Behaviour of α-elastin in bulk and at aqueous surfaces

Transcription

Behaviour of α-elastin in bulk and at aqueous
surfaces
A.R.Lindsay
School of Physics
University of Exeter
Behaviour of α-elastin in bulk and at aqueous
surfaces
Submitted by Amanda R. Lindsay, to the University of Exeter as a thesis for
the for the degree of Doctor of Philosophy in Physics, April 2011.
This thesis is available for Library use on the understanding that it is copyright
material and that no quotation from this thesis may be published without proper
acknowledgement.
I certify that all material in this thesis which is not my own work has been
identified and that no material has previously been submitted and approved for the
award of a degree by this or any other University.
...................................... (Amanda R. Lindsay)
i
‘But I should like to know-’ Pippin began
‘Mercy!’ cried Gandalf. ‘If the giving of information is to be the cure of your inquisitiveness, I
shall spend all the rest of my days in answering you. What do you wish to know?’
‘The names of all the stars, and of all living things, and the whole history of Middle-earth and
Over-heaven and of the Sundering Seas,’ laughed Pippin. ‘Of course. What less? But I am not in
a hurry tonight’
Exchange between Peregrin (Pippin) Took and Gandalf the White from ‘The Two
Towers - The Lord of the Rings Part 2’ By J.R.R. Tolkien
ii
Abstract
The purpose of this work was to examine the behaviour of the soluble elastin
derivative, α-elastin, under a variety of conditions. Although studies of α-elastin in
solution have been made, confining the protein molecules to a two dimensional state
in a monolayer allows probing of different conformational states.
Bulk viscometry experiments indicated, consistent with previous work, that Ca2+
affects α-elastin differently to Na+ . The intrinsic viscosity of α-elastin in water was
0.0073 mL/mg at room temperature and it was seen to increase with decreasing
temperature. In 0.1 M calcium chloride it was seen that the radius of gyration of
the elastin increased by 6% with a 17◦ C rise in temperature, whereas in water and
0.1 M sodium chloride the increase was only 2%.
When confined to the surface it was demonstrated that α-elastin monolayers on
water behave viscoelastically in the surface pressure range 12-20 mN/m , viscoelastic
behaviour was also seen on 0.1 M CaCl2 in the surface pressure range 14-18 mN/m.
Examination of the dissipative component of the complex modulus showed phase
transition occurring between 8 and 10 mN/m on both water and calcium chloride
subphase. From the value of the dissipative component below the phase transition,
the transition was identified as semi-dilute to concentrated. Fitting to Eyring’s
model allowed calculation of the area per segment in motion of the α-elastin on
water Am = 48 Å2 /segment, and on calcium chloride Am = 76 Å2 /segment. These
values are consistent with calculated areas per molecule which indicate that at a
given surface pressure an α-elastin molecule on calcium chloride takes up at least
2.4 times are much space as a molecule on water.
Quasi-static compressions and extensions of α-elastin monolayers were carried out
on three different subphases at three different values of pH. This gave the surface
pressure - area isotherm for each. Dilational modulus calculation indicated a phase
iii
transition at around 8 mN/m. Below 5 mN/m the monolayer was in the semidilute regime, between good and Θ solvent conditions, the Flory exponent, ν =
0.671 ± 0.002. This indicated that the observed phase transition is semi-dilute to
concentrated which confirms the surface viscometry results. The salt solutions were
seen to provide the α-elastin with conditions closer to good solvent with νN aCl =
0.70 ± 0.04. and νCaCl2 = 0.7 ± 0.2.
Fitting an exponential to the decay of surface pressure at constant area indicated
that time constant for the decay was consistent for all three subphases, τ = 0.001 ±
0.0003 s. This consistency shows that the relaxation of the monolayer is not limited
by electrostatic interactions between the monolayer and the subphase.
By altering the temperature of the subphase it was seen that α-elastin forms
monolayers up to 40◦ C, which is above the temperature at which bulk solutions of
α-elastin coacervate and drop out of solution. Thus, coacervation in two dimensions
does not occur under the same external conditions as in three dimensions.
By examining the pressure at the surface of very dilute solutions of α-elastin it
was seen that monolayers did not spontaneously form at a water surface However,
with stirring a monolayer formed quickly which resulted in a surface pressure rise
of 15 ± 2 mN/m. When α-elastin was added below a lipid monolayer and stirred
it was seen that it was able to insert at surface pressure above 30 mN/m. Using
fluorescently labelled lipids it was seen that the label did not affect the elastin
insertion into the lipid monolayer. It was seen that the α-elastin insertion into a
PC:NBD-PC monolayer disrupted the domain structure of the monolayer.
iv
Acknowledgements
Studying for a PhD is hard work and I could not have completed this project
alone so many thanks are due, to;
My supervisor, Peter Petrov for offering me this project in the first place and for
all the support he has given me in the last four and a half years.
Peter Winlove for support, encouragement and advice.
Pietro Cicuta and his students: Elodie Aumaitre and Betta Spigone for welcoming me at the Cavendish and allowing me to take the surface viscometry data
and dynamic surface monolayer data presented in Chapters 4 and 5 of this work,
respectively.
My mentor John Inkson for advice and a listening ear especially through the hard
times.
All the members, past and present, of the Exeter BioPhysics group for the laughter, zoo trips, barbecues, and meals out. Special mentions to: Irman, although we
never researched together, your advice was invaluable; to John, for helping me with
the microscope; Jess ‘the Queen of TPF Microscopy’ for taking me some pictures of
elastin; and to James, without your project mine would not have been, your humour
during the last days saved my sanity.
Dick Ellis and Ellen ‘the patron saint of knowing where everything is’ Green who
were responsible for purifying my elastin and who have made my lab work a lot
easier.
Tim Naylor for organising the funding that has allowed me to spend the last four
and a half years proding protein films.
All members of Exeter University’s Science Fiction, Tolkien and Games Societies
v
for providing diversions when my brain was in need of “defragmentation” and for
understanding my erratic attendance.
To all members of KSMBDA, and the Exeter Aikido Club in particular, for letting
me bounce them around the mat and bouncing me around in turn as a very effective
way of coping with stress. Sensei Alex, you will always be remembered.
Matt for being a shoulder to sigh on and a listening ear when the pressures were
getting to me. I’ll make it up to you when its your turn to write up.
My sister, Susi, for always being able to make me smile and for keeping me
grounded and reminding me which way was up.
My parents, Celia and Terry, for supporting me, getting hard to find papers for
me, convincing me I could reach the stars and showing enthusiasm for what I was
doing even when my explanations left them understanding only one word in three.
Also to my mum again for proof reading this thesis, any mistakes that remain are
purely my own, and for explaining grammar to me in the first place.
vi
Contents
Quotation
i
Abstract
iii
Acknowledgements
v
Contents page
vii
List of Figures
xi
1 Introduction
1
1.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
1.2
Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2
2 Background
4
2.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4
2.2
Introducing Elastin . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
2.2.1
Forms of Elastin . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.2
α-Elastin
2.2.3
α-elastin coacervation
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
. . . . . . . . . . . . . . . . . . . . . . 20
vii
2.3
2.4
2.5
2.2.4
Effects of Calcium on Elastin . . . . . . . . . . . . . . . . . . 24
2.2.5
Lipid-Elastin Interactions . . . . . . . . . . . . . . . . . . . . 33
2.2.6
The Elastin Receptor . . . . . . . . . . . . . . . . . . . . . . . 36
2.2.7
Mechanical properties of Elastin . . . . . . . . . . . . . . . . . 36
2.2.8
Elastin as a Biomaterial . . . . . . . . . . . . . . . . . . . . . 38
2.2.9
Biological Problems Relating to Elastin . . . . . . . . . . . . . 39
Langmuir Monolayers . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.3.1
Development of the Langmuir Methodology . . . . . . . . . . 40
2.3.2
Reading a Π-A Isotherm . . . . . . . . . . . . . . . . . . . . . 42
2.3.3
Monolayer Phase Calculations . . . . . . . . . . . . . . . . . . 45
2.3.4
Surface Rheology . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.3.5
Elastic Moduli
. . . . . . . . . . . . . . . . . . . . . . . . . . 52
Viscometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.4.1
Introduction to Viscosity . . . . . . . . . . . . . . . . . . . . . 59
2.4.2
System Considerations for Viscosity Measurement . . . . . . . 60
2.4.3
Poiseulle’s Law . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.4.4
Surface Viscosity . . . . . . . . . . . . . . . . . . . . . . . . . 65
Behaviour of Polymers in Solution . . . . . . . . . . . . . . . . . . . . 68
2.5.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
2.5.2
Thermodynamics of a Polymer Solution . . . . . . . . . . . . . 68
3 General Methodology
72
3.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.2
Elastin Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.2.1
α-Elastin Characterisation . . . . . . . . . . . . . . . . . . . . 74
viii
3.3
Other Materials and Equipment . . . . . . . . . . . . . . . . . . . . . 76
3.4
Langmuir Troughs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.4.1
Cleaning of the Trough . . . . . . . . . . . . . . . . . . . . . . 78
3.4.2
Spreading a Monolayer . . . . . . . . . . . . . . . . . . . . . . 79
4 Viscometry and Surface Rheometry
80
4.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.2
Bulk Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.3
Surface Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.4
Bulk Viscometry Results . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.4.1
4.5
Intrinsic Viscosity Calculations . . . . . . . . . . . . . . . . . 89
Surface Viscometry Results . . . . . . . . . . . . . . . . . . . . . . . 92
4.5.1
Surface Shear Modulus . . . . . . . . . . . . . . . . . . . . . . 92
4.5.2
Phase Lag Calculations . . . . . . . . . . . . . . . . . . . . . . 103
4.5.3
Shear Viscosity Results . . . . . . . . . . . . . . . . . . . . . . 107
4.5.4
Phase Change Discussion . . . . . . . . . . . . . . . . . . . . . 111
4.5.5
Eyring’s Model . . . . . . . . . . . . . . . . . . . . . . . . . . 112
5 Elastin Monolayers
115
5.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.2
Quasi-Static Methodology . . . . . . . . . . . . . . . . . . . . . . . . 116
5.3
Dynamic Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . 119
5.3.1
5.4
Analysis of the Oscillation Experiments Data . . . . . . . . . 121
Quasi-Static Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
5.4.1
Nanopure Water Subphase . . . . . . . . . . . . . . . . . . . . 123
ix
5.5
5.4.2
Ionic Subphases . . . . . . . . . . . . . . . . . . . . . . . . . . 127
5.4.3
Effect of pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
5.4.4
Relaxation Experiments . . . . . . . . . . . . . . . . . . . . . 140
5.4.5
Flory Analysis of Quasi-Static Data . . . . . . . . . . . . . . . 143
Dynamic Measurements . . . . . . . . . . . . . . . . . . . . . . . . . 146
5.5.1
Dynamic Elastic Moduli . . . . . . . . . . . . . . . . . . . . . 148
5.5.2
Temperature Effects . . . . . . . . . . . . . . . . . . . . . . . 150
6 Elastin-Lipid Interactions
154
6.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
6.2
Microscropy Methodology . . . . . . . . . . . . . . . . . . . . . . . . 155
6.3
6.2.1
Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
6.2.2
Microscope Trough . . . . . . . . . . . . . . . . . . . . . . . . 156
6.2.3
Fluorescence Microscopy . . . . . . . . . . . . . . . . . . . . . 157
Elastin Adsorption Results . . . . . . . . . . . . . . . . . . . . . . . . 157
6.3.1
α-elastin in Subphase . . . . . . . . . . . . . . . . . . . . . . . 157
6.3.2
Lipid-Elastin Π-A Measurements . . . . . . . . . . . . . . . . 164
6.3.3
Π-A Characteristics of Fluorescently Labelled Lipids . . . . . 174
6.3.4
Fluorescence Microscopy . . . . . . . . . . . . . . . . . . . . . 176
7 Conclusions and Future Work
179
7.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
7.2
Discussion Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
7.2.1
Rheometry Results and Discussion . . . . . . . . . . . . . . . 180
7.2.2
Elastin Monlayers Results and Discussion . . . . . . . . . . . . 180
x
7.2.3
7.3
7.4
Lipid-Elastin Interactions Results and Discussion . . . . . . . 181
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
7.3.1
Rheometry Conclusions . . . . . . . . . . . . . . . . . . . . . . 183
7.3.2
Elastin Monolayers Conclusions . . . . . . . . . . . . . . . . . 184
7.3.3
Lipid-Elastin Interactions Conclusions . . . . . . . . . . . . . 185
Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
7.4.1
Surface Rheometry . . . . . . . . . . . . . . . . . . . . . . . . 185
7.4.2
α-elastin monolayers . . . . . . . . . . . . . . . . . . . . . . . 186
7.4.3
Lipid-Elastin Interactions . . . . . . . . . . . . . . . . . . . . 187
Bibliography
187
xi
List of Figures
2.1
A schematic of elastin exon structure . . . . . . . . . . . . . . . . . .
