectively. The

radius data was arranged in a histogram and Eq. 3.1 was used to fit the data to estimate

the mean radius and the standard deviation.

Figure 3.5 shows the data of the pore radius of a) 25 nm, b) 50 nm, c) 100 nm and d)

200 nm pores, respectively. The figure also contains the results of fitting Eq. 3.1 to the data

(dots and lines). The dots represent the expectation values and the lines describe Gaussian


0 50 100 150 2000.405

0.420

0.435

0.450

0.465

0.480

0.495

Frac

tion

of in

ner l

eafle

t lip

ids

Radius (nm)

Figure 3.3: Fraction of inner lipids for unilamellar vesicles as a function of vesicle radius.The area above the line indicates multilamellar vesicles or combination of unilamellar andmultilamellar vesicles.


Figure 3.4: Scanning electron micrograph of surface of PCTE membranes with nominal poreradii of a) 25 nm, b) 50 nm, c) 100 nm, and d) 200 nm


distribution with the mean radius and standard deviation resulting from the fits. The mean

radius and the standard deviation (written in term of polydispersity) are summarized in

Tab. 3.1.

Nominal Radius (nm) Measured Radius (nm) Polydispersity

25 31.4 0.15350 36.5 0.101100 65.2 0.086200 155.2 0.105

Table 3.1: The mean radius and the polydispersity of the pores.

The results show a significant difference between the nominal pore radius with the mea-

sured pore radius of PCTE membranes. Membrane with nominal pore radius of 25 nm has

the measured mean radius 25.6 percent larger than the nominal pore radius, while mem-

branes with nominal pore radii of 50, 100 and 200 nm, have the measured mean radii of

27.0, 34.8 and 22.4 percent, respectively, smaller than the nominal pore radii. In addition,

the polydispersity decreases with increasing nominal pore size.

3.3.2 Vesicle Size Distribution

POPC lipids were used in experiments on pore size distribution dependence of vesicle size

distribution. Samples of POPC lipid dispersion were extruded through PCTEs with nominal

pore radii ranging from 25 to 200 nm continued by measuring size distribution in order to

see the effect of the pore size distribution on the vesicles distribution. The samples were

extruded at applied pressures of 300 psi when 25 nm and 50 nm pores were used, and 250

and 50 psi when 100 nm and 200 nm pores, respectively, were used. The pressures applied


0

20

40

60

80

100

120

140 a. Nominal Rpore = 25 nm

Num

ber

of p

ores

15 20 25 30 35 40 45 500

20

40

60

80

100

120b. Nominal Rpore = 50 nm

R (nm)

Num

ber

of p

ores

40 50 60 70 80 900

20

40

60

80

100

120

140

160 c. Nominal Rpore = 100 nm

N

umbe

r of

por

es

R (nm)

100 120 140 160 180 200 220 2400

50

100

150

200

250

300

350 d. Nominal Rpore = 200 nm

R (nm)

Num

ber

of p

ores

Figure 3.5: Histogram of pore radius data for a) 25 nm, b) 50 nm, c) 100 nm and d)200 nm pores. The dots represent the expectation values for each interval resulting fromfitting Eq. 3.1 to the radius data and the lines describe Gaussian distribution with the meanradius and standard deviation taken from the fit results.


are the optimum pressures to produce vesicle size as small as possible. Previous studies

showed that radius of vesicles decreases with the pressure [3].

Size distribution of vesicles produced were characterized using DLS. The expression for

intensity-intensity autocorrelation function of Eq. 2.3 was fit to the correlation function data

where Eq. 2.12 was used for field-field autocorrelation function and the number-weighted

radius distribution Gn(R) was assumed to follow the Schulz distribution. The number-

weighted mean radius and the polydispersity were determined from fitting procedure.

Figure 3.6 shows a) the mean radii and b) the polydispersities of the vesicles as a

function of nominal pore radius. For comparison, the mean radii and polydispersities of the

pores from Tab. 3.1 are also shown. The bars in Fig. 3.6a are the standard deviations of

the distributions. The figure shows that the mean radii of vesicles produced using 25 nm

and 50 nm pores are larger than the measured mean radii of the pores. In contrast, the

mean radii of the vesicles produced using 100 nm and 200 nm pores are smaller than the

measured mean radii of the pores. The figure also shows that the polydispersity of the

vesicles increases with the radius of the pores up to nominal radius of 100 nm and levels

off for larger pores. The polydispersity of the vesicles produced using larger pores is large,

although the polydispersity of the pores is small. This suggests that the polydispersity of

the vesicle is not dependent on the polydispersity of the pores but is only dependent on

pore size.

To compare the size distributions of the pores and the vesicles, probability distribution

functions representing pore size distribution (Gaussian) and vesicle size distribution (Schulz)


0 50 100 150 2000.0

0.1

0.2

0.3

0.4

Nominal Rpore

(nm)

50

100

150

R

(nm

)

Figure 3.6: a) The mean radii and b) the polydispersities of the extruded vesicles and thepores as a function of nominal pore radius.


0.00

0.03

0.06

0.09

0.12

G(R

)G(R

)

a. Rpore = 25 nm

0.00

0.03

0.06

0.09

0.12

b. Rpore = 50 nm

0 25 50 750.00

0.03

0.06

0.09

0.12

R (nm)

G(R

)

G(R

)

c. Rpore = 100 nm

0 50 100 150 200 2500.00

0.01

0.02

0.03

d. Rpore = 200 nm

R (nm)

Figure 3.7: Comparison of the radius distributions of the vesicles and PCTE pores forvesicles extruded through PCTEs with nominal radii of a) 25 nm, b) 50 nm, c) 100 nm, andd) 200 nm. The dashed lines represent radius distribution of the pores determined from thefits in Fig. 2, while the solid lines are the number-weighted radius distribution of vesiclesresulting from fitting Eq. () to the correlation function data of DLS measurements


are plot in Fig. 3.7. Figures 3.7a, b, c and d show radius distributions of 25 nm, 50 nm,

100 nm and 200 nm pores, respectively, and associated vesicles produced.

Most of size distribution results in these studies agree within the uncertainty with those

observed by Mayer et al. from freeze-fracture electron microscopy, but not from DLS [2].

This is expected because the mean radius in these measurements is number-weighted and so

is the mean radius from electron microscopy. The difference between the DLS results from

these studies and the one from Mayer’s group is probably due to the different analysis of DLS

data. Unfortunately, there is no explanation what kind of analysis they were using. However,

as explained in Chapter 2, for polydisperse vesicles, the mean radius and polydispersity

from DLS analysis vary depending on whteher the analysis is intensity-or number-weighted

bases. The number-weighted mean radius is smaller than the intensity-weighted one. It is

suspected, therefore, that the DLS results from Mayer’s group is intensity-weighted base,

thus, larger than DLS results from these studies. Since most of DLS results from these

studies agree well with electron microscope results from Mayer’s group, the pore sizes used

in both measurements are most likely similar. Only the results from 100 nm pore show

significant difference in the mean radius that could be due to the different size of the pores

used.

The facts that the polydispersity of the vesicles is dependent on the size of the pores but

not on the polydispersity of the pores raise the question of how size of the pores influencing

the polydispersity of extruded vesicles. One possible explanation is that the pre-extruded


samples contain small vesicles that can pass through large pores easily without being rup-

tured. When small pores were used the polydispersity is small because the small vesicle

size is comparable to the size of the pores. However, when large pores were used, the poly-

dispersity is large since the small vesicles is smaller than the size of the pores; the vesicles

formed is, therefore, a combination of small vesicles and the one comparable to the size of

the pores. The argument of the small vesicle presence is supported by previous studies on

the vesicle size distribution using fractionation of vesicles through column chromatography

[49] and also using freeze-fracture electron microscopy [2].

The presence of small vesicles may be caused by freeze-thaw-vortex (FTV) process.

If it was the case, producing vesicles using larger pores without FTV process will help

getting rid of the small vesicles thus reducing the polydispersity of vesicles and increasing

the vesicle mean radius to be comparable to the mean radius of the pores. Experiments

were done to compare the size distribution of vesicles produced by extruding samples of

lipid dispersion through 100 nm pores with and without an FTV process and comparing the

radius distribution of the vesicle to the radius distribution of the pores. Figure 3.8 shows

the radius distribution of the 100 nm vesicles prepared with and without FTV process and

radius distribution of the 100 nm pores. Applying FTV process produces more small vesicles

indicated by small mean radius and large polydispersity. However, there are still many

vesicles smaller than the size of the pores shown by vesicles without FTV process that must

be caused by other means. Small vesicles, then, probably produced spontaneously when

lipid was dispersed in water, or as a consequence of the process of extrusion.


0 50 100 1500.000

0.015

0.030

0.045

0.060

0.075 with FTV; Rn = 58.5 nm, n = 0.353

without FTV; Rn = 63.7 nm, n = 0.28

100 nm pores; Rn = 65.2 nm, n = 0.086

Gn(R

)

Radius (nm)

Figure 3.8: Radius distribution of 100 nm vesicles prepared with and without FTV processand radius distribution of the 100 nm pores.


3.3.3 Vesicle Lamellarity

POPC and NBD-PC lipids were used in experiments on pore size distribution dependence

of vesicle lamellarity. Samples of POPC:NBD-PC lipid dispersion were extruded 10 times

each through 50 nm and 100 nm pores continued by a fluorescence quenching assay.

Figure 3.9a shows the intensity of the POPC:NBD-PC vesicles before and after the

introduction of sodium dithionite and Triton X-100 for samples of 50 nm and 100 nm

vesicles. For comparison, the data for samples of FTV and pre-extruded vesicles is also

shown. The fraction of inner NBD-PC representing fraction of inner lipids was calculated

using Eq. 3.2 and is shown in Fig. 3.9b. In Fig. 3.9c the fraction of inner NBD-PC is

plot as a function of vesicle radius. Values of 44.9, 58.8 and 200 nm were used as the

radii for 50 nm, 100 nm and preextruded vesicles, respectively. The radius of FTV vesicles

is very large but the value of 250 nm was used for clarity. The unilamellar line is also

included in Fig. 3.9c. The figure shows that only 50 nm vesicle data point is very close

to the unilamellar line, consequently, the 50 nm vesicles are unilamellar. There could be

a combination of unilamellar and bilamellar or only bilamellar for FTV, pre-extruded and

100 nm vesicles due to the fact that their data points are above the unilamellar line.

Using the value of inner lipid fraction and the vesicle radius for each sample, the fraction

of unilamellar and bilamellar vesicles was determined through Eq. 3.5. The results are

summarized in Tab. 3.2.

The results in Tab. 3.2 indicate that FTV vesicles are bilamellar, pre-extruded and

100 nm vesicles contains combination of unilamellar and bilamellar vesicles, while 50 nm


0 50 100 150 200 250 3000.0

0.2

0.4

0.6

0.8

1.0

0 50 100 150 200 2500.3

0.4

0.5

0.6

0.7

0.8

0.9

Inte

nsity

(Nor

mal

ized

)

Time (s)

Un-extruded Preextruded vesicle 100 nm vesicles 50 nm vesicles

Adding Sodium Dithionite

Adding Triton X-100

un-extruded pre-extruded 100nm 50nm0

20

40

60

80

Frac

tion

of in

ner l

ipid

s (%

)

Samples

c.

b.

Frac

tion

of in

ner l

ipid

s

Vesicle radius (nm)

Unilamellar line Data Expected value for unilamellar vesicles

a.

Figure 3.9: a) The intensity (normalized) of POPC:NBD-PC vesicles before and after in-troduction of sodium dithionate to quench the outer NBD-PC and Triton X-100 to lyse thevesicles, consequently quenching inner-NBD PC. b) Fraction of inner lipids calculated usingEq. 3.2. c) Fraction of inner lipids as a function of vesicle radius in the presence of theunilamellar line.


1 2 3 4 6 8 1035

40

45

50

55

Frac

tion

of in

ner l

ipid

s (%

)

Number of passes

100 nm pores 50 nm pores

Figure 3.10: Fraction of inner lipids as a function of number of passes for vesicles extrudedthrough 50 nm and 100 nm pores.


Samples Fraction of 1 bilayer vesicles (%) Fraction of 2 bilayer vsicles (%)

FTV 0 100Pre-extruded 51 49100 nm vesicles 72 2850 nm vesicles 100 0

Table 3.2: Fraction of unilamellar and bilamellar vesicles.

vesicles are unilamellar. Number of unilamellar vesicles increases by 13.8% and 53.8% when

the pre-extruded vesicles went through 100 and 50 nm pores, respectively. The results

indicate that small pores are more effective in rupturing multilamellar vesicles than large

pores.

The experiments were also done to observe the influence of the number of passes through

the extruder on the lamellarity of the vesicles. Figure 3.10 shows the percentage of the inner

lipids as a function of number of passes through extruder for 50 and 100 nm vesicles. The

fraction of unilamellar and bilamellar vesicles was determined through Eq. 3.5 and the results

are summarized in Tab. 3.3. It takes only 1 pass through extruder to reach an equilibrium

value of vesicle lamellarity. The results leads to a following possible scenario of vesicle

rupturing during extrusion. At the first pass through the extruder, the MLVs break up into

smaller MLVs and LUVs when large and small pores were used, respectively. The size of

both MLVs and LUVs are comparable to the size of the pores. For the size of smaller MLVs

is comparable to the size of the pores, they can squeeze easily through the pores without

rupturing at the second and more passes. This may suggest that larger extruded vesicles

tend to have more than one bilayer.


Pass 50 nm pores 100nm pores

1 bilayer (%) 2 bilayers (%) 1 bilayer (%) 2 bilayers (%)

1st 99 1 68 322nd 0 100 66 343rd 0 100 69 314th 0 100 69 316th 0 100 71 298th 0 100 72 2810th 0 100 69 31

Table 3.3: The fraction of unilamellar and bilamellar vesicles for 50 and 100 nm vesiclesmeasured at each pass through the extruder.

3.4 Summary

The results of these studies shows that although size of the extruded vesicles can vary

depending on the size of the pores used, only small vesicles have narrow size distribution

and are unilamellar. The polydispersity of the vesicles produced using larger pores is large,

although the polydispersity of the pores is small. The polydispersity of the vesicles, hence,

most likely due to the presence of small vesicles that produced spontaneously when lipid was

dispersed in water, or as a consequence of the process of extrusion. The multilamellarity of

extruded vesicles when produced using large pores, on the other hand, is presumably due

to the break-up of MLVs to the smaller MLVs that are comparable to the size of the pores.

The smaller MLVs then can squeeze easily through pores without rupture. The results also

reveal that larger extruded vesicles tend to have more than 1 bilayer.

Chapter 4

Effect of Sterols on the Mechanical

Strength of Lipid Membranes

4.1 Introduction

This chapter discusses the effect of sterols on the mechanical strength of lipid membranes.

The strength of membranes was probed by measuring the lysis tension of vesicles made from

mixtures of POPC and various sterols, including cholesterol, lanosterol and ergosterol.

All cells are surrounded by a membrane, plasma membrane, which encapsulates their

internal materials separating them from external chemical environment. Plasma membrane

is relatively permeable to water, but impermeable to large molecules and ions. Transport

of ions and large molecules from the outside into the cell requires a special channel or

transporter system. One of the essential function of the plasma membrane, therefore, is

54

CHAPTER 4. EFFECT OF STEROLS ON LIPID MEMBRANES 55

as a physical barrier that maintains cell integrity. In addition, many cellular enzymes,

protein molecules which function as biochemical catalysts, attach to the membrane and

many biological/biochemical processes occur only with the aid of these enzymes. Plasma

membrane, hence, also function as a substrate for enzyme assisted functional processes in

cells. Plasma membrane contains lipids including sterols such as cholesterol and proteins

with ratio varying depending on the source of the membranes.

A membrane lipid can be structurally characterized as having a polar domain that is

hydrophilic connected to a non-polar domain that is hydrophobic. There are enormous

diversity of the membrane lipids varying respect to the composition of polar and/or non-

polar domains. It is the hydrophobic and hydrophilic nature of the lipid molecules that

causes them to associate into different equilibrium shapes when exposed to an aqueous

environment. When lipid molecules are dispersed in water, they will organize to such a

shape that lowers the free energy of the system. Exposing a hydrophobic molecule to water

causes the rupture of hydrogen bonds between water molecules adjacent to the hydrophobic

molecule, thus increasing the energy. Alternatively, the hydrogen bonds can reorient forming

a hydrogen bonding network around the hydrophobic molecule. This causes a restriction

on hydrogen bonding network, thus decreasing the entropy. Either way, the free energy of

the system is increased. On the other hand, interaction between hydrophilic headgroups

and water molecules is stronger than self interaction among the hydrophilic headgroups and

among water molecules. Interaction between water molecules and hydrophilic headgroup,

thus, decreases the free energy of the system. For these reasons, when a lipid is dispersed in


Figure 4.1: Shapes of lipid molecules : a) a truncated cone, b) a cylinder and c) an invertedcone. ao is surface area occupied by a lipid molecule and v is volume of the hydrocarbonchain/chains.


water, the molecules will aggregates into such a shape to shield the hydrophobic tails from

and expose the hydrophilic headgroup to water to decrease the free energy of the system.

There are variety of structures such as : micelles, bilayer or inverted micelles which are

formed when lipid molecules are dispersed in aqueous environment. The structures formed

are dictated by the geometry of the lipid molecule which can be described as a truncated

cone, a cylinder, or an inverted cone as shown in Fig. 4.1. The structures form depend on

the value of volume of the hydrocarbon chain/chains v, surface area occupied by a lipid

molecule ao, and a maximum length the hydrocarbon chain can assume lc through shape

factor defined as vaolc

[30, 53]. The shape factors less than 13 and between 1

3 and 12 tend to

aggregate into spherical micelles and cylindrical micelles, respectively. The shape factors

from 12 to 1 and greater than 1 tend to form a bilayers and inverted micelles, respectively.

As an example, Fig. 4.2 shows a molecular structure of POPC, the main lipids use in

these studies. The polar domain that is neutral consists of a positively charged choline group

attaches to a negatively charged phosphate group. The non polar domain consists of two

hydrocarbon chains : one is 16 carbon fatty acid called palmitic acid, and the other is 18

carbon fatty acid called oleic acid with one double bond between carbons 9 and 10. For the

presence of one double bond in its hydrocarbon chains POPC is called a mono-unsaturated

lipid. Lipids with no double bond in hydrocarbon chains are called saturated lipids. The

POPC molecule favors the formation of bilayer when they are dispersed in water. Once the

bilayer is formed, it tends to close up into a vesicle to prevent the exposure of hydrocarbon

chains around the perimeter of the bilayer to water [54].


13

CH

O

N + O -

PO O

O

CO

O

C O

CH

2

16

2

4

6

8

10

12

14

3

5

7

9

11

13

15

2

4

6

8

3

5

7

11

12

13

14

15

16

1718

Figure 4.2: The molecular structure of POPC, the most abundant phospholipid in plasmamembrane. The polar domain that is neutral consists of a positively charged choline groupattaches to a negatively charged phosphate group. The non polar domain consists of twohydrocarbon chains : one is saturated-16 carbon fatty acid called palmitic acid, and theother is unsaturated-18 carbon fatty acid called oleic acid with one double bond betweencarbons 9 and 10.


Ener

gy

Area per molecule a

Attractive energy ~ a Repulsive Energy ~ C/a Total Energy ~ a + C/a

Minimum energya = ao

Figure 4.3: Energy as a function of area per molecule a. The equilibrium value of a, ao

occurs when dE/da = 0.


A bilayer has an area per molecule a which is set by forces originally due to lipid-lipid

and lipid-water interactions. The compressive force which tends to decrease a is the results

of hydrophobic interaction. This force can be represented by energy per molecule of the

form γa, where γ is the surface tension of water and lipid interface. This force is balanced

by steric repulsive force that tends to increase a. It is mainly due to the interaction between

hydrocarbon tails. The steric force is strong at short distance and falls of rapidly as a

increases and can be represented by potential in a form C/a, where C is a positive constant.

The energy per molecule from these two terms can be written as [30],

E = γa + C/a. (4.1)

E has minimum at the value of a at equilibrium, ao. C can be determined in terms of γ

and ao by setting dE/da = 0. Equation 4.1 can then be expressed as,

E = 2γao +γ

a(a − ao). (4.2)

Figure 4.3 shows the energy per molecule E as a function of a.

Vesicles are often used to study the properties of lipid membranes. Three properties

of the membrane that are relevant to these studies and have been studied extensively are

area expansion (compressibility) modulus Ka, lysis tension γlysis and bending rigidity κc.

When tension or compression is applied to the bilayer at equilibrium, the area per molecule

will deviate from the equilibrium value ao. Fig. 4.4.a shows an area dilation under the

influence of isotropic tension τ . The change in area ∆a is characterized by an isothermal


area expansion (compressibility) modulus Ka, defined as

Ka =τ

α, (4.3)

where α is the fractional change in membrane area equal to ∆aao

produced by isotropic

membrane tension τ .

If tension applied is larger than tension bilayer can withstand in keeping the lipid

molecules together, the vesicle will rupture. The tension require to rupture or lyse the

vesicle called the lysis tension γlysis. It is a measure of the mechanical strength of the

bilayer. Vesicle bilayer ruptures when ∆a is in the range of 2 - 5 % [53].

In order to force a flat membrane to a curved shape membrane, a stress associated with

curvature must be applied. Fig. 4.4.b shows the change in membrane curvature due to

the bending moment M (torque per unit length). This change in membrane curvature is

characterized by the bending rigidity kc of a bilayer,

M = kc∆c = kc∆(1

R1+

1

R2), (4.4)

where ∆c = c−c0 is the change in membrane curvature, c0 is the spontaneous curvature and

R1 and R2 are the principal radii of curvature at a given point on the membrane surface.

Studies show that Ka, γlysis and kc are affected by sterols and in particular knowledge of

sterol effect on these membrane properties comes from cholesterol. For example, cholesterol

increases the membrane lysis tension [3, 55], area expansion modulus [56], and bending

rigidity [57, 58]. In addition, cholesterol also changes the phase behavior of membranes that

seems to be universal for membranes made from both saturated lipids such as DPPC [59] and


τ

τ

τ

τ (a)

M

MM

M

(b)

Figure 4.4: Membrane shape changes as a response to mechanical forces. a. Isotropic areadilation due to isotropic tension τ , b. bending or curvature change due to bending momentM .


Figure 4.5: The molecular structure of cholesterol, lanosterol and ergosterol.


mono-unsaturated lipids such as SEPC [60], PPetPC and POPC [61]. Cholesterol induces

the acyl chain order, while breaks up the lateral packing of the liquid crystalline state-

lipid matrix, consequently increasing lipid lateral mobility. The more ordered acyl chains

resemble the gel state, while the lipid lateral mobility associates with the liquid crystalline

state. Cholesterol, therefore, induces a new state of lipid membrane, a liquid order state.

Lipid membrane in this new state, thus, has the mechanical properties of the gel state

and the lateral mobility of the liquid crystalline state. Both properties are important for

functioning of the biology membranes.

Cholesterol is a major sterol in plasma membranes of mammalian cells [62]. Other

sterols of interest are lanosterol and ergosterol. Lanosterol predominates in some procaryotic

membranes, while ergosterol is abundant in plasma membranes of lower eukaryotes such as

certain protozoa, yeast, fungi and insects such as Drosophila [63]. The molecular structure

of cholesterol, lanosterol and ergosterol is shown in Fig. 4.5. The main structure of these

sterols is quite similar. Cholesterol consists of a fused tetracyclic ring, a hydroxyl covalently

attached to the first ring, and an extended hydrocarbon side chain attached to the fourth

ring. Lanosterol differs from cholesterol as it has three additional methyl groups and a

double bond on its side chain. On the other hand, ergosterol differs from cholesterol by a

methyl group on its side chain and two extra double bonds, one on its side chain and the

other on the second ring.