7
2.2
The structures of Desmosine and Isodesmosine . . . . . . . . . . . . . 14
2.3
A schematic of the formation of a desmosine crosslink . . . . . . . . . 15
2.4
A schematic of the formation of an elastic fibre
2.5
Schematic of the formation of intermolecular bonds . . . . . . . . . . 23
2.6
Schematic of the formation of intramolecular bonds . . . . . . . . . . 23
2.7
Calcium binding to α-elastin at different pHs. . . . . . . . . . . . . . 28
2.8
The structure of PC and PS . . . . . . . . . . . . . . . . . . . . . . . 33
2.9
A Langmuir Trough . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
. . . . . . . . . . . . 16
2.10 An idealised Π-A isotherm . . . . . . . . . . . . . . . . . . . . . . . . 43
2.11 Arrangements of proteins on a liquid surface . . . . . . . . . . . . . . 44
2.12 Shear modulus shape change . . . . . . . . . . . . . . . . . . . . . . . 54
2.13 Schematic for the Newtonian definition of viscosity . . . . . . . . . . 59
2.14 Schematic for the definition of Poiseuille’s Law . . . . . . . . . . . . . 63
3.1
Titration curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.1
Schematic of a capillary viscometer . . . . . . . . . . . . . . . . . . . 81
4.2
Schematic of Surface Viscometer . . . . . . . . . . . . . . . . . . . . . 83
xii
4.3
Surface Rheometer . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.4
Flow time against concentration at 5◦ C . . . . . . . . . . . . . . . . . 85
4.5
Flow time against concentration at RT . . . . . . . . . . . . . . . . . 85
4.6
Relative viscosity at 5◦ C . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.7
Relative viscosity at RT . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.8
Intrinsic viscosity of α-elastin in water at 5◦ C . . . . . . . . . . . . . 90
4.9
G against frequency as pressure varies. . . . . . . . . . . . . . . . . . 94
0
0
4.10 G against frequency as pressure varies. . . . . . . . . . . . . . . . . . 94
00
00
4.13 G∗ against frequency as pressure varies. . . . . . . . . . . . . . . . . . 98
4.14 G∗ against frequency as pressure varies. . . . . . . . . . . . . . . . . . 98
0
4.15 G against pressure as frequency varies. . . . . . . . . . . . . . . . . . 99
0
00
00
4.19 G∗ against pressure as frequency varies. . . . . . . . . . . . . . . . . . 101
4.20 G∗ against pressure as frequency varies. . . . . . . . . . . . . . . . . . 101
4.21 Phase lag against frequency on water . . . . . . . . . . . . . . . . . . 104
4.22 Phase lag against frequency on calcium chloride . . . . . . . . . . . . 104
4.23 Phase lag against surface pressure on water . . . . . . . . . . . . . . . 106
4.24 Phase lag against frequency on calcium chloride . . . . . . . . . . . . 106
4.25 Shear viscosity against frequency on water . . . . . . . . . . . . . . . 108
4.26 Shear viscosity against frequency on calcium chloride . . . . . . . . . 108
xiii
4.27 Shear viscosity against surface pressure on water . . . . . . . . . . . . 109
4.28 Shear viscosity against surface pressure on calcium chloride . . . . . . 109
00
4.29 G against Π showing a phase change . . . . . . . . . . . . . . . . . . 111
4.30 Fitting to Eyring’s model, water subphase . . . . . . . . . . . . . . . 113
4.31 Fitting to Eyring’s model, 0.1 M CaCl2 subphase . . . . . . . . . . . 113
5.1
The quasi-static trough set-up . . . . . . . . . . . . . . . . . . . . . . 117
5.2
Sample isotherm plotted against time . . . . . . . . . . . . . . . . . . 118
5.3
Sample isotherm plotted against area . . . . . . . . . . . . . . . . . . 118
5.4
The dynamic trough set-up
5.5
Raw Π-A data from an oscillation experiment . . . . . . . . . . . . . 121
5.6
Quasi-static dilational moduli . . . . . . . . . . . . . . . . . . . . . . 123
5.7
Quasi-static dilational moduli . . . . . . . . . . . . . . . . . . . . . . 123
5.8
Π-A on NaCl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
5.9
Π-A on CaCl2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
. . . . . . . . . . . . . . . . . . . . . . . 120
5.10 Quasi-static dilational moduli . . . . . . . . . . . . . . . . . . . . . . 129
5.12 Second Cycle Π-A’s . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
5.13 An isotherm at pH 9.0 . . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.14 An isotherm at pH 3.5 . . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.17 An isotherm at on NaCl pH 9.0 . . . . . . . . . . . . . . . . . . . . . 137
5.18 An isotherm on NaCl pH 3.5 . . . . . . . . . . . . . . . . . . . . . . . 137
5.19 Dilational Moduli at pH 9.0 . . . . . . . . . . . . . . . . . . . . . . . 137
xiv
5.20 Dilational moduli at pH 3.5 . . . . . . . . . . . . . . . . . . . . . . . 137
5.21 An isotherm at pH 9.0 . . . . . . . . . . . . . . . . . . . . . . . . . . 138
5.22 An isotherm at pH 3.5 . . . . . . . . . . . . . . . . . . . . . . . . . . 138
5.25 Surface relaxation on water . . . . . . . . . . . . . . . . . . . . . . . 140
5.26 Surface relaxation on a CaCl2 subphase . . . . . . . . . . . . . . . . . 141
5.27 Surface relaxation on an NaCl subphase . . . . . . . . . . . . . . . . 141
5.28 Surface pressure decay on water. . . . . . . . . . . . . . . . . . . . . . 142
5.29 Surface pressure decay on CaCl2 . . . . . . . . . . . . . . . . . . . . . 142
5.30 Surface pressure decay on NaCl. . . . . . . . . . . . . . . . . . . . . . 142
5.31 Flory fit to a Π-A isothem . . . . . . . . . . . . . . . . . . . . . . . . 143
5.32 Linear fit to a Π-ε characteristic . . . . . . . . . . . . . . . . . . . . . 146
5.33 Dynamic surface pressure measurements . . . . . . . . . . . . . . . . 147
5.34 Dilational and Shear Elastic moduli . . . . . . . . . . . . . . . . . . . 148
5.35 Dynamic and QS Dilational moduli on water . . . . . . . . . . . . . . 149
5.36 Dynamic and QS Dilational moduli on CaCl2
. . . . . . . . . . . . . 149
5.37 Dynamic Dilational Moduli at different temperatures . . . . . . . . . 151
5.38 Π-A on water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
6.1
The fluorescence imaging trough . . . . . . . . . . . . . . . . . . . . . 157
6.2
Π-t of adsorped monolayer . . . . . . . . . . . . . . . . . . . . . . . . 158
6.3
Π-A of adsorped monolayer . . . . . . . . . . . . . . . . . . . . . . . 158
6.4
Π-t of an adsorbed α-elastin monolayer. . . . . . . . . . . . . . . . . . 161
6.5
Π-A of an adsorbed α-elastin monolayer. . . . . . . . . . . . . . . . . 161
xv
6.6
Dilational modulus of an adsorped α-elastin monolayer. . . . . . . . . 161
6.7
Pressure - time isotherm showing elastin insertion. . . . . . . . . . . . 164
6.8
Π-A of PC monolayer . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
6.9
Π-A of PC monolayer with α-elastin . . . . . . . . . . . . . . . . . . 166
6.10 Dilational Modulus from PC monolayer . . . . . . . . . . . . . . . . . 168
6.11 Dilation moduli from PC monolayer . . . . . . . . . . . . . . . . . . . 168
6.12 Π-A of a PS monolayer . . . . . . . . . . . . . . . . . . . . . . . . . . 170
6.13 Π-A of a PS monolayer with α-elastin . . . . . . . . . . . . . . . . . . 170
6.14 Dilational Modulus from PS monolayer . . . . . . . . . . . . . . . . . 171
6.15 Dilation moduli from PS monolayer . . . . . . . . . . . . . . . . . . . 171
6.16 Comparison of PC monolayer Π-A’s . . . . . . . . . . . . . . . . . . . 173
6.17 Comparison of PS monolayer Π-A’s . . . . . . . . . . . . . . . . . . . 173
6.18 Π-A’s of an PC:NBD-PC monolayer with α-elastin
. . . . . . . . . . 175
6.19 Π-A’s of an PC:NBD-PC:PS:NBD-PS monolayer with α-elastin . . . 175
6.20 Image of a PC monolayer . . . . . . . . . . . . . . . . . . . . . . . . . 177
6.21 Image of a PC monolayer after addition of α-elastin . . . . . . . . . . 177
6.22 Image of a 1:1 PC:PS monolayer . . . . . . . . . . . . . . . . . . . . . 178
6.23 Image of a PC:PS monolayer after elastin added . . . . . . . . . . . . 178
xvi
Chapter 1
Introduction
In this chapter I give a basic introduction to the experimental work and set out the
structure this thesis will take and give some detail of the contents of each chapter.