In recent years, the effect of lanosterol and ergosterol in comparison to cholesterol on

lipid membranes has become a focus of many studies [64, 65, 66, 67, 68, 69, 70]. Ergosterol


has been shown to be more effective than cholesterol in inducing acyl chain ordering in

membrane of DPPC, a saturated lipid, but less effective in breaking-up the lateral packing

in the liquid crystalline state of these membranes [68]. In DPPC membranes with 40 mol%

sterols, all sterols increase the area expansion modulus Ka, where ergosterol shows the

largest effect and lanosterol the smallest [64]. The same studies also show limited change

in the lysis tension of the membrane with the introduction of sterols. Lanosterol, on the

other hand, is less effective than cholesterol in inducing the acyl chain order of membrane

of DMPC, a saturated lipid [66]. The ordering effect, however, is different when mono-

unsaturated lipids are used. Henriksen et al. show that cholesterol is most influential

followed by lanosterol, while ergosterol the least influential in inducing acyl chain ordering

in POPC vesicles [70]. They show that the area expansion modulus Ka as well as the

bending rigidity κc increase with the introduction of sterols, where cholesterol was found to

result in the most significant increase in both values, followed by lanosterol and ergosterol

[67, 70]. Furthermore, the studies show universal behavior amongst Ka, κc and the order

parameter M1 when sterols are introduced into POPC membranes.

Studies of the mechanical strength of lipid membranes as a function of lanosterol and

ergosterol content, however, is lacking, especially for unsaturated lipid membranes. This

is in spite of the fact that the mechanical strength of the membrane is one of the most

important factors in studies of both model cells and carriers in wide range of application. The

mechanical strength of membrane is related to cell viability which depends on maintaining

an intact plasma membrane. As carriers, in drug delivery for example, the mechanical


strength is related to a resistance to stress induced by lytic agents in the host.

This chapter describes experiments designed to observe the effect of cholesterol, lanos-

terol and ergosterol on the mechanical strength of lipid membranes. The membrane strength

was probed by measuring the lysis tension of vesicles made from mixtures of POPC and

various sterols. The lysis tension was determined from the minimum pressure required to

extrude vesicles through small pores. All sterols increase the lysis tension of POPC mem-

branes, where cholesterol shows the largest effect followed by lanosterol and ergosterol.


4.2.1 Vesicle Preparation

Materials used in these experiments were POPC, cholesterol, ergosterol and lanosterol.

POPC and cholesterol (98% purity) were purchased from Avanti Polar Lipids, Inc., while

ergostrol (98% purity) and lanosterol (97% purity) were purchased from Sigma-Aldrich Co.

Samples of POPC:sterols were prepared by mixing POPC and each sterol in powder form

at the required concentrations. The mixed powder was then dissolved in a benzene:methanol

mixture (4:1). The solution was gently shaken until all powder disappeared. Finally, the

solution was lyopholized, i.e., it was frozen in liquid nitrogen and then evaporated under

vacuum for about 6 hours.

The powder remaining after lyophilization was hydrated using purified water from a

Milli-Q water purification system (Millipore, Bedford MA) in the ratio of 1 mg of lipid

per 1 ml of water. The suspensions were then taken through an FTV process before being


extruded through 400 nm pores (pre-extrusion). The pre-extruded vesicles then were diluted

to a concentration of 0.1 mg/ml and were used for experiments to determine lysis tension.

4.2.2 Lysis Tension Determination

Determination of lysis tension of vesicles follows methods explained in the previous studies

[3] where the lysis tension is related to the minimum pressure required to burst or lyse a

vesicle at the entrance of a pore. Figure 4.6 show a diagram of a vesicle at the entrance of

a pore. Rp and Rv are radii of pore and vesicle, respectively, while P1 is pressure applied,

P2 is pressure inside the vesicle and P0 is the atmosphere pressure. To enter the pore, a

large vesicle must reduce its volume by bursting. As a large vesicle of radius Rv is pulled

into a pore of radius Rp, a surface tension γ is induced in the membrane. The vesicle bursts

when the induced surface tension is greater then the lysis tension of the bilayer. A relation

between applied pressure P , γ, Rp and Rv can be derived by applying the Laplace relation

between pressure and curvature P = 2γH [71], where H and γ are the mean curvature

and surface tension, respectively, to small and large fragments of vesicles in Fig. 4.6. The

Laplace pressure for the large fragment is given by,

P2 − P1 =2γ

Rv, (4.5)

and for the small fragment is given by,

P2 − P0 =2γ

Rp. (4.6)

Substracting Eq. 4.5 from Eq. 4.6 yields

P = 2γ(1

Rp− 1

Rv), (4.7)


Polycarbonatemembrane filter

Polycarbonatemembrane filter

0P

2P

vR

pR

1P

Figure 4.6: Schematic diagram of a vesicle at the entrance of a pore.


where the pressure difference P = P1 − P0. The lysis tension γlysis can be estimated when

the values of the minimum pressure required to rupture the vesicles, the radii of the pores

and of the vesicles are available.

The radii of pores and vesicles were determined using SEM and DLS, respectively. The

detail of DLS and SEM measurements have been described in Chapters 2 and 3, respectively.

The minimum pressure was determined from the flowrate of vesicle suspension measure-

ments. When vesicle suspensions are pushed through pores with a pressure larger than Pmin,

the flowrate of the suspension is proportional to the pressure applied following Darcy’s law

for flow through Np cylindrical tubes,

Q =NpπR4

p(P − Pmin)

8ηL, (4.8)

where η is the viscosity of the solvent, Rp and L are the radius and the length of the

cylindrical pore, respectively. Applying pressures smaller than Pmin results in no flow of

suspension through the pores. In this case, the pressure is not enough to cause tension

sufficient to rupture the vesicles. The vesicles then just block the pores. The minimum

pressure, then, can be defined as a pressure at and below which the flowrate of the vesicle

suspension is zero. Since the flowrate is proportional to the pressure, the minimum pressure

can be determined by measuring the flowrate at several different pressures and fitting a line

to the flowrate pressure data to extrapolate to zero flowrate pressure, which is the minimum

pressure.

The flowrate of the suspension was measured by taking an amount of solution in a certain

interval time and weighing the solution. This was achieved using a flowmeter.


Sterol % Sterol Mean Radius (nm)

0 195 ± 1.3

Cholesterol 10 193 ± 1.220 211 ± 2.030 226 ± 1.6

Lanosterol 10 215 ± 1.520 187 ± 1.530 157 ± 1.7

Ergosterol 10 228 ± 2.820 199 ± 2.230 177 ± 0.7

Table 4.1: The mean radius of pre-extruded vesicles. The uncertainties represent the stan-dard deviations from 5 measurements.


Lysis tension of vesicles made from mixtures of POPC:cholesterol, POPC:lanosterol, and

POPC:ergosterol was determined. Equation 4.7 was used for lysis tension calculation, where

γ and P were replaced by γlysis and Pmin, respectively,

γlysis =Pmin

2

(

Rp Rv

Rv − Rp

)

. (4.9)

In order to calculate the lysis tension of vesicles, the radii of pre-extruded vesicles Rv and the

pores Rp, and the minimum pressure required to rupture vesicles Pmin need to be measured.

The radius of pre-extruded vesicles was measured using DLS. Equation 2.3 was fit to

the autocorrelation function data where Eq. 2.6 was used as the field-field autocorrelation

function. The mean radius of pre-extruded vesicles were determined from the fits and the

results from samples of POPC, POPC:cholesteol, POPC:lanosterol and POPC:ergosterol

with different concentrations of sterols are summarized in Table 4.1.


25 30 35 40 450

20

40

60

80

100

120

Num

ber o

f Por

es

R (nm)

Data Fit

Figure 4.7: Histogram of pore radius data for PCTE with nominal pore radius of 50 nm.


0.00

0.25

0.50

0.75

1.00

0.00

0.25

0.50

0.75

1.00

0 25 50 75 100 1250.00

0.25

0.50

0.75

1.00

a. Cholesterol

Water POPC POPC + 10 % sterol POPC + 20 % sterol POPC + 30 % sterol

b. Lanosterol

Q(c

P m

l/s)

c. Ergosterol

P (psi)

Figure 4.8: Viscosity corrected flowrate as a function of pressure for POPC:sterol with theconcentration of sterols varying from 10 to 30 mol%. A linear function was fit to the datato determine the zero flowrate pressure defined as the minimum pressure Pmin. The datafor water is shown for comparison. The lines are the linear fits.


Sterol % Sterol Pmin (psi)

0 56.2 ± 5

Cholesterol 10 69.3 ± 3.620 78.5 ± 5.330 88.0 ± 4.9

Lanosterol 10 67.1 ± 2.920 71.7 ± 2.930 76.0 ± 4.1

Ergosterol 10 63.4 ± 2.020 67.3 ± 6.230 67.3 ± 6.3

Table 4.2: The minimum pressure for vesicle solution made from the mixture ofPOPC:sterols. The values were found from the linear fitting procedure as shown in Fig.4.8.

The radius of the pores was measured using SEM. There are 471 pores analyzed from

SEM micrographs. The radius data was displayed in a histogram and a Gaussian distribution

was fit to the data to determine the mean radius of the pores. Figure 4.7 show the histogram

and the Gaussian fit. The mean radius resulting from the fits is 33.5 ± 8.8 nm.

The minimum pressure Pmin for each sample was determined from the flowrate measure-

ments of vesicle suspension at different pressures. The data and the linear fits are shown in

Fig. 4.8 and the results for Pmin are summarized in Table 4.2.

The lysis tension of POPC vesicles with different concentration of sterols was then cal-

culated using Eq. 4.9 and the results are shown in Fig. 4.9. The figure shows that all sterols

increase the lysis tension of POPC membrane, where cholesterol shows the largest effect,

followed by lanosterol and ergosterol. The increase in lysis tension due to cholesterol and

lanosterol is almost linear while the increase due to ergosterol levels-off above 20 mole%.

The increase in lysis tension of POPC membrane by cholesterol is consistent with results


0 10 20 30

8

10

12

Lysi

s Te

nsio

n (m

N/m

)

Sterol concentration (mole %)

Cholesterol Lanosterol Ergosterol

Figure 4.9: The lysis tension of POPC vesicles as a function of sterol concentration.


45 50 55 60 65 70 75

8

9

10

11

12

200 220 240 260 280 300 320 340 360

8

9

10

11

12

30 40 50 60 70 80 90

8

9

10

11

12

M1 (103s-1)

a.

b.

Lysi

s te

nsio

n (m

N/m

)Ly

sis

tens

ion

(mN

/m)

Ka (mN/m)

Cholesterol Lanosterol Ergosterol

c.

Ly

sis

tens

ion

(mN

/m)

c (KbT)

Figure 4.10: The lysis tension of membrane as a function of a) order parameter M1, b)bending rigidity κc and c) area expansion modulus Ka. Values of M1, κc and Ka are takenfrom Table 1 of ref. [70].


from previous studies on POPC [3] and other unsaturated lipids such as SOPC [55]. How-

ever, there is less information for lysis tension of unsaturated lipid membranes in the pres-

ence of lanosterol and ergosterol. Nevertheless, there are results on the effect of cholesterol,

lanosterol and ergosterol on other properties of POPC membrane including order parame-

ter M1, bending rigidity κ and area expansion modulus Ka [67, 70]. Sterols increases order

parameter M1, bending rigidity κ and area expansion modulus Ka of the POPC membranes

following the sequence cholesterol > lanosterol > ergosterol [67, 70]. This is correlated to

the increase in the lysis tension that also follows the sequence cholesterol > lanosterol >

ergosterol. The correlation between lysis tension γlysis of POPC membranes from these

studies with order parameter M1, bending rigidity κ and area expansion modulus Ka of

POPC membrane can be shown by plotting γlysis as a function of M1, κ and Ka. Fig-

ure 4.10 shows the plot of γlysis as a function of M1, κ and Ka, where values of M1, κ and

Ka are taken from Table 1 of ref. [70]. The data collapse nicely for all graphs indicating

their strong relations.

The relation between lysis tension and order parameter shown in Fig. 4.10.a is almost

linear. The linear relation between lysis tension and order parameter indicates that lipid

chain order determines membrane mechanical strength. Since increase in the chain order

attributes to increase in the membrane thickness [75], the linear relation between lysis

tension and order parameter suggests that lysis tension is related to the thickness of the

membranes. In the process of membrane rupture there is competition between the applied

tension that forms pores and the edge energy associated with the pore formation [76]. The


edge energy is the material property which is related to the strength of the membranes. The

increase in the thickness of the membrane as a consequence of the presence of sterols should

make pore formation more difficult because of increased edge energy.

The relation between lysis tension and bending rigidity shown in Fig. 4.10.b is also almost

linear. Linear relation between lysis tension and both order parameter and bending rigidity

is reasonable since bending rigidity also shows a linear dependence on order parameter [70].

The relation between lysis tension and area expansion modulus from Fig. 4.10.c is not

as simple as lysis tension-order parameter and lysis tension-bending rigidity relations. From

Eq. 4.3 and the fact that bilayer ruptures when ∆a is in the range of 2 - 5 % [53], the linear

relationship between lysis tension and area expansion modulus is expected. This is not the

case as shown in Fig. 4.10.c. The Ka-γlysis relation for POPC:sterols can be compared to

Ka-γlysis relation for SOPC:sterol measured in the previous studies [55]. Figure 4.11 shows

the plot of lysis tension as a function of area expansion modulus for SOPC-cholesterol system

where the data is taken from Table 2 in ref. [55]. The figure shows the increase of lysis tension

with the area expansion modulus up to a plateau after reaching Ka of 300 mN/m and at

Ka larger than 1000 mN/m. The plateau after reaching Ka of 300 mN/m is comparable

to the data in Fig. 4.10.c showing a change in slope after reaching Ka of 300 mN/m. Both

SOPC:cholesterol and POPC:sterols data show a linear relation between lysis tension and

area expansion modulus for the value of Ka below 300 mN/m with exception for pure POPC

data where the value of Ka is the lowest.

Using the values of τlysis and Ka from Fig. 4.10.c, ∆a/ao at lysis can be predicted using


200 400 600 800 1000 1200

5

10

15

20

25

30

lysi

s (mN

/m)

Ka (mN/m)

Figure 4.11: The lysis tension of SOPC membrane as a function of area expansion modulus.Both values are taken from Tabel.2 in ref. [55].


Ka (mN/m) τlysis (mN/m) ∆aao

213 7.84 0.037216 8.6 0.040230 9.2 0.040233 9.4 0.040240 9.5 0.039241 9.6 0.040260 10.1 0.039276 10.8 0.039281 11.2 0.040354 11.9 0.034

Table 4.3: Estimation of ∆a/ao at lysis.

Eq. 4.3 and the results are shown in Table 4.3. The value of ∆a/ao at lysis is 3.7 % for pure

POPC and 4 % average in the presence of sterols with the exception of 3.4 % for the largest

Ka. The increase in ∆a/ao with the presence of sterols is consistent with results shown by

Needham and Nunn with lipid in the presence of cholesterol[55]. The results in Table 4.3

(with the exception of last data point) show that the value of ∆a/ao at lysis is almost the

same for all sterols.

4.4 Summary

These studies show that cholesterol, lanosterol and ergosterol increase the mechanical strength

of POPC membranes as indicated by an increase in the membrane lysis tension. Despite

their similarity in molecular structure these sterols impact the lysis tension of the lipid

membrane to varying extent. Cholesterol exhibits the largest effect followed by lanosterol


and ergosterol. The lysis tension of POPC membrane increases almost linearly as a func-

tion of sterol content for cholesterol and lanosterol up to 30 mole% sterols. However, the

lysis tension levels-off after 20 mole% for POPC:ergosterol system. The increase in the lysis

tension of POPC membranes by sterols is correlated to the increase in the chain order of

the lipids by sterols.

Chapter 5

Vesicle Aggregation Induced by

cytidine

5’-triphosphate : phosphocholine

cytidylyltransferase

5.1 Introduction

Cytidine 5’-triphosphate(CTP):phosphocholine cytidylyltransferase (CCT) is a key regu-

latory enzyme in the synthesis of phosphatidylcholine (PC). It catalyzes the formation of

cytidine 5’-diphosphate choline, the head-group donor in the synthesis of PC. This enzyme is

81

CHAPTER 5. VESICLES AGGREGATION INDUCED BY CCT PROTEIN 82

amphitropic : it interconverts between a soluble form where it is inactive and a membrane-

bound form where it is active. The equilibrium between soluble and membrane-bound

forms is determined by lipid composition of the membrane and by the phosphorylation

state of the enzyme. A membrane containing anionic lipids, for example, stabilizes the ac-

tive membrane-bound form of the enzyme [77, 78, 79], while phosphorylation on the enzyme

favors its inactive soluble form [77]. CCT is important to compensate the degradation of PC

due to the action of phospholipases, enzymes that catalyzes the hydrolysis of phospholipid.

When inactive CCT senses a decrease in the amount of PC in a membrane, it will bind

to the deficient PC membranes and become active [80]. Studies show that CCT responds

quickly to PC catabolism [81, 82, 83]. In cells, CCT is distributed in nucleus where it is not

or less active, and endoplasmic reticulum (ER) where it is active [84, 85]. The distribution

of CCT between nucleus and ER is in response to the stimulation the PC synthesis.

CCT has a dimeric structure [86]. Each monomer consists of 367 residues that can

be classified into four functional domains : domains N, C, M and P. Domain N (residues

1-72) consists of cluster of positively charged amino acids that specify nuclear localization

[87]. The catalytic domain, domain C (residues 73-236) [88], together with domain N, is

involved in the dimerization [89]. The membrane binding domain, domain M (residues 237-

300), contains three consecutive repetitions of 11 residues (residues 256-288) that form a

random coil in the absence of lipid vesicles and an amphipathic α helix in the presence of

anionic lipid vesicles [90]. Domain P (residues 301-367) is rich in serine-proline and is highly

phosphorylated [91].


CCT binds to and becomes activated by membranes containing anionic lipids [77, 78, 79],

diacylglycerols (DAG) and phosphatidylethanolamines (PE) [92, 93, 94]. These activating

lipids can be categorized into 2 class lipids. For membranes made up anionic lipids, the

binding of CCT to the membrane occurs by electrostatic interaction between the positively

charged binding site of CCT and the negatively charged surface of the membrane [78, 79, 95].

These lipids are known as class I lipids. For membrane composed DAG and PE, the binding

of CCT is postulated to occur because CCT relieves the negative curvature strain which

exists in these membranes [93, 94]. These lipids are known as class II lipids. Long chain fatty

acids, which are charged and promote formation of the membrane with negative curvature

also promote binding and activation of CCT. CCT poorly binds to and shows very weak

activity in membranes containing Egg PC [96, 77, 97]. Egg PC, therefore, is often categorized

as a non-activating lipid.

Experiments on anionic lipid membranes show that there are 2 steps for CCT activation

: binding to the membrane and intercalation of the binding site into the hydrophobic area

of the membrane [95]. Although CCT binds to membranes in both liquid crystalline and gel

phases, it is active only in a liquid crystalline phase membrane. This suggests that additional

interaction other than binding is required for CCT activation, which can be achieved only

in liquid crystalline phase membrane. These results have lead to a model of activation

CCT where domain M is the inhibitor of the catalytic domain. The inhibition is relieved

when domain M intercalates into the hydrophobic region of the membrane [80]. This model

is supported by the finding showing that CCT activity is proportional to the amount of


activating lipids present, while fragment of CCT that lacks the membrane binding domain,

CCT-236 is always active independent of lipid concentration [98].

The binding affinity of CCT and the activity of CCT depends on the lipid composition

of the membrane. The binding affinity for class I lipids is larger than that for fatty acids or

class II lipids by up to a factor of ten [77, 78, 93, 94]. Class I lipids are also more potent

than class II lipids and long chain fatty acids in activating CCT [78, 97]. This suggests that

CCT-membrane binding affinity and CCT activity are correlated.

The binding of CCT to the membrane of a single vesicle is thought to occur by binding

of both domain Ms of the CCT dimer as shown in Fig. 5.1a. However, it is also possible that

each domain M can bind to the surface of different vesicles, cross bridging the two vesicles

as shown in Fig. 5.1b. The cross-bridging mode will result in the aggregation of vesicles. It

has been observed that addition of CCT to small unilamellar anionic vesicles increases the

turbidity of the samples, suggesting the occurrence of aggregation [99].

This chapter contains studies on the effect of CCT on the aggregation of large unilamellar

vesicles. DLS was used to explore the possible aggregation of vesicles made from non-

activating and activating lipids, where activating lipids include class I and class II lipids.

There was no aggregation observed for non-activating lipid vesicles after addition of CCT

for all concentration of CCT used in these experiments. In contrast, aggregation occurred

when CCT was introduced to samples of class I lipid vesicles. The required concentration

of CCT monomer to stimulate the aggregation is as small as 3 per vesicle. This suggests

that CCT cross-bridging vesicles is a mechanism for the aggregation. In comparison, there


a.

b.

M M

M

M

PP

P

P

N/CN/C

N/C N/C

Figure 5.1: Two modes of CCT binding to membrane : a) Domain M of each monomer bindsto the membrane of a single vesicle, b) CCT cross-bridges two vesicles through binding ofdomain M of each monomer to the membrane of two different vesicles.


was no indication of aggregation when CCT was introduced to samples of class II lipid

vesicles. The results suggest that binding of CCT to class I lipid membranes was dominated

by cross-bridging mode, while to class II lipid membranes either two domain Ms bind on

the same vesicle, or only one monomer binds while the other is unbound.


5.2.1 Materials

The lipids used in these experiments were L-α Phosphatidylcholine (Egg PC), L-α Phos-

phatidylglycerol (Egg PG),1,2-Dioleoyl-sn-Glycero-3-Phosphoethanolamine (DOPE) and 1-

Palmitoyl-2-Oleoyl-sn-Glycero-3-Phospho-L-Serine (POPS) and oleic acid (OA). Egg PC

and Egg PG were purchased from Northern Lipids Inc. (Vancouver, BC), OA from Sigma,

while DOPE and POPS were from Avanti Polar Lipids Inc. (Alabaster, AL). Rat liver

CCTα and CCTα236 (encoding amino acids 1-236) were expressed by baculovirus infection

of Trichoplusia ni cells and were purified as described previously [99]. The pure enzymes

were stored in aliquots at -80◦ C and their concentrations were determined by quantitative

amino acid analysis (Alberta Peptide Institute, Edmonton, AB). A 57-mer peptide (residues

237-293 of rat liver CCTα) was synthesized by K. Piotrowska at the University of British

Columbia Peptide Service Laboratory (Vancouver, BC) as described previously [99]. The

CCTα, CCTα236 and peptide were prepared by Dr. Svetla Taneva from Prof. Rosemary

Cornell’s Lab (Dept. of Molecular Biology and Biochemistry SFU).


5.2.2 Vesicle Preparation

Lipids in powder form were dissolved in chloroform. Aliquots of lipid or lipid mixture

solutions were put in a round bottom flask before drying by rotary evaporation. The dried

lipid films were then placed under vacuum overnight to remove traces of residual solvent.

The lipid films were dispersed in a mixture of 10 mM Tris, pH 7.4, 1 mM ethylen-

diaminetetra-acetic acid (EDTA), 150 mM NaCl, and 10 mM dithiothreito (DTT) to a

concentration of 4 mM. The suspension was taken through a freeze-thaw-vortex process as

explained in the previous chapters. The lipid suspension was then diluted to concentration of

0.1 mM and extruded through 2 polycarbonate membrane filters with nominal pore radius

of 25 nm held in an extruder (Lipex Biomembrane, Vancouver BC). The extrusion was

repeated 14 times at a pressure of 200 psi.