1.1
Introduction
Elastin is a protein of the extracellular matrix, in its normal state it forms resilient
elastic fibres found in many different tissues. The work in this thesis uses α-elastin, a
soluble form of elastin and examines its behaviour as a surface monolayer. α-elastin
is a soluble form of elastin but is similar in structure to the naturally occurring
precursor.
The behaviour of α-elastin in bulk has been examined by a variety of techniques
and its behaviour is fairly well established; however, the behaviour of elastin an
a monolayer has not been studied. Elastin is biologically significant not only for
its role in providing elasticity to a variety of tissues, but also because it is able
to self-assemble to form fibres. Elastin has also been implicated in the starting of
atheroscleroic build up in arteries.
In this work, experiments using bulk solutions of α-elastin were used to characterise the properties of the solution which could then be compared to the the
1
two-dimensional properties which were examined by applying the solution to waterair interface. Altering the subphase then allowed the effect of changing the inter
and intra-molecular interactions to be examined.
1.2
Thesis Outline
Chapter 2 covers the background to the project. Elastin is described in Section
2.2. The role of elastin in tissue and the forms it can take are detailed, followed
by a discussion of the specific form of elastin (α-elastin) that is used in the project.
The interesting phase transition that elastin in solution undergoes with increasing
temperature is looked at. Then there is a description of why examining the interactions between elastin and calcium, and elastin and lipids is interesting and of
physiological relevance. Then the deleterious effects of mutations and other problems with elastin are discussed. Finally the mechanical properties of elastin and the
uses of elastin as biomaterial are discussed. The Langmuir trough methodology is
described in Section 2.3, starting with the history of the technique. Then a variety
of different measurement methodologies are described. Then how to read a surface
pressure-area isotherm is described, detailing what different features of such a curve
indicate. Then the mathematics needed to analyse a monolayer is described last
in the section covering both the elastic moduli and the phase calculations. Section
2.4 details the background behind the viscometry measurements made in this work.
This begins with an introduction to viscosity, goes on to detail Poiseuille’s Law for
flow in a capillary tube, and also details the methods used to measure viscosity in
this work.
Chapter 3 describes the general methodology used in this work. The process by
which α-elastin is extracted from porcine aortas and purified is described in section
3.2. The other materials used in this experimental work are detailed in section
2
3.3. Finally general points relating to the use of a Langmuir Trough are detailed in
section 3.4.
Chapter 4 details the methodology used in the viscometry experiments and, the
results obtained from those experiments are discussed. The bulk viscometry methodology is outlined in Section 4.2. The surface viscosity methodology is described in
in section 4.3. The results of these experiments are then detailed and discussed;
bulk viscosity results are presented in section 4.4, and surface viscosity results are
presented in section 4.5.
Chapter 5 describes the methodology used in the Langmuir trough experiments
where an α-elastin monolayer was probed under quasi-static and dynamic conditions
on a variety of subphases; and then the results of the experiments are detailed and
discussed. This chapter is introduced in section 5.1. The methodology used for the
quasi-static experiments is described in section 5.2 and the methodology used for the
dynamic experiments is described in section 5.3. The results of these experiments
are then detailed and discussed; quasi-static results are presented in section 5.4, and
the dynamic results are presented in section 5.5.
Chapter 6 details the methods used to observe lipid-elastin interactions; the results of these experiments are detailed and discussed. This chapter is introduced in
section 6.1. The methodology used to image the lipid-elastin monolayers is described
in section 6.2. The results of the experiments are presented and discussed in section
6.3.
Chapter 7 is collected discussion of all the results detailed herein and gives possible
directions for continuation of this work. This chapter is introduced in section 7.1,
a summary of results and discussion is presented in section 7.2. Conclusions drawn
from these results are summarised in 7.3 and suggestions for future work are given
in section 7.4.
3
Chapter 2
Background
2.1
Introduction
This chapter introduces elastin and discusses the forms it occurs in naturally as well
as the experimentally useful forms. Some of the theories and methodologies used to
probe surface and bulk solution properties are also detailed.
The nature and role of elastin in a biological system is discussed in section 2.2.
This section discusses elastin in a general sense before going on to describe the
different forms elastin can take in section 2.2.1. This leads on to a discussion about αelastin, the form of elastin used in this work, specifically in section 2.2.2. The phase
transition α-elastin undergoes, known as coacervation is described in section 2.2.3.
Then the interactions between calcium ions and elastin is discussed in section 2.2.4
and the interactions between lipids and elastin in section 2.2.5. The mechanical
properties of elastin are discussed in section 2.2.7. Then the properties of elastin
as a biomaterial are discussed in section 2.2.8. This introduction concludes by
considering the effects that elastin mutations have in the body in section 2.2.9
The history and experimental uses of Langmuir monolayers is described in section 2.3, and viscometry is detailed in section 2.4. Lastly the behaviour of polymers
in solution is described in section 2.5
4
Section 2.3 discusses the Langmuir trough methodology and its theoretical basis.
The background of this methodology is described in section 2.3.1. What is shown
by a surface pressure - area isotherm is detailed in section 2.3.2. Then, looking
at information that can be obtained from surface pressure - area measurements,
the phases of a monolayer and how these phases relate to the arrangement of the
molecules on the surface are discussed in section 2.3.3. A general introduction to
the background and uses of surface rheology is given in section 2.3.4. Lastly the
elastic moduli of monolayers and the methods for calculating them are described in
section 2.3.5.
Section 2.4 discusses the methodology and mathematics used to measure viscosity.
Firstly viscosity is explained and defined in section 2.4.1, then Newtonian and nonNewtonian fluids, types of flow, and types of viscosity are discussed in section 2.4.2.
Poiseuille’s Law for flow in a capillary is explained and defined in section 2.4.3 which
also discusses the capillary viscometer, one method of measuring bulk viscoity. The
way viscosity affects a two dimensional system is discussed in section 2.4.4 which
also describe the oscillating ring viscometer, one method of measuring surface shear
viscosity.
Lastly the general behaviour of polymers in solution is described in section 2.5.
This section begins with a general introduction to the behaviour of polymers in
solution and how solutions form in section 2.5.1. The thermodynamics of a polymer
in solution are detailed in section 2.5.2.
2.2
Introducing Elastin
Elastin is a protein of the extracellular matrix, and in its fibrous form it is found in
tissues such as skin, blood vessels and lungs, where it is responsible for their longrange elastic properties. It provides resiliency to these tissues and the ability to
5
be be repetitively and reversibly deformed. This protein is found in all vertebrates
except agnathans (jawless fish) [1]. In most species a single gene encodes elastin
[2], one exception to this in Zebrafish [3]. In humans the gene for elastin is on
chromosome 7, it is made up of 2361 base pairs and contains 34 exons and introns
[4, 5].
This protein stops being produced at puberty and the elastin already present continues to function for the rest of the organism’s life, more than 70 years in the case
of humans. Two methods have been used to calculate the age of elastin, the first is
by examining
14
C content of lung tissue [2, 6]. The second method is racemisation,
that is, by studying the transition between chiral forms of amino acids. Data from
Aspartic acid was used [6, 7]; aspartic acid is (like all biologically active molecules)
only active in the l(evorotory)-form; it transitions to the d(extrorotatory)-form predictably over time if there is no tissue turnover. The concentration of d-aspartic acid
has been seen to be linear with age and thus there must be very little or no elastin
turn over in the life of a human. This demonstrates the remarkable toughness and
resiliency of this protein, far beyond that of any man-made material.
Elastin is a fibrillar polymeric material and is mainly composed of the amino acids
alanine, glycine, proline and valine [8, 9]. All four of these amino acids have relatively
small side chains, and in the whole protein there are very few permanently charged
side chains, which leads to a high degree of conformational freedom. The interactions
between the side chains are responsible for the cross-linking (discussed in more detail
in section 2.2.2 and especially figures 2.2 and 2.3). Elastin is an unusual protein as
it is amphiphilic; this is shown in figure 2.1. The hydrophilic regions are also the
regions in which the covalent cross-linking between the elastin monomers occurs.
These regions typically contain polyalanine sequences. The hydrophobic regions are
dominated by polypenta- or polyhexapeptide repeat sequences [10]. Overall elastin
is hydrophobic.
6
Figure 2.1: A schematic of the exon structure of human tropoelastin. The green and turquoise
domains contain the crosslinking regions while the white domains are the hydrophobic regions.
Human elastin does not have exons 34 or 35 which are found in other mammalian elastins. Image
taken from [2].
Figure 2.1 shows the exon structure of human tropoelastin (the elastin precursor
described is in more detail in section 2.2.1). The word “exon” is a shortened form
of “expressed region” and is the name used to describe the coding section of RNA
or a gene. The crosslinking domains (shown in green or turquoise) are differentiated
by whether they are enriched in proline (KP) or alanine (KA). The grey is the Nterminal end of the protein and the blue represents the C-terminal end (which will
be discussed in greater depth in section 2.2.1) [2].
The primary structure, the sequence of amino acids that make up the protein,
is known for elastin. The complete or partial genomic sequence of elastin from 29
different species is freely available as of September 2010; the sequence for human
tropoelastin is presented in [11]. Comparisons between the structures of different
species’ elastin are made in [12], and references therein, observing through comparisons the regions in the elastin sequences which have undergone more recent
duplication events compared with the rest of the sequence. It was also observed
that the the fundamental unit of duplication is a pair of cross-linking and hydrophobic domains. The secondary structure of elastin, the local organisation of the amino
acid chain, is not yet understood. Work carried out Debelle et al. [13] and, Green
et al. [14], much more recently, use near infra red Raman spectroscopy to assign
structural forms to the protein. The former work [13] found bovine elastin to con-
7
tain 10% α-helices, and 45% β-sheets with the rest undefined. The latter work [14]
observed the content of a relaxed dehydrated elastin fibre to be: 36.3% β-turns,
17.6% α-helices, and 46.1% unordered.