5.2.3 Vesicle Size Measurements

The mean radius and the polydispersity of the vesicles before and after introduction of CCT

were measured using DLS. The autocorrelation function measurement was repeated 5 times

at a scattering angle of 90◦. Equation 2.3 was fit to the intensity-intensity autocorrelation

function using Eq. 2.10 for the field-field autocorrelation function. The intensity-weighted

mean radius Ri and polydispersity σi were determined from the fitting procedure. The

intensity-weighted distribution was determined instead of the number-weighted distribution

because the aggregated structures within the samples are very polydisperse. As explained

in Chapter 2, the use of number-weighted distributions for polydisperse samples is limited.


The use of intensity-weighted distribution is, however, still acceptable for the purpose of

these studies as it allows the observation of the change of vesicle radius and polydispersity.

The radius distribution is assumed to follow a Schulz distribution as described by Eq. 2.2.

For DLS samples, CCT from 5.5 µM stock solution was introduced to a 0.1 mM vesicle

suspension to produce CCT concentration in the range of 1-100 nM. The ratio of CCT

monomers to the vesicles was determined from concentration of CCT, concentration of

lipids and number of lipid molecules per vesicle. The number of lipid molecules per vesicle

was determined using an average vesicle radius as determined from these experiments, a

bilayer thickness of 4 nm [100], and a headgroup area of 0.68 nm2 for Egg PG, Egg PC

and POPS [101] and 0.28 nm2 for OA [102]. The lipid concentration was determined by the

method of Bartlett [103].

To confirm the results from DLS measurements, negatively stained transmission electron

microscopy (TEM) was conducted on some samples. 8 µL of samples of vesicles with and

without CCT were pipetted onto copper formfar coated grids (300 mesh). After removing the

excess material by blotting with filter paper, the specimen grid was stained with 8 µL of 2 %

aqueous uranyl acetate for a few seconds. The excess stain was then removed by filter paper.

The micrographs were taken from at least 2 grids for each samples. The microscope used was

a Hitachi H-760 transmission electron microscope (Hitachi, Tokyo, Japan) located at Bio-

imaging facility, The University of British Columbia. The measurements were conducted

by Dr. Svetla Taneva from Prof. Rosemary Cornell’s Lab (Dept. of Molecular Biology and

Biochemistry SFU).


5.3 Results

5.3.1 Effect of CCT on Activating and Non-Activating Lipid Vesicles

The effect of CCT on the aggregation state of activating and non-activating lipid vesicles was

studied by observing the change in their size and polydispersity using DLS before and after

introduction of CCT. Egg PG and Egg PC were chosen for activating and non-activating

lipids, respectively, as Egg PG membrane is very strong activator of CCT, while Egg PC

membrane is very weak in activating CCT.

Figure 5.2 shows intensity-intensity autocorrelation function data for vesicles made from

a) Egg PC and b) Egg PG, before and after introduction of CCT protein. There is no

difference in decay time of the intensity-intensity autocorrelation function when CCT was

introduced to Egg PC vesicles, although the ratio of CCT per vesicle was as high as 85.

In contrast, an increase in the decay time is observed, when CCT was introduced to Egg

PG vesicles with a ratio as small as 3 CCT per vesicle. As explained in Chapter 2, DLS

distinguishes small and large particles through the decay time of intensity-intensity autocor-

relation function : the larger the particles the longer the decay time. These results suggest

that CCT protein induces aggregation of Egg PG vesicles, but not Egg PC vesicles. The fig-

ure also shows that DLS is a sensitive tool for detecting a difference between un-aggregated

and aggregated vesicle states through the difference in decay time of the intensity-intensity

autocorrelation function : a change in decay time is detectable for as low as 3 CCT per

vesicle.

The mean radius and polydispersity of vesicle samples before and after introduction of


1E-4 1E-3 0.01 0.1 1 10

0.0

0.1

0.2

0.3

0.4

0.0

0.1

0.2

0.3

0.4

(ms)

g2 - B

asel

ine

0 CCT/vesicle 3 CCT/vesicle 10 CCT/vesicle 15 CCT/vesicle 21 CCT/vesicle 30 CCT/vesicle 40 CCT/vesicle 48 CCT/vesicle

b. Egg PG + CCT

a. Egg PC + CCT

0 CCT/vesicle 5 CCT/vesicle 21 CCT/vesicle 42 CCT/vesicle 85 CCT/vesicle

g2 - B

asel

ine

Figure 5.2: Intensity-intensity correlation function data from DLS measurements of vesiclesmade from a) Egg PC and b) Egg PG before and after introduction of CCT protein.


40

60

80

0 20 40 60 800.00

0.25

0.50

0.75

R (n

m)

Number of CCT monomer/vesicle

Egg PC Egg PG

b.

a.

Figure 5.3: a) The mean radius and b) the polydispersity of vesicles made from Egg PCand Egg PG before and after introduction of CCT protein.


CCT were determined by fitting Eq. 2.3 to the intensity-intensity autocorrelation function

data, where Eq. 2.10 was used as the the field-field autocorrelation function. Figure 5.3

shows a) the mean radius and b) the polydispersity as a function CCT concentration for

Egg PC and Egg PG vesicles. As can be predicted from Fig. 5.2, CCT does not induce

aggregation of vesicles made from EggPC, a weak CCT activator. However, CCT induces

aggregation for vesicles made from Egg PG, a strong CCT activator.

When CCT was introduced to the vesicle suspensions, there is a possibility of an initial

high local concentration of CCT. This CCT dispersion problem would cause the binding

of CCT to only few vesicles in the area where CCT was concentrated. To check for this

possibility, DLS measurements were conducted on the system of CCT-vesicles prepared by

two different order of introducing CCT; the first one is by adding CCT on dilute vesicle

suspension as explained in the methods and the second one by adding concentrated vesicle

suspension to a dilute CCT solution. The second procedure ensures the dispersion of CCT

in vesicle suspension. The results of the measurements show no difference in the mean radius

and polydispersity of CCT-vesicles prepared by these two procedures. Therefore, there is

no issue of CCT dispersion problems in the experiments.

To confirm the results from DLS measurements, effect of CCT on aggregation state of

Egg PG vesicles was also observed using TEM. Figure 5.4 shows a typical TEM micro-

graph of a) pure Egg PG vesicles and b) Egg PG vesicles plus CCT with concentration of

6 CCT/vesicle. The mean radius and polydispersity of this particular sample were deter-

mined by analyzing at least 300 vesicles. Figure 5.5 shows the radius data whose micrograph


Figure 5.4: A typical TEM micrographs of a) pure Egg PG vesicles and b) Egg PG vesiclesplus CCT with concentration of 6 CCT/vesicle.


0

5

10

15

20

a. Egg PG

Num

ber o

f ves

icle

s (%

)

20 40 60 80 100 120 140 160 180 300 400 500 600 7000

5

10

15

20

R (nm)

b. Egg PG + CCT (6 monomer/vesicle)

Figure 5.5: Histogram of vesicle radius data for a) Egg PG and b) Egg PG plus CCT withconcentration of 6 CCT/vesicle.


representatives are shown in Fig. 5.4, arranged in a histogram. The mean radius of Egg

PG increases from 46.0 ± 24.0 nm for pure Egg PG to 56.0 ± 28.0 nm when CCT with

concentration of 6 CCT/vesicle was introduced. The values are comparable to the mean

radius from DLS which increases from 43.6 ± 10.1 nm for pure Egg PG to 50.1 ± 17.4 nm

when CCT with concentration of 7 CCT/vesicle was introduced. The uncertainties in the

mean radius values are the standard deviations of the distributions. Results of TEM mea-

surements for other concentrations of CCT shows that the mean radius increases with CCT

concentration, which is consistent with those from DLS.

To examine the amount of Egg PG in Egg PG:Egg PC membranes required for vesicle

aggregation to occur, DLS was used to monitor the radius and polydispersity of vesicles made

from the mixtures of Egg PC and Egg PG in the presence of CCT. The concentrations of

Egg PG used for these experiments were 15, 25, 33, 50 and 100 mol %. Figure 5.6 shows

the plot of a) the mean radius and b) the polydispersity of vesicles made from Egg PC:Egg

PG mixtures before and after introduction of CCT. The figure shows that no aggregation

occurs in mixtures containing 15 and 25 mol % of Egg PG. As indicated by an increase in

vesicle size and polydispersity, aggregation starts to occur when the concentration of Egg

PG reaches 33 mol %.

DLS experiments are also conducted to observe the ability of parts of CCT including

CCT-236 (CCT without domain M) and 57-mer peptide (only domain M) in inducing vesi-

cles aggregation. Figure 5.7 shows the mean radius and the polydispersity of Egg PG vesicles

before and after introduction of CT-236 or peptide. Similar data for CCT from Fig. 5.8 is


40

50

60

70

80

0 50 100 150 2000.15

0.30

0.45

0.60

0.75

15 mol% Egg PG 25 mol% Egg PG 33 mol% Egg PG 50 mol% Egg PG 100 mol% Egg PG

R (n

m)

a.


b.

Figure 5.6: a) The mean radius and b) the polydispersity of Egg PC:Egg PG mixture beforeand after introduction of CCT.


45

60

75

0 15 30 45

0.25

0.50

0.75

b.

CCT CCT-236 Peptide

R(n

m)

a.

Number of protein or peptide monomer/vesicle

Figure 5.7: a). The mean radius and b) the polydispersity of Egg PG vesicles before andafter introduction of CCT or CCT-236 or peptide.


shown for comparison . The figure shows that parts of CCT, CCT-236 and 57-mer peptide,

do not cause vesicle aggregation. Therefore, the whole structure of CCT is required for

aggregation of Egg PG vesicles.

5.3.2 Effect of CCT on Class I and Class II Lipid Vesicles

The effect of CCT on the aggregation state of class I and class II lipid vesicles was studied by

observing the change in their size and polydispersity using DLS before and after introduction

of CCT. POPS, in addition to Egg PG represents class I lipids, while EggPC:DOPE (40:60)

and EggPC:DAG (85:15) represent class II lipids. In addition, experiments were also done on

vesicles made from Egg PC:OA (50:50) which features both class I and II lipids. Figure 5.8

shows a) the mean radius and b) the polydispersity of vesicles made from Egg PG, POPS,

EggPC:DOPE, EggPC:DAG and Egg PC:OA as a function of CCT concentration. The

figure shows that CCT induces aggregation of vesicles made from class I lipids, including

Egg PG and POPS. However, the aggregation did not take place when CCT was introduced

to vesicles made from Egg PC:DAG, Egg PC:DOPE and Egg PC:OA. Nevertheless, there

was a marginal increase in the radius of Egg PC:OA of 10 % when CCT with concentration

more than 80 CCT per vesicle was added.


40

60

80

100

0 20 40 60 800.00

0.25

0.50

0.75

1.00

R (n

m)

Egg PG POPS Egg PC:DOPE Egg PC:DAG Egg PC:OA


b.

a.

Figure 5.8: a) The mean radius and b) the polydispersity of vesicles made from Egg PG andPOPS (class I CCT activators), Egg PC:DAG (85:15) and Egg PC:DOPE (40:60) (class IICCT activators), Egg PC:OA (a special CCT activator that features both class I and II),as well as Egg PC (a weak CCT activator).


5.4 Discussion

5.4.1 CCT-Membrane Binding Strength and Vesicle Aggregation

The results showing no aggregation of vesicles when CCT was introduced to the Egg PC

vesicles, is reasonable due to the fact that CCT binds poorly to the Egg PC membrane

[96, 77]. Although CCT binds to class I and II lipid membranes, aggregation takes place

only for class I lipid vesicles. As class I lipids are more potent than class II lipids in binding

CCT [78, 97], the aggregation of CCT seems to be correlated to the CCT binding strength

to the membrane. This is supported by the results of aggregation of vesicles made from

Egg PC:Egg PG mixture where concentration of Egg PG was varied. Since membrane-

CCT binding strength depends on the amount of anionic lipids in the membrane, varying

the concentration of Egg PG changes the membrane binding strength of CCT. The results

shows that no aggregation takes place at small concentration of Egg PG and aggregation

starts to occur when the concentration of Egg PG reaches 33 mol %. This suggests that

aggregation can take place when the CCT binding strength is sufficient.

CCT concentration of approximately 80 CCT per vesicle marginally increased the radius

of vesicles containing 33 % Egg PG or 50 % OA (15 or 10 %, respectively). The CCT binding

strength required for the aggregation can be estimated by determining the binding strength

of CCT on these 2 membranes. The CCT binding strength can be determined using the

relation,

∆G = −R T ln Kx, (5.1)

where the partition coefficient Kx is the ratio of bound-to free proteins. Kx value for


Egg PC:Egg PG with 33 mol% Egg PG and Egg PC:OA are in the range of 5 - 8 x 105

[97]. This value corresponds to free energy of binding ∆Gbinding of -7.7 to -8 kcal/mol.

The results suggest that the binding strength must be at least approximately 8 kcal/mole

(10 kbT/molecule) to allow the occurrence of aggregation.

Previous studies on kinetics of binding of myelin basic protein [104] and mitochondrial

creatine kinase octamer [105] on vesicle membrane as well as vesicle aggregation induced by

the proteins show that aggregation is a relatively slow process compared to binding. Based

on these results, for aggregation to take place, time for interaction between protein and

membrane must be longer than time scale for aggregation. The requirement for sufficient

binding strength for CCT to the membranes, therefore, presumably is caused by the fact

that the stronger binding may increase interaction time between CCT and membrane. Con-

sequently, strong binding of CCT to class I lipid membrane may enable more interaction

time which is more than aggregation time scale. On the other hand, weak binding of CCT

on class II lipid membrane may result in short interaction time where CCT dissociates faster

than aggregation time scale.

5.4.2 Cross-Bridging Mechanism for Aggregation

The fact that it takes only a small number of CCT monomers (3 per vesicle) to stimulate

the aggregation of anionic lipid vesicles favors the cross-bridging mechanism for aggregation.

This is different from aggregation induced by other proteins such as : annexin IV [106],

myelin proteolipid [107], and cytochrome c [108], where proteins cover vesicle surface causing

protein oligomerization [106, 107] and/or vesicle charge neutralization [108].


The results showing that CCT induces aggregation of vesicles made from class I but

not for class II lipids reveal the existence of a different mode of binding of CCT on the

membranes for these two activators. In membranes of class I lipids, the mode where each

monomer of CCT binds to the membranes of two vesicles in a cross-bridging mode as shown

in Fig. 5.1b. dominates. However, in membranes of class II lipids, both monomers probably

bind to the membrane of one vesicle as shown in Fig. 5.1a. Alternatively, only one monomer

binds to the membrane while the other remains unbound.

The results showing that CCT induces aggregation of vesicles made from class I but not

for class II lipids and that CCT can be activated by both class I and II lipids indicate that

cross-bridging is not a required step for CCT activation. Although Class II lipids activate

CCT, there is no indication of cross-bridge mode on vesicles containing these lipids.

Binding of CCT in a cross-bridging mode which is dominant in class I lipid membrane,

ensures the binding of two domain Ms, while in non-cross-bridging mode as the case for

class II lipid membrane, there is a possibility that one domain M is unbound. This raises a

question whether the higher activity of CCT in class I lipid membrane is due to the fact that

both domain M engaged in membrane, while in class II lipid membrane only one domain M

bound to membrane. However, these experiments cannot distinguish the mode of binding as

depicted in Fig. 5.1b from the other where only one monomer binds. Further experiments

using CCT heterodimers containing only one domain may answer this question.

The ability of CCT in cross-bridging anionic vesicles is a new aspect of the interactions

between CCT and lipid membranes that has not been studied previously. The in vivo


function of CCT in relation to this ability is not known. However, in Endoplasmic reticulum

and nuclear membranes where CCT translocates, there are 20-30 mol % anionic lipids [109,

110]. This amount is comparable to the threshold concentration of 33 mol % for anionic lipids

in membrane to promote vesicle cross-bridging. This suggests that CCT has a potential to

cross-bridge cell membranes.

5.4.3 Electrostatic Repulsion of Charged Membranes versus Aggregation

The results showing that CCT induces aggregation of negatively charged vesicles (Egg PG

and POPS vesicles) raise an issue of the contribution of the repulsive electrostatic interaction

between charged vesicles. For negatively charged lipid vesicles in an electrolyte solution,

cations in the solution approach the membrane surface, while the anions are repelled from

the surface. Ballance between electrostatic condition as well as thermal motion of the ions

results in a diffusion of high concentration of cations and low concentration of anions near

the membrane surface which is known as the diffusive double layer. As cations bind to the

surface, they neutralize the surface charge, reducing the repulsive electrostatic interaction

between vesicles. Far from the surface, the concentration of ions is the same as the bulk

concentration. The repulsive electrostatic double layer interaction is a long range interaction.

The vesicles also feel a long range van der Waals interaction which is attractive. The

combination is described by the Derjaguin-Landau-Verwey-Overbeek (DLVO) theory [?].

For negatively charged membranes composed of vesicles with radius R in a monovalent

electrolyte solution, the electrostatic energy between surfaces of two vesicles is expressed by


[111],

U =64πkbTRn

κ2

(

tanh

(

eΨo

4kbT

))2

exp(−κD), (5.2)

where e, κ and Ψo are the electron charge, the Debye constant and the surface potential

of the membrane, respectively and n is the density of monovalent electrolyte. The Debye

constant κ is determined using the relation,

κ =

(

2ne2

ǫrǫokbT

)

, (5.3)

where ǫr and ǫo are the relative permittivity of the medium and permittivity of vacuum, re-

spectively. The relation between surface potential Ψo and charge density σ of the membrane

is written as [111],

σ =2ǫrǫoκkbT

esinh

(

eΨo

2kbT

)

. (5.4)

In Eq. 5.4, charge density σ is defined as a cation-bound surface, thus smaller than the

original charge density σo. The relation between σo and σ is written as [112]

σ =σo

1 + Kn exp(

− eΨo

2kbT

) , (5.5)

where K is the binding constant for the monovalent cation. The surface potential can be

determined by equating Eqs. 5.4 and 5.5, if K and σo are known.

The van der Waals interaction energy between surfaces of two vesicles separated by

distance D is written as [113],

V = V (R, R) + V (R − t, R − t) − 2V (R, R − t), (5.6)


where t is the thickness of the bilayer and,

V (R, R − t) = −A

6

( 2R(R − t)

(D + 2R) − (R + (R − t))2

+2R(R − t)

(D + 2R) − (R − (R − t))2

+ ln(D + 2R) − (R + (R − t))2

(D + 2R) − (R − (R − t))2

)

. (5.7)

Thus, the total interaction energy of two vesicles is given by,

G = U + V. (5.8)

There are also solvation and steric interactions at short range, but for monovalent electrolyte

such as NaCl used in these experiments, the interaction can be described sufficiently by

DLVO theory [112].

Figure 5.9 shows the graphs of U, V and G as a function of distance between surfaces

of two vesicles for Egg PG vesicles in 150 mM NaCl as used in these experiments. Values

used in these calculations include σo of approximately 0.235 C/m2, equivalent to 1 electron

per headgroup, a binding constant K for Na+ of 0.7 M−1 [114], and a bilayer thickness

of 4 nm [100], while the radius of the vesicles was estimated from DLS experiments. The

figure shows that two vesicles repel each other over all length scales with a maximum value

at a separation distance around 0.2 nm. Figure 5.9 indicates that the ions present, which

can screen the surface charge are not sufficiently concentrated to induce vesicle aggregation,

which is consistent with observations of pure POPS and Egg PG vesicle dispersions. The

figure also suggests that, for aggregation to take place, the size of CCT must be comparable

to or larger than 4 nm, the distance where the system experiences more repulsive interactions


0 1 2 3 4 5

-150

-100

-50

0

50

100

150

Inte

ract

ion

ener

gy (k

BT)

D (nm)

van der Waals interaction Electrostatic interaction Total interation

Figure 5.9: Electrostatic and van der Waals energy between vesicles as a function of surfaceseparation D.


when the vesicles approach each other. The size of the CCT dimer has been estimated to

be 10 nm as determined from the structure of glycerophosphate cytidylyltransferase [88]

and domain M [115]. Thus, there is a repulsive interaction between vesicles at a very

short distances that CCT must overcome, but the CCT dimension is sufficient to make this

possible.

5.5 Summary

This chapter contains studies on the effect of CCT on aggregaion of vesicles containing class

I and class II CCT activators. Aggregation occurs when CCT is introduced to samples

of class I lipid vesicles. In contrast, there is no indication of aggregation when CCT is

introduced to samples of class II lipid vesicles. The results suggest that binding of CCT to

class I lipid membranes was dominated by the cross-bridging mode, while in class II lipid

membranes either two domain Ms bind to the same vesicle or only one monomer binds while

the other is unbound. It is thought that the cross-bridging mode is correlated to the strength

of the CCT binding to membrane which presumably enables interaction time between CCT

and membrane. The cross-bridge occurs when unbound domain M whose one domain M

engaged on one vesicle collides with other vesicle. This requires binding time scale larger

than collision or cross-bridging time scale.

Chapter 6

Conclusion

This thesis discusses two biophysical topics : studies of the effect of sterols on the mechanical

strength of lipid membranes and the interaction between CCT and membrane lipids.

Studies on the effect of sterols on lipid membranes show that sterols, including choles-

terol, lanosterol and ergosterol, increase the mechanical strength of the membranes, where

cholesterol shows the largest effect followed by lanosterol and ergosterol. The increase in

the strength of lipid membrane by sterols is correlated to the increase in the chain order

of the lipids by sterols. The increase in the membrane strength by sterols indicates their

contribution to cell viability, which depends on maintaining an intact plasma membrane.

Studies on vesicle aggregation induced by CCT show that CCT induces aggregation of

vesicles made from class I lipids. In contrast, CCT does not induce aggregation of vesicles

made from both non activating and class II lipids. The occurrence of aggregation depends

108

CHAPTER 6. CONCLUSION 109

on the binding strength of CCT on the membrane. The results suggest that CCT cross-

bridges two vesicles. The cross-bridging mechanism shown by CCT is a new aspect of CCT

interaction with lipid bilayer.

Bibliography

[1] R.R.C. New. Influence of liposome characteristics on their properties and fate. In Li-

posomes as tools in basic research and industry, J.R.Philippot, F.Schuber. Eds. (CRC

Press, Boca Raton, 1995), pp.4-20.

[2] L.D. Mayer, M.J. Hope and P.R. Cullis. Vesicles of variable sizes produced by a rapid

extrusion procedure. Biochim. Biophys. Acta. 858:161-168, 1986.

[3] P.J. Patty and B.J. Frisken. The pressure-dependence of the size of extruded vesicles.

Biophys. J. 85:996-1004, 2003.

[4] B. Mui, P.R. Cullis, E. Evans and T.D. Madden, “Osmotic properties of large unil-

amellar vesicles prepared by extrusion,” Biophys. J. 64, 443-453 (1993).

[5] Y. Talmon, J.L. Burns, M.H. Chesnut, D.P. Siegel, “Time-resolved Cryotransmission

electron microscopy,” J. Electron Microsc. Tech. 14, 6-12 (1990).

[6] M. Almgren, K. Edwards, and G. Karlsson, “Cryo transmission electron microscopy of

liposomes and related structures,” Colloid. Surf. A. 174, 3-21 (2000).

110

BIBLIOGRAPHY 111

[7] B.A. Korgel, J.H. van Zanten, and H.G. Monbouquette, “Vesicle size distribution mea-

sured by field-flow-fractionation couple with multiangle light scattering ,” Biophys. J.

74, 3264-3272 (1998).

[8] J.H. van Zanten, and H.G. Monbouquette, “Characterization of vesicles by classical

light scattering,” J. Colloid Interface Sci. 146, 330-336 (1991); Phosphatidylcholine

vesicle diameter, molecular weight and wall thickness determined by static light- scat-

tering,” J. Colloid Interface Sci. 165, 512-518 (1994).