It is believed that the molecular and domain structure of elastin gives it its elastic properties, while the interaction between the hydrophobic regions providing a
restoring force (to be detailed further in section 2.2.7 below). It has also been
suggested that the hydrophobic regions are responsible for elastin’s ability to coacervate, a phenomena in which fibres are reversibly precipitated from a solution of
soluble elastin when the solution is heated. It seems that this ability to self-organise
plays a key role in the formation of elastin fibres in vivo. However, elastin fibres in
the body are not simply the product of elastin coacervation as they involve complex interactions with microfibrillar glycoproteins. Coacervation resulting in fibre
formation has been observed in both tropoelastin [15] and α-elastin [16, 17]; these
soluble forms of elastin are described in more detail in section 2.2.1. It seems likely
that the process of coacervation is necessary to the assembly of elastic fibres in the
extracellular matrix [18, 19]. The process of coacervation is discussed in more detail
in section 2.2.3.
Due to the complicated structure of even an elastin monomer, various simplified
elastin-like polypeptides have been studied [20, 21, 22, 23] and modelled [24]. In the
work of Li et al. [24] the model was of a pentapeptide repeat totalling 90 peptides. It
was used to simulate coacervation. The work of Abatangelo and Daga-Gordini [20]
looks at the binding of calcium ions to elastin peptides, while the work of Kaibara et
al. [25] studied the ion transport properties of an elastin peptide coacervate layer.
These are discussed in more detail in section 2.2.4. The work of Bellingham et al.
[21] looks at polypeptides which mimic various sections of human elastin. It was
shown that the only peptide which did not coacervate was one which contained no
hydrophobic domains (which supports the theory of coacervation discussed in section
8
2.2.3). It was also seen that increasing the NaCl content of the solvent decreased the
temperature at which the onset of coacervation occurred, and increased the amount
of coacervation at a given temperature. Increasing the concentration of the peptide
in solution also decreased the temperature at which coacervation occurred.
Elastin peptides are also interesting in their own right. They can be released
from mature elastin and have been shown to modulate the behaviour of various cells
including fibroblasts (which are responsible for extracellular matrix synthesis), endothelial cells (which line blood vessels), macrophages and monocytes (white blood
cells) and several cancer cell lines. The behaviour of elastin peptides in biological
systems is beyond the scope of this project but further details can be found in [26]
and references therein.
As mentioned above elastin is found in many tissues and in many different species.
Different research groups source their elastin from different species: cow [16], pig
([27], this work), horse [28], sheep [29], human [30]. The work done in Toronto
[21, 22] is based on human elastin. Work has also been carried out comparing elastins
from different species [1, 12, 31]. The work of Chalmers et al. [1] compares nine
types of elastin (pig, turkey, cow, salmon, dogfish shark, white shark, frog, turtle,
alligator) showing that elastin has become more hydrophobic up the evolutionary
scale. It is suggested that hydrophobicity controls the assembly of elastin fibrils, and
thus aids the assembly of elastin fibres, and also affects the mechanical properties
which are discussed in more detail in section 2.2.7. The assembly of elastin into
fibres is discussed in more detail in section 2.2.2. The work of Debelle and Alix
[31] compares the peptide structures of tropoelastin in six species (chicken, human,
cow, rat, sheep and mouse). Their comparisons lead towards a molecular model
for elastin, water-swollen elastin in particular. The work of He et al. [12] looks at
similarities and differences between the elastin genome of different species and also
compares their structure with that of other proteins such as spider silk and collagen.
9
2.2.1
Forms of Elastin
In tissues elastin forms polymeric cross-linked insoluble fibres, but it has other forms
as well. The soluble forms of elastin are useful as they can form a solution which
can be applied to any surface to form monolayers, which can then be probed by a
variety to techniques to gain further insights into intermolecular interactions. There
are four types of soluble elastin that are commonly used: tropoelastin, α-elastin,
β-elastin and κ-elastin.
Tropoelastin
Tropoelastin is the precursor form of elastin. It occurs naturally in the body, it is
the un-crosslinked form in which elastin is synthesised. It was first isolated in 1969
[32]. Tropoelastin has a molar mass of between 60 and 70 kDa depending on the
species of origin; mouse tropoelastin is the high end of this range and human the low
[2]. The amino acid composition is compared to that of mature elastin in [33], the
primary difference between the two is that tropoelastin contains more lysine than
mature elastin, where the lysine has formed crosslinks (described further in section
2.2.2). The tropoelastin does not have crosslinks. Tropoelastin is soluble despite
90% of the amino acids in each molecule being hydrophobic [34], this phenomena is
also observed in α-elastin and will be discussed in more detail in section 2.2.3.
Copper deficiency is known to prevent the crosslinking of both elastin and collagen. This is because the presence of copper ions is required for the activity of
the cross-linking enzyme (lysyl oxidase). Thus, it is possible to extract tropoelastin
from the blood vessels of copper deficient animals. However, currently tropoelastin
is usually synthesised in vitro by microbial biosynthesis [35]. In a non copper deficient animal elastin does not exist in the tropoelastin form for very long, as almost
as soon as it it produced it crosslinks to form fibres. This assembly process is not
10
completely understood. The coacervation of tropoelastin has been studied [36, 37].
The review paper by Vrohovski and Weiss [38] examines how coacervation is affected
by temperature, tropoelastin concentration, pH and NaCl concentration.
Particular attention has been paid to the carboxyl (C-) terminus of the protein
as it has been seen that this region (encoded by exons 29-36 in bovine tropoelastin,
cf figure 2.1 [39]) is necessary to tropoelastin self-organisation prior to cross-linking.
It is notable that the C-terminus of tropoelastin is one of the few hydrophilic regions
of the protein [40]. It was observed that targeting the C-terminus of tropoelastin
with a domain specific antibody results in a lack of elastic fibre formation. A similar antibody attack on the amino (N-) terminus of the same protein does not have
this affect on fibre assembly [41]. Further experiments have examined the sequence
exon by exon through the C-terminus, noting that deleting exons 31-36 does not
affect the ability of the protein to deposit on to microfibrils. However, deleting exon
30 (as well as exons 31-36) results in a large reduction in the amount of protein
associating in the extra-cellular matrix. Increasing the deletion to include exon 29
(thus replicating the earlier work) again resulted in no elastin assembly [39]. This
suggests that coacervation might not be the only driving force behind elastin fibre
assembly as previously thought. As detailed later (in section 2.2.3) coacervation
is hydrophobically driven and the deletions described above do not affect the hydrophobic regions of the protein [39]. The discovery of the importance of this region
of tropoelastin has led to further study and modelling of the structures formed by
this peptide sequence under different conditions [42].
The secondary structure of human and bovine tropoelastin was predicted by
Debelle et al. [43] using the GOR1 (Garnier-Osguthorpe-Robson) methodology as
18 ± 5 % α-helices, 63 ± 17 % β-sheets, 13 ± 13 % β-turns and 6 ± 6 % random
coil. This agrees fairly well with experimental analyses carried out by later by the
same group [13, 40] and by Green et al. [14].
11
Due to tropoelastin’s status as the elastin precursor it is a valuable test subject
for looking at how a variety of molecules may affect the behaviour, formation or
function of elastin in the body. It has been suggested that proteoglycans (PGs)
and glycoaminoglycans (GAGs) may participate in the formation of elastic fibres
and also may remain in the fibre [44] (see also the references therein). It has been
observed that in lysyl oxidase inhibitor (which reduces / prevents the formation of
elastin fibres, explained in more detail in section 2.2.2) treated animals the elastic
fibres that were produced had an abnormally high amount of GAGs present [45].
However, it has been observed that normal human arterial wall contains PGs and
the amount is higher than in atherosclerotic lesions [46]. Further evidence to support
the existence of PGs or GAGs in healthy elastin fibres is the work carried out by
Gheduzzi et al. [44].
α-elastin
α-elastin is derived from fibrous elastin made soluble by treating with oxalic (ethandioic)
acid. This is the form of elastin that was used in the subsequent experiments reported in this thesis. More information on the preparation of α-elastin, including
the extraction of fibrous elastin from tissue follows in sections 2.2.2 and 3.2.
Other Elastins
β-elastin is a lower molar mass elastin derivative that is also created in the production of α-elastin. It is separable from α-elastin by ultrafiltration or fractionation
by coacervate precipitation; this is possible because β-elastin does not coacervate.
β-elastin has a molar mass of around 6 kDa [8, 16].
κ-elastin is a protein derived from elastin. It is made soluble by treating the
elastin with potassium hydroxide. One preparation method is to hydrolyse fibrous
12
elastin, extracted from tissue as described in section 3.2, with 1M KOH in 80% v/v
ethanol-water at 37◦ C under constant mechanical stirring. The supernate, which
contains solubilised elastin peptides, is removed after the first hour of the hydrolysis
and replaced with an equal volume of the same solvent. This process is repeated for
subsquent hours (in the work cited [47] the peptides extracted at hour four were used
in subsequent experiments). After the supernate is removed it is neutralised with
percholoric acid at 4◦ C and then centrifuged to remove the perchlorate salts. The
post-centrifuge supernate was dialysed against distilled water and freeze dried [47].
The secondary structure of κ-elastin in dry powder form and in solution has been
examined by infrared Raman spectroscopy [13], and by Circular dichroism [40]. In
dry form the κ-elastin was 13% α-helix, 46% β-strand and the rest an undetermined
mix of β-turn and unordered elastin. In solution it was found that 47% was βstrand, 31% was β-turn and 22% was unordered. The peptides making up κ-elastin
have a molecular weights greater than 40 kDa [47]. κ-elastin has been observed to
have biogical effects on certain cell types [48] (and references therein); this will be
discussed in greater detail below in section 2.2.6.
Crosslinking in Elastin
Native elastin contains about 17 chains held together by four or more crosslinks.
For a while the source of the covalent cross-linking between the chains was a puzzle.