[9] J.C. Selser and R.J. Baskin, “A light scattering characterization of membrane vesicles,”

Biophys. J. 16, 337-356 (1976).

[10] F.R. Hallett, J. Watton and P. Krygsman, “Vesicle sizing-number distribution by dy-

namic light scattering,” Biophys. J. 59, 357-362 (1991).

[11] S. Kolchens, V. Ramaswami, J. Birgenheier, L. Nett and D.F. O’Brien, “Quasi-elastic

light scattering determination of the size distribution of extruded vesicles,” Chem. and

Phys. Lipids 65, 1-10 (1993).

[12] A.J. Jin, D. Huster and K. Gawrisch, “Light scattering characterization of extruded

lipid vesicles,” Eur. Biophys. J. 28, 187-199 (1999).

[13] B.J. Frisken, C. Asman and P.J. Patty, “Studies of vesicle extrusion,” Langmuir 16,

928-933 (2000).

[14] D.E. Koppel, “Analysis of macromolecular polydispersity in intensity autocorrelation

spectroscopy : the method of cumulants,” J. Chem. Phys. 57, 4814-4820 (1972).

BIBLIOGRAPHY 112

[15] B.J. Frisken, “Revisiting the method of cumulants for analysis dynamic light scattering

data,” J. Appl. Opt. 40, 4087-4091 (2001).

[16] S. Provencher, “Contin : a general purpose constrained regularization program for

inverting noise linear algebraic and integral equations ,” Comp. Phys. Comm 27, 229-

242 (1982);“ A constrained regularization method for inverting data represented by

linear algebraic or integral equations ,” Comp. Phys. Comm 27, 213-227 (1982).

[17] F.R. Hallett, T. Craig, J. Marsh and B. Nickel, “Particle size analysis : number distri-

butions by dynamic light scattering,” Can. J. Spectr.34, 63-70 (1989).

[18] G. Bryant, C. Abeynayake and J.C. Thomas, “Improve particle size distribution mea-

surements using multiangle dynamic light scattering. 2. Refinements and applications,”

Langmuir 12, 6224-6228 (1996).

[19] L.H. Hanus and H.J. Ploehn, “Conversion of intensity-average photon correlation spec-

troscopy measurements to number-average particles size distributions.1. Theoretical

development,” Langmuir 15, 3091-3100 (1999).

[20] J.C. Thomas, “The determination of log normal particle size distributions by dynamic

light scattering,” J. Colloid Interface Sci. 117, 187-192 (1987).

[21] C.B. Bargeron, “Measurements of continuous distribution of spherical particles by in-

tensity correlation spectroscopy : analysis by cumulants,” J. Chem. Phys. 61, 2134-2138

(1974).

BIBLIOGRAPHY 113

[22] D.S. Horne, “Determination of the size distribution of Bovine Casein micelles using

photon correlation spectroscopy,” J. Colloid Interface Sci. 98, 537-548 (1984).

[23] T.W. Taylor, S.M. Scrivner, C.M. Sorensen and J.F. Merklin, “Determination of the

relative number distribution of particle sizes using photon correlation spectroscopy ,”

Appl. Opt. 24, 3713-3717 (1985).

[24] J.C. Selser, “Letter : A light scattering method of measuring membrane vesicle-number

average size and size dispersion,” Biophys. J. 16, 847-848 (1976).

[25] J. Pencer and F.R. Hallett,”Effect of vesicle size and shape on static and dynamic light

scattering measurements,” Langmuir 19, 7488-7497 (2003).

[26] B.J. Berne and R. Pecora, “Dynamic light scattering with applications to chemistry,

biology and physics”, (Oxford University Press, Oxford, 1993).

[27] P. Stepanek, “Data analysis in dynamic light scattering”, In Dynamic light scattering

: the method and some applications, W. Brown. Ed. (Oxford University Press, Oxford,

1993), pp.177-241.

[28] D.G. Hunter and B.J. Frisken, “The effects of lipid composition and extrusion pressure

and temperature on the properties of phospholipid vesicles,” Biophys. J. 74, 2996-3000

(1998).

[29] G. Maulucci, M. De Spirito, G. Arcovito, F. Boffi, A.C. Costellano and G. Briganti,

“Particle size distribution in DMPC vesicles solutions undergoing different sonication

times,” Biophys. J. 88, 3545-3550 (2005).

BIBLIOGRAPHY 114

[30] J. Israelachvili, “Intermolecular and Surface Forces”, (Academic Press, San Diego,

1992).

[31] P.J. Patty and B.J. Frisken, Private Communication, 2005.

[32] G. Gregoriadis, Ed.“Liposomes as Drug Carrier Recent Trends and Progress”, (John

Wiley and Sons, New York, 1988).

[33] C. Nicolau and A. Cudd, “Liposomes as carriers of DNA.“ Crit. Rev. Therap. Drug

Carr. Systems 6, 239-271 1989.

[34] J.L. Rigaud and B. Pitard,“Liposomes as tools for reconstituion of Biological systems”,

In Liposomes as tools in basic research and industry, J.R.Philippot, F.Schuber. Eds.

(CRC Press, Boca Raton, 1995), pp.71-88.

[35] D.D. Lasic, “Liposomes : from Physics to applications”, (Elsevier, ????????, 1993).

[36] D.D. Lasic,“Novel applications of liposomes,” TIBTECH 16 307-321, (1998).

[37] K. Matsuzaki, O. Murase, K. Sugishita, S. Yoneyama, K. Akada, M. Ueha, A. Naka-

mura, S. Kobayashi,“Optical characterization of liposomes by right angle light scatter-

ing and turbidity measurement,” Biochim. Biophys. Acta 1467 219-226, (2000).

[38] S. Sriwongsitanont and M. Ueno,“Effect of a PEG lipid (DSPE-PEG2000) and freeze-

thawing process on phospholipid vesicle size and lamellarity,”Colloid Polym. Sci. 282

753-760, (2004).

[39] G. Cowan, “Statistical data analysis”, (Oxford University Press, Oxford, 1998).

BIBLIOGRAPHY 115

[40] Chonn, A. and Cullis, P.R. Adv. Drug Deliv. Rev. 30, 73-xx, (1998).

[41] M.C. Woodle, M.S. Newman, and P.K. Working. “Biological properties of sterically

stabilized liposomes”, In Stealth Liposomes, D. Lasic, and F. Martin. Eds. (CRC Press,

Boca Raton, 1995), pp.xx-xx.

[42] Litzinger, D.C., Buiting, A.M.J., van Rooijen, N., and Huang, L. Biochim.Biophys.Acta

99 xx-xx, (1994).

[43] V. Roden and S.M. Goldin. Formation of unilamellar lipid vesicles of controlable di-

mensions by detergent dialysis. Biochemistry 18:4173-4176, 1979.

[44] Hope, M.J., Bally, M.B., Webb G., and Cullis, P.R. Biochim. Biophys. Acta 124,

(1985).

[45] F. Szoka and D. Papahadjopoulos. Procedure for preparation of liposomes with large

internal aqueous space and high capture by reverse-phase evaporation. Proc. Natl.

Acad. Sci. USA 75:4194-4198, 1978.

[46] H.J. Gruber and H. Schindler. External surface and lamellarity of lipid vesicles: a

practice-oriented set of assay methods. Biochim. Biophys. Acta. 1189:212-224, 1994.

[47] J.C. McIntyre and R.G. Sleight. Fluorescence assay for phospholipid membrane asym-

metry. Biochemistry. 30:11819-11827, 1991.

[48] M. Vogel, C. Munster, W. Fenzl, and T. Salditt. Thermal unbinding of highly oriented

phospholipid membranes. Phys. Rev. Lett. 84:390-393, 2000.

BIBLIOGRAPHY 116

[49] L. Ingebrigtsen and L. Brandl. Determination of the size distribution of liposomes

by SEC fractionation and PCS analysis and enzymatic assay of lipid content. AAPS.

Pharm. Sci. Tech. 3:1-7, 2002.

[50] E. Evans and W. Rawicz. Entropy driven tension and bending elasticity in condensed-

fluid membranes. Phys. Rev. Lett. 64:2094-2097, 1990.

[51] W. Rawicz, K. C. Olbrich, T. McIntosh, D. Needham and E. Evans. Effect of chain

length and unsaturation on elasticity of lipid bilayers. Biophys. J. 79:328-339, 2000.

[52] C. Tanford. The hydrophobic effect. 2nd. edn. (Wiley, New York, 1980).

[53] D. Boal. Mechanics of the Cell. (Cambridge University Press, Cambridge, 2002).

[54] . M. Wortis and E. Evans. Membrane self assembly : Mechanical properties and vesicle

shapes. Physics in Canada. 5:281-288, 1997.

[55] D. Needham and R.S. Nunn. Elastic deformation and failure of lipid bilayer membrane

containing cholesterol. Biophys. J. 58:997-1009, 1990.

[56] D. Needham, T.J. McIntosh and E. Evans. Thermomechanical and transition proper-

ties of dimyristoylphosphatidylcholine/cholesterol bilayer. Biochemistry 27:4668-4673,

1988.

[57] H.P. Duwe, J. Kaes, and E. Sackmann. Bending elastic moduli of lipid bilayers : mod-

ulation of solutes. J. Phys. Fr 51:945-962, 1990.

BIBLIOGRAPHY 117

[58] P. Meleard, C. Gerbeaud, T. Pott, L. Fernandez-Puente, I. Bivas, M.D. Mitov, J. Du-

fourcq, J. Bothorel. Bending elasticities of model membranes : influence of temperature

and sterol content. Biophys. J. 72:2616-2629, 1997.

[59] M.R. Vist and J.H. Davis. Phase equilibria of cholesterol/dipalmitoylphospha-

tidylcholine mixtures : 2H nuclear magnetic resonance and differential scanning calo-

rymetry. Biochemistry 29:451-464, 1990.

[60] F.M. Linseisen, J.L. Thewalt, M. Bloom and T. M. Bayerl. H-2-NMR and DSC study

of SEPC-Cholesterol mixtures Chem. Phys. Lipids 65:141-149, 1993.

[61] J.L. Thewalt and M. Bloom. Phosphatidylcholine:cholesterol phase diagram. Biophys.

J. 63:1176-1181, 1992.

[62] L. Finegold, Cholesterol in Membrane Models. (CRC Press, Boca Raton, 1995).

[63] K.E. Bloch. Sterol structure and membrane function. CRC. Crit.Rev.Biochem. 14:47-

92, 1983.

[64] E. Endress, S. Bayerl, K. Prechtel, C. Maier, R. Merkel and T.M. Bayerl. The effect

of cholesterol, lanosterol and ergosterol on lecithin bilayer mechanical properties at

molecular and microscopic dimensions : A solid-state NMR and micropipet study.

Langmuir 18:3293-3299, 2002.

[65] E. Endress, H. Heller, H. Casalta, M.F. Brown, and T.M. Bayerl. Anisotropic motion

and molecular dynamics of cholesterol, lansoterol and ergosterol in lecithin bilayers

studied by quasi-elastic neutron scattering. Biochemistry 41:13078-13086, 2002.

BIBLIOGRAPHY 118

[66] G.V. Martinez, E.M. Dykstra, S. Lope-Piedrafita, and M.F. Brown. Lanosterol and

cholesterol-induced variations in bilayer elasticity probed by 2H NMR relaxation. Lang-

muir 20:1043-1046, 2004.

[67] J. Henriksen, A.C. Rowat and J.H. Ipsen. Vesicle fluctuation analysis on the effect of

sterols on membrane bending rigidity. Eur. Biophys. J. 33:732-741, 2004.

[68] Ya-Wei Hsueh, K. Gilbert, C. Trandum, M. Zuckermann, and J. Thewalt. The ef-

fect of ergosterol on dipalmitoylphosphatidylcholine bilayers : A deuterium NMR and

calorimetric study. Biophys. J. 88:1799-1808, 2005.

[69] K.J. Tierney, D.E. Block and M.L. Longo. Elasticity and phase behavior of DPPC

membrane modulated by cholesterol, ergosterol and ethanol. Biophys. J. 89:2481-2493,

2005.

[70] J. Henriksen, A.C. Rowat, E. Brief, YW Hsueh, J.L. Thewalt, M.J. Zuckermann and

J.H. Ipsen. Universal behavior of membranes with sterols. Biophys. J. 90:1639-1649,

2006.

[71] L.D. Landau and E. Lifshitz, Fluid Mechanics 2nd Ed., (Pergamon Press, Oxford,

1987).

[72] M. Bloom, E. Evans and O.G. Mouritsen. Physical properties of the fluid lipid bilayer

component of the cell membranes : a perspective. Quart. Rev. Biophys. 24:293-397,

1991.

BIBLIOGRAPHY 119

[73] Z. Chen and R.P. Rand. The influence of cholsterol on phospholipid membrane curva-

ture and bending elasticity. Biophys. J. 72:2616-2629, 1997.

[74] . Unpublished.

[75] H. Schindler and J. Seelig. Deuterium order parameters in relation to thermodynamic

properties of phospholipid bilayers. A statitistical mechanical interpretation. Biochem-

istry 14:2283-2287, 1975.

[76] J.D. Litster. Stability of lipid bilayers and red blood cell membranes. Phys. Lett.

53A:193-194, 1975.

[77] R.S. Arnold, A. De Paoli-Roach and R.B. Cornell. Binding of CTP:phosphocholine

cytidylyltransferase to lipid vesicles: Diacylglycerol and enzyme dephosphorylation in-

crease the affinity for negaively charged membrane. Biochemistry 36:6149-6156, 1997.

[78] J.E. Johnson N.M. Rao, S.W. Hui and R.B. Cornell. Conformation and lipid

binding properties of four peptides derived from the membrane-binding domain of

CTP:phosphocholine cytidylyltransferase . Biochemistry 37:9509-9519, 1998.

[79] J.E. Johnson, M. Xie, L.M. Sing, R. Edge and R.B. Cornell. Both acidic and basic amino

acids in an amphitropic enzyme, CTP:phosphocholine cytidylyltransferase, dictate its

selectivity for anionic membranes. J. Biol. Chem. 278:514-522, 2003.

[80] R.B. Cornell and I.C. Northwood. Regulation of CTP:phosphocholine cytidylyltrans-

ferase by amphitropism and relocalization. Trends Biochem. Sci 25:441-447, 2000.

BIBLIOGRAPHY 120

[81] R. Sleight and C. Kent. Regulation of phosphatidylcholine biosynthesis in mammalian

cells. II. Effects of phospholipase C treatment on the activity and subcellular distribu-

tion of CTP:phosphocholine cytidylyltransferase in Chinese hamster ovary and LM cell

lines. J. Biol. Chem. 258:831-835, 1983.

[82] H. Tronchere, M. Record, F. Terce, and H. Chap. Phosphatidylcholine cycle and regu-

lation of phosphatidylcholine biosynthesis by enzyme translocation. Biochim. Biophys.

Acta. 1212:137-151, 1994.

[83] H. Tronchere, V. Planat, M. Record, F. Terce, G. Ribbes, and H. Chap. Phosphatidyl-

choline turnover in activated human neutrophils-agonist-induced cytidylyltransferase

translocation is subsequent to phospholipase-D activation. J. Biol. Chem. 270:13138-

13146, 1995.

[84] Y. Wang, T.D. Sweitzer, P.A. Weinhold, and C. Kent. Nuclear localization of soluble

CTP:Phosphocholine Cytidylyltransferase. J. Biol. Chem. 268:5899-5904, 1993.

[85] I.C. Northwood, A.H.Y. Tong, B. Crawford, A.E. Drobnies, and R.B. Cornell. Shut-

tling of CTP:Phosphocholine Cytidylyltransferase between the nucleus and endoplasmic

reticulum accompanies the wave of phosphatidylcholine synthesis during the G0 and

G1 transition. J. Biol. Chem. 274:26240-26248, 1999.

[86] R.B. Cornell. Chemical cross-linking reveals a dimeric structure for

CTP:phosphocholine cytidylyltransferase. J. Biol. Chem. 264:9077-9081, 1989.

BIBLIOGRAPHY 121

[87] Y. Wang, T.D. Sweitzer, P.A. Weinhold and C. Kent. Nuclear localization of soluble

CTP:phosphocholine cytidylyltransferase. J. Biol. Chem. 268:5899-5904, 1993.

[88] C.H. weber, Y.S. Park, S. Sanker, K. Kent, and M. Ludwig. A prototypical cytidy-

lyltransferase : CTP:glycerol-3-phosphate cytidylyltransferase from Bacillus subtilis.

Structure 7:1113-1124, 1999.

[89] M. Xie, J.L. Smith, Z. Ding, D. Zhang and R.B. Cornell. Membrane binding modulates

the quaternary structure of CTP:phosphocholine cytidylyltransferase. J. Biol. Chem.

279:28817-28825, 2004.

[90] J.E. Johnson and R.B. Cornell. Membrane binding amphipathic α helical peptide de-

rived from CTP:phosphocholine cytidylyltransferase . Biochemistry 33:4327-4335, 1994.

[91] J.I.S. MacDonald and C. Kent. Identification of phosphorylation sites in rat liver

CTP:phosphocholine cytidylyltransferase . J. Biol. Chem. 269:10529-10537, 1994.

[92] H. Jamil, G. Hatch, and D.E. Vance.Evidence that binding of CTP:phosphocholine

cytidylyltransferase to membranes in rat hepatocytes is modulated by the ratio of

bilayer- to non bilayer-forming lipids. Biochem. J. 291:419-427, 1993.

[93] G.S. Attard, R.H. Templer, W.S. Smith, A.N. Hunt, and S. Jackowski. Modulation of

CTP:phosphocholine cytidylyltransferase by membrane curvature elastic stress. Proc.

Natl. Acad. Sci. USA. 97:9032-9036, 2000.

BIBLIOGRAPHY 122

[94] S.M.A. Davies, R.M. Epand, R. Kraayenhof, and R.B. Cornell. Regulation of

CTP:phosphocholine Cytidylyltranferase activity by physical properties of lipid mem-

branes: An important role for stored curvature strain energy. Biochemistry 40:10522-

10531, 2001.

[95] R.S. Arnold and R.B. Cornell. Lipid regulation of CTP:phosphocholine cytidylyltrans-

ferase : electrostatic,hydrophobic, and synergistic interactions of anionic phospholipids

and diacylglycerol . Biochemistry 35:9917-9924, 1996.

[96] R.B. Cornell and D.E. Vance. Binding of CTP: phosphocholine cytidylyltransferase to

large unilamellar vesicles Biochim. Biophys. Acta. 919:37-48, 1987.

[97] S.G. Taneva, P.J. Patty, B.J. Frisken and R.B. Cornell. CTP:Phosphocholine Cytidylyl-

transferase binds anionic Phospholipid vesicles in a cross-bridging mode. Biochemistry

44:9382-9393, 2005.

[98] J.A. Friesen, H.A. Campbell and C. Kent. Enzymatic and cellular characterization of

a catalytic fragment of CTP:phosphocholine cytidylyltransferase α . J. Biol. Chem.

19:133849-13389, 1999.

[99] S. Taneva, J.E. Johnson and R.B. Cornell. Lipid-induced conformational switch in the

membrane binding domain of CTP:Phosphocholine Cytidylyltransferase: A circular

dichroism study. Biochemistry 42:11768-11776, 2003.

[100] M.D. Houslay, K.K. Stanley. Dynamics of biological membranes : Influence of synthe-

sis, structure and function. Willey, New York, 1982.

BIBLIOGRAPHY 123

[101] R.W. Evans, M.A. Williams, and J. Tinoco. Surface area of 1-palmitoyl phosphatidyl-

cholines and their interactions with cholesterol. Biochem. J 245:455-462, 1987.

[102] T.H. Ha, D.K. Kim, M.-U. Choi, and K. Kim. Influence of poly(ethylenimine) on the

monolayer of oleic acid at the air water interface. J. Colloid Interface Sci. 226:98-104,

2000.

[103] G.R. Barlett. Phosphorous assay in column chromatography. J. Biol. Chem. 234:466-

468, 1959.

[104] M.B.A. ter Beest and D. Hoekstra. Interaction of myelin basic protein with artifi-

cial membranes. Parameters governing binding, aggregation and dissociation. Eur. J.

Biochem. 211:689-696, 1993.

[105] O. Stachowiak, M. Dolder and T. Wallimann. Membrane-binding and lipid vesi-

cle cross-linking kinetics of the mitocondrial creatin kinase octamer. Biochemistry

35:15522-15528, 1996.

[106] M.A. Kaetzel, Y.D. Mo, T.R. Mealy, B. Campos, W. Bergsma-Schutter, A. Brisson,

J.R. Dedman, and B.A. Seaton. Phosphorylation mutants elucidate the mechanism of

Annexin IV-mediated membrane aggregation Biochemistry 40:4192-4199, 2001.

[107] O.A. Bizzozero, and T.A. Howard Myelin proteolipid protein-induced aggregation of

lipid vesicles : efficacy of the various molecular species. Neurochem. Res. 27:1269-1277,

2002.

BIBLIOGRAPHY 124

[108] S. Bernad, S. Oellerich, T. Soulimane, S. Noinville, M.H. Baron, M. Paternostre, and

S. Lecomte. Interaction of horse heart and thermus thermophilus type cytochromes

with phospholipid vesicles and hydrophobic surfaces. Biophys. J. 86:3863-3872, 2004.

[109] C. D’Antuono, M. del Carmen-Fernandez Tome, N. Sterin-Speziale and D.L. Bernik.

Lipid-protein interactions in rat renal subcellular membranes: A biophysical and bio-

chemical study.Arch. Biochem. Biophys. 382:39-47, 2000.

[110] T. Neitcheva and D. Peeva. Phospholipid composition, phospholipase A2 and spin-

gomyelinase activities in rat liver nuclear membrane and matrix. Int. J. Biochem. Cell

Biol. 327:995-1001, 1995.

[111] H. Ohshima, in : H. Ohshima and K. Furusawa (Eds). Electrical phenomena at inter-

face, 2nd ed., Marcel Dekker, New York, 1998.

[112] S. Ohki and H. Oshima. Interaction and aggregation of lipid vesicles (DLVO theory

versus modified DLVO theory). Colloids and Surfaces B. 14:27-45, 1999.

[113] K. Ohsawa, H. Ohshima and S. Ohki. Surface-potential and surface-charge density

of the cerebral-cortex synaptic vesicle and stability of vesicle suspension. Biochim.

Biophys. Acta 648:206-214, 1981.

[114] M. Eisenberg, T. Gresalfi, T. Riccio and S. McLauchlin. Title Biochemistry 18:5213-

5223, 1979.

BIBLIOGRAPHY 125

[115] J.S. Dunne, R.B. Cornell, J.E. Johnson, N.R. Glover, and A.S. Tracey. Structure of the

membrane binding domain of CTP:Phosphocholine Cytidylyltransferase Biochemistry

35:11975-11984, 1996.

[116] A. Lykidis, I. Baburina and S. Jackowski. Distribution of CTP:phosphocholine Cytidy-

lyltranferase (CCT) isoforms. J. Biol. Chem. 274:26992-27001, 1999.

[117] S. Jackowski. Cell cycle regulation of membrane phospholipid metabolism. J. Biol.

Chem. 271:20219-20222, 1996.

[118] D.P. Veitch and R.B. Cornell. Substitution of serine for glycine-91 in the HXGH

motif of CTP:phosphocholine cytidylyltransferase implicates this motif in CTP binding.

Biochemistry 35:10743-10750, 1996.

[119] P.C. Choy and D.E. Vance. Lipid requirements for activation of CTP:phosphocholine

cytidylyltransferase. J. Biol. Chem. 253:5163-5167, 1978.

[120] Theory of breakdown (rupture) of free films. Kolloidn. Zh 24:370-374, 1962.

using vesicles to study the effect of sterols on the … · 2007-03-30 · using vesicles to study...

Documents