The only known source of crosslinking at the time was between cystine residues
but the cystine concentration in α-elastin is not high enough to account for the
bonding. Cystine only makes up 0.4% of the dry weight of the protein [8]. It wasn’t
until 1963 that more detailed analysis was done. Inital investigations revealed the
presence of two unknown compounds which were shown to be similar to each other
in composition but differed in their degree of unsaturation [49]. Further work by
the same group [50] resolved the structure of these two new compounds, desmosine
13
and isodesmosine, these structures are shown in figure 2.2.
(b) Iso-desmosine.
(a) Desmosine.
Figure 2.2: Structures of the amino acids desmosine (a) and iso-desmosine (b). Both structures
copied from [50] with slight alterations.
As mentioned in Section 2.2.1 copper definciency prevents elastin from cross linking, because copper is critical for lysyl oxidase activity, which is required to form
the desmosine or isodesmosine crosslinks. Lysyl oxidase catalyses the removal of the
-NH2 group from the lysine and its replacement with a =O. This converted form of
lysine is known as allysine, as it is an aldehyde. Figure 2.3 schematically shows how
a desmonsine crosslink is formed from a combination of allysine and unmodified lysine. Two allysine side groups can condense to form the bivalent adol condensation
product (ACP) (pathway 1 in Figure 2.3), while a lysine and an allysine can undergo
a Shiff base reaction to produce dehydrolysinonorleucine (dLNL) (pathway 2 in Figure 2.3). ACP and dLNL then condense to form desmosine (for more information
see reference [2] and references therein).
It was originally suggested [51] that fibrous elastin, like rubber, is a crosslinked
random network. However, in rubber every repeat unit has the potential to cross
link, while in elastin cross links only form from lysine residues. Only about 40 of the
800 amino acids making up tropoelastin are lysine, thus the probability that a given
residue can cross-link is 1 in 20. For a system of random tropoelastin molecules the
probability that a contact between a pair of molecules would involve a lysine in each
14
chain, and thus have the potential to form a crosslink, is 1 in 400. The probability
that a given chain would have twenty potentially crosslinking contacts in a random
system is (40/800)2 .(39/799)2 .(38/798)2 . . . (21/781)2 ≈ 10−57 . Given a molecular
correlation time (the time taken for a pair of molecules to achieve a new correlation)
of just 80 ns then it would take 1040 years to go through all of possible correlations.
This is the time scale for a system
of random molecular tropoelastin to
achieve just twenty potential links
and not all of the contacts will necessarily result in a crosslink forming.
Figure 2.3:
In reality, most of the forty lysine
A schematic of how a desmosine
crosslink forms from a mixture of allysine and lysine.
residues are involved in crosslinking;
therefore, there must be some form of
Image taken from [2].
molecular ordering involved in the crosslinking of tropoelastin to form elastin fibres.
Thus the system formed can not be a random network [52]. Coacervation is usually
believed to be responsible for this ordering.
Elastic Fibres
Elastic fibres in the body are made-up of an elastin core that is surrounded by a
mantle of microfibrillar proteins [38]. Each fibre is several microns in diameter [35]
while the microfibrils are 8-12 nm in diameter [53]. The fibrils, which define the
elastic network of a tissue, contain fibrillin but no elastin [53]. A schematic of the
elastic fibre assembly process is shown in figure 2.4, taken from [54] Images obtained
by Wachi et al. [55] show that the fibrillin and elastin are not in the same place and
that the formation of the fibrils occurs even when a mutated form of tropoelastin is
used which is unable to bind to them. Further details on elastic fibres can be found
in [54] and references therein.
15
Figure 2.4: Fibrillin is assembled pericellularly into arrays which are seem to mature into
crosslinked microfibrills. To create an elastic tissue (blue arrow) tropoelastin is then deposited
on the microfibrils and then the tropoelastin is crosslinked as described above. Image taken from
[54].
There has been debate over the function of the non elastin components of an
elastic fibre. In 2002 Urry et al. [35] assigned them no functional role. However, in
2007 Lillie and Gosline [56] concluded that the fibrils bear about 10% of the stress
applied to a fibre up to 50% extension. This suggests that it may be very tricky to
assign certain properties to the microfibrils if they only act over a certain range.
Structural Models of Elastin
As mentioned previously the structure of elastin is currently not fully resolved and
the way that structure relates to its properties is an area of much research. It has
been suggested [57] that the glycine content of elastin (around 27 g per 100 g dry
weight of protein [8]) is related to the flexibility of the polymer chains. It seems
that many elastomeric proteins share this feature, for example spider dragline silk
(around 42 g per 100 g [58]), so it appears likely that the shared structure of these
16
proteins is related to their shared properties.
Theoretical work to model elastin is focussed in two directions, a top down approach starting with the properties and trying to find a model that fits these properties (to be discussed in more detail in section 2.2.7); and a bottom up approach
which starts by looking at the peptide structure of elastin and trying to model the
properties from there. This approach only works because elastin is not a globular
protein [23]. In a globular protein the conformation of a particular section of the
polypeptide chain is dependant on the whole chain [59], while in elastin short peptide sequences are able to assemble in a manner similar to the intact sequence. Thus
elastin can be described as a fractal protein [23]. Sections of elastin which have been
examined include: Exon 5 of tropoelastin (a sequence of 11 peptide) produced fibrils
[23], Exon 30 of human tropoelastin (a sequence of 25 peptides) produced amyloid
fibres [60]; further work was carried out on shorter sections of this exon [61]. Exons
2-7 (the N-terminal) of tropoelastin were able to coacervate and form elastin-like
fibres [61], Exons 1-7 (the N-terminal with signal peptide) of tropoelastin formed
mainly amyloid fibres [61]; this is interesting given the work of Brown-Augsburger
et al. [41] mentioned earlier which indicated the N-terminal of tropoelastin was not
needed for elastic fibre formation. Exon 20 of tropoelastin formed a supra molecular
structure similar to that observed in elastin [62], and Exon 24 of tropoelastin was
seen to aggregate into short rigid filaments which then form bundles [62]. It appears
from the different aggregations produced there is more than one mechanism involved
in elastin’s ability to form fibres.
Circular dichroism studies on tropoelastin peptide sequences in a variety of solvents [61] have shown that the secondary structure changes depending on the solvent. Tropoelastin is largely in a β-strand structure (a sequence of amino acids
whose peptide backbones are extended) in water but changes to an α-helix (a hydrogen bonded coil where the bond links peptides spaced four apart) in aqueous
17
2,2,2-trifluroethanol.
Work focusing on one of the hydrophobic regions [33], showed that a β-turn
(a hydrogen bonded coil in which the bond links peptides which are spaced three
apart) structure with a small amino acid (often glycine) on the inside of the turn
and the bulkier amino acids, proline and valine, on the outside of the turn is most
favourable. This led to a structural model for elastin involving these β-turns as oiled
coils (detailed in section 2.2.7). The conformation of a tetrapepide sequence, found
many times in elastin, has been examined using molecular dynamics [63] and the
break-down of the secondary structure components of a variety of tropoelastins has
been predicted theoretically using the Garnier-Osguthorpe-Robson (GOR1) method
[31].
2.2.2
α-Elastin
α-elastin was first produced in 1951 [64]. It has has a molar mass of around 70 kDa
varying between species (the porcine α-elastin used in this work has a molar mass of
approximately 67 kDa [16, 17]). The amino acid content of α-elastin is very similar
to that of mature elastin [65], consisting predominantly of alanine, glycine, proline
and valine. It contains many more N-terminal (amine) residues than elastin as a
result of the splitting of the peptide chains [8]. Titration experiments were carried
out on α-elastin by Bendall [66] over a range of pHs between 2 and 12 and showed
that the there were also more free carboxyl groups in α-elastin than mature elastin.
Bovine α-elastin contains an average of 17 chains each with an average of 35
N-terminal residues on each [8]. α-elastin differs in structure from the soluble prefibrous form, tropoelastin which is secreted by cells before it is assembled into functional elastin fibres. This difference can be observed in the secondary structures of
the protein (tropoelastin in [13, 14, 40] discussed in section 2.2.1 and α-elastin in
18
[14, 65] to be discussed next).
The secondary structure of α-elastin has been examined by circular dichroism [65]
and Raman spectroscopy [14]. It was seen that there is an α-helix in the structure of
the protein when in solution, making up 29% of the elastin’s structure in the amide
III band [14], comparing to no α-helical structure being observed in κ-elastin [13]. It
has also been observed that the conformation of α-elastin in solution depends on the
solvent in which it is disolved, for example α-elastin in trifluroethanol has a much
more ordered structure than α-elastin disolved in water [67]. Likewise, circular
dichrosim measurements have shown that using a water-ethanol mix as a solvent
increases the order in α-elastin and the degree of order increases with the amount
of ethanol in solution [65]. In the same work ([65]) is was seen that a sulphuric acid
reduced the order in α-elastin, likely due to a combination of protonation of the
elastin by the acid (as described in [68]) and possible denaturing of the protein.
Raman spectroscopy of strained (60% strain) and relaxed elastin fibres under
hydrated and dehydrated conditions [14] shows that in the amide I band there is
very little difference in the secondary structure of a strained and unstrained fibre.
It was also observed that the difference in water content between the hydrated and
dehydrated fibres was not that great, 11% as opposed to 6%. There was also seen
to be a decrease in the β-turn fraction when the elastin fibres were hydrated, 24%
hydrated compared to 26% dehydrated. There was a similar decrease in the fraction
of unordered elastin and a slight decrease in the fraction of α-helix structure. The
differences in structure between stretched and unstretched fibres was much smaller
than this. It was observed that α-elastin became more unordered when it was
dissolved in water.
α-elastin contains many non-polar amino acid residues which should hinder its
ability to dissolve in a polar solvent such as water. It is however highly soluble in
cold water. It is believed that this is due to intramolecular associations occurring
19
between the non-polar parts of the protein and decreasing its area of hydrophobic
exposure. As the temperature increases the number of intramolecular bonds reduces
allowing the molecule to unfold so intermolecular bonds are able to form, which leads
to coacervation (discussed in more detail in Section 2.2.3).
α-elastin is charged and its observed charge in solution varies with pH due to
bound ions. It is negatively charged at high pH and positively charged at low
pH. Bendall observed that α-elastin has no bound H+ at pH 5 at 20◦ C [66]. The
isoelectric point (the pH at which a specific molecule carries no net electrical charge)
of α-elastin is pH 4.7 (at ionic strength 0.02 and 0◦ C) [16]. The isoionic point (the
pH at which a specific molecule has no net electrical charge and no adhering ions) is
pH 6.93 ± 0.82 in deionised water [27]. The difference between these measurements
reflects the fact that at the isoelectric point there are ions bound to the molecule
which affects the net charge observed from the molecule.
α-elastin as used in this work
The experimental work in this thesis used α-elastin derived from porcine aorta and
extracted using the methodology of Partridge et al. [16], which is further detailed
in Section 3.2. At the end of this process the α-elastin is a fine creamy-coloured
powder and is stable in its dry form. Solutions are then prepared using nanopure
water.
2.2.3
α-elastin coacervation
Coacervation is a reversible phase transition by which a dissolved material aggregates
and drops out of solution. It is sometimes referred to as ‘inverse solubility’ as it
exhibits the opposite temperature dependence to solubility, when compounds are
more likely to coacervate at higher temperatures rather than less likely. α-elastin
20
coacervates at around 30 ◦ C but this is dependant on concentration and various
other factors. Coacervation occurs when it becomes energetically possible for the
molecules of elastin to interact.
The coacervation of α-elastin has been studied since the protein was first identified [16]. The point of coacervation is easily identified as the transparent solution
becomes cloudy and opaque due to the elastin fibres being insoluble. Coacervation occurs when the temperature is raised past Tco (the coacervation temperature)
and reverses when the temperature is lowered below Tre (the reversal temperature).
Tco =Tre to ± 0.1◦ C [16]. When α-elastin coacervates it forms filaments 500 nm in
length. These filaments are able to self-organise to form fibres which are 10,000 20,000 nm in length and are similar to the native elastin fibre [17]. Unsurprisingly
there is a change in the conformation of the α-elastin during coacervation [69], it was
observed that 50% of coacervated α-elastin is in the α-helix conformation (cf α-helix
percentage of α-elastin in solution in section 2.2.2) and the circular dichroism curve
altered significantly.
Early work [52] examined the ability of various tropoelastin derived peptides to
coacervate. It was seen that a poly-tetrapeptide (Val-Pro-Gly-Gly)n did not coacervate even under conditions of high temperature, high concentration and large n up
to 40 repeats. A poly-pentapetide (Val-Pro-Gly-Val-Gly)n coacervated starting at
30◦ C, while a poly-hexapeptide (Val-Ala-Pro-Gly-Val-Gly)n was seen to precipitate
irreversibly. Increasing the hydophobicity of this polypeptide through the addition
of aromatic groups led to coacervation. This work [52] was used to study the molecular details of coacervation. Gheduzzi et al. [44] looked at the affects of heparan
sulphate, a linear polysaccharide which occurs as a proteoglycan and is responsible
for regulating a variety of biological processes including coagulation of the blood. It
was established, using solutions of tropoelastin, that heparan sulphate lowered the
temperature of the onset of coacervation.
21
As mentioned in section 2.2.2 it is believed that intramolecular bonds in α-elastin
break at higher temperatures which increases its surface area and renders it less soluble. Once the protein has ‘unrolled’ it is able to form intermolecular bonds which
allows it to form fibres. The degree of order of the α-elastin has increased in going from a randomly distributed solution to a fibre. Despite its amphiphilic nature
elastin is largely hydrophobic with around 60% of its amino acids having lipophillic
side chains. When exposed to these lipophillic regions, water forms a surrounding
hydrogen-bonded shell, this is known as clathrate water, shown in turquoise in figures 2.5 and 2.6. This shell is disrupted by the increase in temperature so water
is forced to interact with the side chains [17]. This assembly process is schematically illustrated in figures 2.5 and 2.6. Figure 2.5 shows in schematic form the way
a polypentapeptide associates during coacervation with the formation of inter and
intramolecular bonds. The peptides must be present in sufficent concentration (>
20 mg ml−1 ) for this process to occur, the temperature must also be greater than
30◦ C [52]. At lower concentrations the polypentapeptide was not observed to coacervate, likewise Urry observed [52] that the polytetrapeptide did not coacervate at
all; in both of these cases the solution exhibited a conformational change which is
schematically shown in figure 2.6. This conformational change occurs at 50 - 60◦ C
and involves the formation of intramolecular bonds. Thus, during the formation
of inter- and intra-molecular bonds the ordered clathrate water shell around the
hydrophobic side chains is disrupted causing a large increase in entropy.
As mentioned above in section 2.2.1, the study of the coacervation of tropoelastin
derived polypeptides by Urry et al. [52] led to analysis of the molecular details of
coacervation. It is believed that poly-penta and poly-hexapeptides have a β-spiral
secondary structure made up of a 10 atom hydrogen bonded ring (a β-turn). The
hydrogen bonds are considered to be dynamic with the probability of bonding between valines in the same peptide being 0.8 and between glycines 0.6. The resulting
22
β-spiral has hydrophobic and hydrophilic spiralling ridges. The formation of these
bonds is shown in figure 2.6. The hydrophobic ridges allow for intermolecular association by the disruption of clathrate water shells in the manner described and
illustrated above.
As mentioned earlier in section 2.2 the
work of Chalmers et al. [1] shows that
the hydrophobicity of elastin increases up
the evolutionary scale, which also implies
that the temperature at which coacervation occurs should decrease over the same
Figure 2.5: A schematic of the formation scale which would promote elastic fibre forof intermolecular bonds between hydrophobic
side chains which leads to coacervation, copied
from [52]. The grey regions are hydrophobic
mation. Comparing the most hydrophobic of their samples (turkey and pig) and
side chains which are shown surrounded by a
the least hydrophobic (white shark and
‘cage’ of clathrate water. The turquoise indi-
salmon) showed little difference in elastic
cated the region of clathrate water that be-
storage modulus suggesting that the elastin
comes disordered bulk water after the bond is
fibres, once formed, function in a similar
formed and coacervation occurs.
way.
Partidge and Whiting [70] showed that
blocked α-elastin was still able to coacervate, thus the charges of the C- and N-
Figure 2.6: A schematic of the formation
terminals are not required for coacerva-
of intramolecular bonds between hydrophobic
side chains (from [52]). The grey regions are
tion. In fact, blocking decreased the tem-
hydrophobic side chains which are shown sur-
perature at which coacervation occurred
rounded by a ‘cage’ of clathrate water. The
by 20◦ C between pH 6-7. Adding ethy-
turquoise indicated the region of clathrate wa-
lene glycol reduced the temperature at
ter that becomes disordered bulk water after
which the coacervation occurred by a fur-
the conformation change.
23
ther 20◦ C. It was postulated that this was
a result of an alteration of the structure of the water around the elastin as well as
changes in the α-elastin tertiary structure.
At an ionic strength of 0.01, α-elastin coacervates most easily at a pH between
5 and 6 and a temperature of 25◦ C. At a higher ionic strength (0.1) and lower pH
(2-3) α-elastin can be made to coacervate at temperatures as low as 20◦ C [16]. It
has also been observed coacervation can be completely inhibited at temperatures as
high as 50◦ C in a 1 mM solution of the ionic surfactant sodium dodecyl sulphate.
[71].
2.2.4
Effects of Calcium on Elastin
Background
Over the years there has been much interest in the effects of calcium on elastin. It
has been seen that the calcium content of elastic fibres in humans increases over time
[72, 73]. The work of Elliot and McGrath [72] looked at calcium distribution through
layers of the human aorta and found that in all layers the amount of Ca2+ increases
with age; this indicates the increase in calcium content is not just a surface effect.
The calcium is specifically bonded to the elastin within the fibre and the bond
is strong enough to survive sodium hydroxide extraction of the elastin from the
tissue [73]. The work of Gonçalves et al. [74] focuses on using ultrasound to observe
stenosis in the human carotid artery (a risk factor for stroke). They note that elastin
rich build-ups are transparent to ultrasound but those that are calcium rich are not.
This occurs as a result of calcium attenuation of the ultrasound. Interestingly, in
this case it is the ultrasound transparent deposits that are the ones with the most
risk associated with them. Why elastin rich build ups should be more dangerous is
not clear. It is possible that the elastin rich phase represents a later period of build
24
up.
As calcium ions are positively charged they naturally interact electrostatically
with the negatively charged regions of the elastin; binding of Ca2+ to ionised carboxylic groups has been observed [75]. It has also been seen that Ca2+ binding
to these groups is not dependant on the calcium concentration and changes the
conformation of α-elastin in solution. In fact, it is able to inhibit a concentration
dependant conformation transition [76].
In trifluroethanol it was observed that the enthalpy change of binding between
calcium and α-elastin is 1.9 ± 0.1 kcal/g ion, as this value is positive it indicates that
driving force for the binding must be of entropic origin [77]. The entropy change
evaluated from this and the binding constant is of order of 71 J/mol K. However,
the entropy change calculated in a binding process only refers to the species actually
involved in the binding thus changes in the organisation of the solvent (similar to
the disruption of clathrate water described earlier) would account for the entropy
increase as by binding there is a less ordered system created.
Early work in this area [78] showed that while calcium binding to elastin is linearly
dependant on elastin concentration (in the range 1-5 mg/mL) it is not linearly
dependant on calcium concentration. At low concentrations (below 1×10−4 M) there
is a linear dependence but as the concentration of calcium increases the amount of
bound calcium levels off, and above 2.8×10−4 M there was no increase in the amount
of bound calcium with increasing calcium concentration.
Neutral Site Binding
Given that elastin and calcium ions are both charged, and there is a pH dependence
of the binding of calcium to elastin [78], it seems logical to suggest that charged
groups are responsible for the binding. However, it has also been noted that [79, 80]
25
calcium ions are able to bond to a neutral site on the elastin and that this is a
possible mechanism for atherosclerosis. In order to ensure that any calcium binding
was taking place at uncharged sites α-elastin was dissolved in trifluroethanol and
acidified with trifluroacetic acid (it is known that using a strong organic acid in
an organic solvent leads dissolved polypeptides to have only neutral or positively
charged side chains [80]). In the work of Urry and Urry et al. [79, 80] it was shown
that the addition of the trifluroacetic acid did not change the circular dichroism
pattern of the α-elastin, which indicates that the changes observed are due to the
addition of the CaCl2 .
Later work by the same group [67, 81], continued to investigate this phenomenon.
It was shown that there are at least two conformationally distinct binding sites on
the elastin molecule [67]. It is possible to block the amino and carboxyl groups
to produce the N-formyl-O-methyl ester form of elastin [67, 81]. With the charged
groups blocked any effect of charge on binding is neutralised. The formation of the
N-formyl-O-methyl ester form of elastin is detailed in [82]. In the work of Starcher
et al. [67] it was shown that blocking of the carboxyl and amino groups does not
greatly affect conformation changes induced by the binding. However, there is a 10%
reduction in the amount of calcium found to bind to blocked elastin. It is suggested
that this is the proportion of calcium which does not bind to neutral site. Thus it
is believed that while a small fraction of the calcium ions bind electrostatically, the
vast majority of binding is at neutral sites. This leads to the charge neutralisation
theory for the calcification [79] of elastin. As the binding sites are electrostatically
neutral, adding calcium ions to them means that the fibre will become positively
charged. Space-charge saturation limits the number and proximity of the bound
ions. However positively charged calcium containing sites will attract negatively
charged ions such as phosphate (a major component of arterial calcification [83]) and
carbonate. When a given site is neutralised, neighboring sites are no longer limited
26
from binding calcium ions and thus are able to bind more ions. The affinity between
neutral sites and calcium ions is the driving force behind the fibre calcification.
If a calcium ion binds to the peptide backbone or is shared between two peptide
chains the fibre’s rigidity will be increased and its elasticity decreased, various suggested structures of peptide / cation complexes are given in [79]. It is known that
during arteriosclerosis (the first step in atherosclerosis) there is a loss of elasticity
which leads to the loss of the windkessel effect (the secondary pumping provided by
elastic recoil in the arteries) [84]. The theory that calcium binding always occurs
suggests that there should be a natural process to inhibit calcium binding in the
body and that when calcification does occur it should be chronic, with the amount
of calcification increasing with age. As already mentioned above, this agrees with
experimental observations [73].
Using equilibrium dialysis it has been seen that in trifluroethanol, calcium ions
bind to α-elastin in numbers that can only be explained if the interaction between
Ca2+ and α-elastin occurs at many sites along the protein and these sites do not
interfere with each other [77].
Conditions of Binding
As mentioned previously in section 2.2 work has been done investigating the binding
of calcium ions to elastin peptides [20]. In this work, α-elastin prepared in the
manner of Partridge et al. [16] was further hydrolysed with KOH; after purification
this was shown to produce several peptides, of which two, E I and E III, made up
10% by weight of the purified elastin. It was seen that the binding of calcium to
these peptides was pH dependant. Twice as much calcium was bound to E I at pH 8
as compared to pH 5. It was also observed that blocking the C-terminal of the petide
greatly reduced the binding of calcium to the peptide, which suggests that binding
to this peptide is largely electrostatic in nature. The same year the same group [75]
27
also investigated the binding to α-elastin. They observed very similar results, as
figure 2.7 shows approximately twice as much calcium bound to the α-elastin at pH
8 as at pH 5. The binding is also observed to increase with calcium concentration
which is in contradiction to what was observed by Molinari Tosatti et al. [78] and
Starcher et al. [67].
A separate investigation has also observed that binding of calcium ions to elastin
occurs preferentially above a pH of 5.5 [78], so the effect of pH on both salt solutions
and pure water was also examined. It has also been suggested that Ca2+ may induce
reversible conformational changes in elastin in a way that Na+ does not [85]; this
is an indication that the elastin-Ca2+ interaction is different from the interactions
with other ions. This means that the interaction between calcium and elastin is
likely to be complex one but at the moment its precise nature is unknown. It was
later observed [29] that when the surrounding medium used was a 50% methanol in
water mixture calcium binds down to pH 4.
The binding of calcium to αelastin apparently has some dependence on the solvent in which
the elastin is dissolved. In water,
equilibrium dialysis shows that
around 3 to 4 ions bind per 10000
residues. In trifluroethanol, however, there are around 12 bound
ions per 100 residues [77].
Figure 2.7: A graph showing the amount of calcium
In has been observed that al-
binding to α-elastin at two different values of pH. The
cohols (methanol, ethanol and
circles are pH 8 and the triangles pH 5. Graph taken
from [75].
propanol) increase the binding of
calcium to fibrous elastin [29]. In
28
a solution of 60% alcohol (there was little variation between the three alcohols examined) nearly six times as much calcium bound to the elastin as when plain water was
used. In the same work a similar effect is observed with α-elastin. When disolved in
water 20-30 µmol Ca2+ binds per gram, but when the solvent is 40% methanol-H2 O
over 200 µmol Ca2+ binds per gram.
Further work on conformational changes [29, 77, 85] looked at a variety of factors affecting the conformation of elastin in solution. The work of Terbojevich et
al. [77] which looks at the α-elastin in trifluroethanol system by observing changes
in the circular dichroism spectra of the system, saw that the onset of the conformational change is not directly proportional to the amount of bound calcium, but
instead depends on the protein concentration. With less elastin present conformation changes start occurring at lower relative concentrations of calcium. In order to
explain this, it was suggested that the concentration dependence of α-elastin’s conformational stability is due to the aggregation of elastin molecules in trifluroethanol.
Apparently this aggregation stabilised an ordered conformation of α-elastin, so that
a large amount of calcium binding is required to disrupt it and cause the conformational change. It seems the aggregation of α-elastin molecules forces the Ca2+
to bind to the surface of the aggregate. Thus, the core, which contains the helical
sections disrupted by this conformational transition, remains largely unaffected. As
the level of calcium binding increased there comes a point when the core is affected
and the elastin undergoes a conformational transition. It is also notable that this
same affect can be induced by adding water to the system. It is suggested [77] that
water affects the conformation of α-elastin by causing an initial dissociation of aggregate structures and, at higher concentrations, coordinates around the carbonyl
groups and causes the collapse of the ordered structure. It is also observed that
the association constant for water molecules with elastin is lower than that of Ca2+
with elastin. The work of Rucker et al. [29] examines the effects of methanol on a
29
system of α-elastin in water, also by circular dichroism. It was observed that the
order in the α-elastin increased with methanol fraction. The work of Hornebeck
and Partridge [85] used a column of elastin fibres and they observed that binding of
calcium ions to the fibres produces a change in the cholesterol binding affinity of the
fibres. It is suggested that a conformation change is induced by the calcium that
leads to the hydrophobic side chains being located at internal water interfaces which
leads to increased interaction with other hydrophobic molecules such as cholesterol.
Build up of cholesterol on elastin fibres in the body is part of the changes observed
in artherosclerosis.
Comparison with other ions
In order to investigate the differences between sodium ions binding to elastin and
calcium ions binding to elastin, Winlove et al. [27] performed titration experiments
and observed that the uptake of Ca2+ and Na+ was similar between pH 2 and
pH 12.5, which suggests that electrostatic interactions were dominating the binding.
They also showed, using radiotracers, that 4.5 µmole of calcium per gram dry weight
of the elastin was bound permanently to the α-elastin whereas none of the sodium
bound permanently, which again indicates a difference between the behaviour of
calcium and sodium ions. Earlier titration experiments by Hornebeck and Partridge
[85] had shown that using a buffer with calcium ions in, rather than with sodium
ions, slowed the passage of cholesterol through bovine aorta.
An investigation of the calcification of elastin fibres in vitro, in particular the
growth of hydroxyapatite (Ca5 (PO4 )3 OH) crystals, has established that although
the calcium content of the surrounding solution affects the rate of growth, it was
not absorbed at all from solutions with concentrations under 5×10−3 M [83]. Given
that in the human body the calcium concentration is 1×10−2 M in plasma [86], this
supports the binding of calcium to elastin in vivo. It has been seen that treating
30
elastin fibres with aluminium chloride prevents the build up of either calcium or
phosphate ions [87]. It is believed that this is due to an alteration in the elastin secondary structure as a result of the aluminium ion’s presence; infra-red spectroscopy
shows that the addition of AlCl3 induces alterations in the amide I band (1600 1700 cm−1 ). It is believed that this alteration is due to a reduction in the number
of non-classical β-sheet structures, which suggests that these structures are involved
in the binding of calcium ions to elastin. It had previously been shown that of 18
ions tested Ca2+ induced by far the greatest conformation change [67]. Like calcium
ions, phosphate ions bind to elastin much more readily at alkaline pHs [83].
Rucker et al. [29] compared the binding of ten metal ions to elastin fibres and
found that Cu2+ and Zn2+ bound in the greatest amounts; neither of these ions were
examined in the work of of Starcher et al. [67] which is discussed above. Neither
copper nor zinc occur in high concentrations in the body under normal conditions.
However, it was seen that of the ions examined calcium was the only one whose
binding was affected by changing the surrounding medium to a 50% methanol-water
mixture [29]. Later, Long et al. [88] reported that elastin peptides interacted with
divalent ions (Sr2+ , Ca2+ , Mg2+ ) selectively over monovalent ions (Na+ , K+ ).
Infrared spectra have been used to analyse the effect of various salts on coacervated N-formyl-O-methyl α-elastin [81]. Focussing on the region between 1400 and
1700 cm−1 the blocked α-elastin had three peaks: the amide I, a result of C=O
stretching; the amide II, a result of N-H deformation; and a non-amide deformation.
CaCl2 was seen to have a large effect on the amide I band, shifting it to smaller
wavenumbers. This shift was reproduced by CaBr2 and was thus an affect of the
calcium ion. The other salts tested caused the amide I band to move to slightly
higher wave numbers. It is concluded that the calcium binding site on the blocked
elastin is peptide carbonyl oxygen, a neutral site which fits with the work of Urry
et al discussed earlier.
31
Using a polypentapeptide, (which is capable of coacervating) as a simplified model
of elastin, as well as α-elastin allowed ion transport studies on membranes made
from α-elastin and polypentapeptide coacervates [25]. The Ca2+ ion is compared
to Mg2+ and Na+ . It was suggested that this particular pentapeptide sequence
is largely responsible for the elastomeric function and coacervate characteristics of
elastin [89]. It was shown that the polypentapeptide membrane behaved differently
to the α-elastin membrane in both NaCl and MgCl2 solutions. In both cases the
transport ratio of the positive and negative ions, and the transmembrane potential
differed between the pentapeptide membrane and the α-elastin membrane suggesting
that the transport properties of the membranes are different. However, the two
membranes behaved very similarly in CaCl2 solution, indicating a transport pathway
specific to the Ca2+ ion. The transmembrane potential was not seen to alter with
pH, in the range 4.5 to 7.5 which suggests that the Ca2+ pathway is controlled by
electrically neutral mechanisms. It was hypothesised that the selective binding of
Ca2+ ions to specific peptide carbonyl groups suppresses the mobility of the calcium
ions and so increases the Ca2+ content of the coacervate membrane [25]. It has
also been observed that the addition of Ca2+ induces alterations in the coacervation
temperature, and thus the self-assembly modes of α-elastin [90]. In this work it
was also found that Na+ promoted the coacervation of α-elastin under the same
conditions for which Ca2+ retarded it. Na+ promotion of coacervation has also been
observed in tropoelastin [38].
Coacervated α-elastin calcifies in a manner similar to elastin fibres. It has been
observed [91] that calcification is a bulk property of the coacervate and is not limited
to the coacervate-solvent interface. The calcification binds the protein together and
reduces its ability to de-coacervate or spread at an air-water interface.
32
2.2.5
Lipid-Elastin Interactions
Lipids
Lipids are amphillic molecules consisting of a polar head and usually two hydrophobic tails. Lipids vary with respect to their head groups, tail length and the degree of
saturation of the acid chains. Phospholipids, lipids with a phosphilated head group,
are a very important group of lipids as they form bilayers which is the basic matrix
of the plasma membrane.
The work discussed in this thesis only uses two types of
lipids and their fluorescently labelled counterparts. These
lipids are phosphatidylcholine (PC) and phosphatidylserine
(PS); the structure of these lipids are shown in figure 2.8.
These two lipids are both shown in their 16:0 form, meaning
that there are 16 carbon atoms in each tail and no double
bonds so the lipid chains are saturated.
Interactions between lipids and elastin
Tropoelastin is produced inside cells while elastin fibres are
assembled in the extracellular region. This implies that
tropoelastin must pass through the lipid membrane surrounding the cell. The mechanism by which this occurs
is largely unknown.
It has been seen that elastin fibres can accumulate lipids
Figure
2.8:
The
structure of phosphatidyl-
over time [84, 92, 93]. The work of Robert [84] shows the
choline (PC) on the left
binding of various lipids to elastin fibres in the aorta. It
and
was shown that cholesterol, cholesterol esters, triglycerides,
phosphatidylserine
(PS) on the right.
fatty acids and phospholipids all interacted with the aortic
33
tissue and were seen to bind more numerously on aortic tissue that already had
lesions. The binding of lipids to elastin fibres may be related to the accumulation
of calcium mentioned in section 2.2.4. This has been proposed as a starting point
for atheroscleroic build-ups in arteries [30, 73]. As mentioned above in section 2.2.4
during the first stages of atherosclerosis the elasticity of the elastic fibres in the
artery decreases, in the same stage there is an increase in the collagen / elastin ratio
and amount of lipid deposition [84]. A much earlier work by Kramsch [94] shows
that the lipid content of an atherosclerotic region of a blood vessel is approximately
three times higher than in a neighbouring normal region. A difference in the amino
acid composition of elastin from the two regions is also observed, the aspartic acid
components shows the biggest variation with an increase of 254% in the atherosclerotic region. The same work also shows an increase in the cholesterol content when
comparing between a newborn and adult in several different species, including humans. This agrees well with later analysis by Elliott and McGrath [72]. The work
of Kramsch [94] does not, however, show a general increase in the lipid content of
human elastin with age; in fact, the highest percentage was in newborns. However,
the oldest samples included in this work were betweeen 30-40 years so it is highly
probably that the lipid accumulation occurs after this age.
It has been seen in rabbits that injection of soluble elastin peptides can lead to
a reduction of athlersclerotic lesions induced by a high cholesterol diet [47] (and
references therein). Jacob et al. looked at the binding of cholesterol to κ-elastin
and to insoluble elastin. It was observed that the binding was dependant on the
cholesterol concentration and also, that adding calcium ions increased the amount
of binding. It was also found that the elastin fibres retain more cholesterol at high
temperatures. While the higher temperature (65◦ C) is not physiological reasonable
is serves to indicate that the the binding is driven by hydrophobic associations.
Work done by Hashimoto et al. [95] on rats fed an atherogenic diet with and
34
without chlorella phospholipid (a mixture containing PC, PS and several other phospholipids [96]) indicated that the addition of the chlorella phospholipid reduced the
amount of elastin and collagen found in deposits in the aorta below the levels found
in the control animals. They suggest that aortic degradation of collagen and elastin
was stimulated in the presence of chlorella phospholipid leading to the reduction in
deposition.
Some work has been carried out investigating the interaction between lipoproteins
and arterial elastin. It was observed that low-density lipoproteins bind most quickly
to elastin while high-density lipoproteins inhibit this process [30]. Investigations
by Lillie and Gosline [93] into the affects of lipids on the mechanical properties of
arterial elastin observed that adding sodium dodecyl sulphate (SDS) increased the
swelling of the elastin network, thus it is believed that the SDS is incorporated
into the elastin network. Analysis of the storage and loss moduli with frequency
show that making additions of SDS to the elastin network increased stiffness of the
elastin. At low concentrations it seems that this stiffening is simply a function of
the incorporation of additional species into the elastin network (they compare with
a reference solution of sucrose), however, at higher concentrations the stiffening
induced by the lipid is greater than that from the reference solution. Using their
data as a starting point they suggest [93] that loading an arterial wall with lipid only
has a minimal affect on the wall’s viscoelastic performance as simply not enough
stiffening occurs and it is unlikely that all the elastin in an artery would be affected
at once.
Work done by Vrhovski et al. [36] observed that the coacervation characteristic of tropoelastin during purification was very different from the coacervation
of pure tropoelastin. The impure material started to coacervate at a much lower
temperature (around 10 ◦ C rather than around 35 ◦ C). It was found that the principle impurity in the impure tropoelastin were fatty acids and that adding a small
35
amount of fatty acids to the purified tropoelastin reproduced the coacervation curve
of the impure tropoelastin. They suggest that impurities which are able to bind
to hydrophobic domains are able to artificially lower the coacervation temperature
as, if there are sufficient hydrophobic contaminants then the tropoelastin starts to
aggregate with the contaminants which encourages the onset of coacervation.
2.2.6
The Elastin Receptor
Elastin peptides have been discussed as a simplified model for elastin; however,
elastin peptides are found in vivo. They occur as a result of the degradation of
elastic fibres [97]. These peptides are able to influence cell function and induce
responses in the cells. Rodgers and Weiss [48] summarise previous work indicating
that a wide variety of peptides have been examined and seen to have a wide range
of effects on a wide range of cells.
Many of these peptides induce a chemotactic response in monocytes [48, 26],
which was taken as an indication that there was a membrane receptor for elastin.
Work by Hinek et al. [98] indicated one sub-unit (67 kDa) of the elastin receptor
was a significant galactoside-binding site, which is involved in the macromolecular
assembly of elastin into elastic fibres. This is sometimes referred to as the elastin
binding protein. Further work by the same group [99] indicated the other two subunits (61 and 55 kDa) were membrane-associated.
2.2.7
Mechanical properties of Elastin
Early work looked at the mechanical properties of blood vessels and attributed
these properties between the collagen fibres and the elastic fibres that make up the
blood vessel walls. The review paper by Burton [100] summarises this. The Young’s
modulus of elastic fibres, obtained from aorta, is 3×105 N/m2 , the tensile strength of
36
1×106 N/m2 and capable of undergoing a 100% extension. As noted earlier, elastin
is a major consituent of elastic fibres but not the only one. Thus the properties of
elastic fibres are related but not equivalent to the properties of elastin.
Elastin is an unusual protein as it has both high molecular mobility and insolubility. The mechanical properties of biological polymers are of great interest both for
biological reasons, how that material responds to stress in the body, and biomaterials
reasons.
The elastic properties of elastin are very important to its role in the body. Blood
vessels are able to expand as a pulse goes though. In expanding, elastic energy is
temporarily stored in fibres of the vessel wall. This energy is quickly returned to
the blood when the elastin recoils. The protein’s elastic recoil acts as a secondary
pump that forces blood downstream and helps smooth out pressure fluctuations
(the Windkessel effect) [1]. A number of studies have focussed on the mechanical
properties of elastin. Atomic force microscopy has been used to perform single chain
force-extension experiments on a polypentapeptide analogue of elastin [35] allowing
measurement of the elastin modulus of the chain.
The mechanical properties of elastin are affected by both temperature and the
water content of the fibres. By studying a variety of different elastins it was seen
that at low temperatures the water content of elastin is highest, and it reduces
with increase in temperature. At 40◦ C the water content was between 30% - 50%
depending on the species of origin [1].
Various models have been used to describe the elastic behaviour of elastin and
they will be outlined here but further details can be found in [38] and references
therein. It is generally believed that the recoil mechanism is entropic in nature, i.e.
stretching decreases the entropy of the system and the return to maximum entropy
causes the fibres to recoil again.
The Random Chain Model considers elastin to be like a typical rubber and made
37
up of random flexible chains with permanent crosslinks. When such a system is
stretched the chains are forced to straighten and align becoming more ordered,
thus decreasing the system’s entropy. This model does not take into account the
necessity of water in the elasticity of elastin. The Liquid Drop Model considers
water swollen elastin as a two phase system made up of globular regions (the oil
droplets) connected by crosslinks in water. Any deformation of the droplets exposes
hydrophobic regions to the surrounding water. The restoring force is given by the
entropy increase needed to keep the hydrophobic regions away from the water. The
Oiled Coil Model considers elastin to be made up of α-helices and hydrophobic ‘oiled
coils’ which, in a manner similar to the liquid drop model, are exposed to water when
the system is stretched and then are restored by the entropy increase required to
move the hydrophobic regions away from the water. The Fibrillar Model considers
elastin to be made up of α-helical crosslinking regions and a looser helix (the βspiral) which runs between the cross links. In this model the entropy alterations
occur due to oscillations in the β-spirals.
Debelle et al. [13] concluded from their Raman spectroscopy of the secondary
structure that bovine elastin was most closely approximated by the liquid drop
model. As previously mentioned later work [40] examined the secondary structure
of human elastin and found it to be very similar to that of the bovine elastin.
2.2.8
Elastin as a Biomaterial
Due to elastin’s toughness and elastic properties it is of interest as a biomaterial
[11]. Elastin is used in living skin substitutes to test the efficacy of burn treatments
[101]. Elastin is also interesting as biocompatible polymer possibly suitable for the
manufacture of biological grafts; some information in this area of research can be
found in review article [102]. It has also been suggested that elastin or elastin-like
38
biomaterials might be of use in making vascular grafts and elastic cartilage [11] (this
reference also contains further information on the topic). It has also been suggested
that elastin could be used to form a scaffold for tissue regeneration [62].
2.2.9
Biological Problems Relating to Elastin
Due to elastin’s important role in the body and its presence in many tissue types it
is unsurprising that mutations which change the properties of elastin usually have
deleterious effects. External factors can also have an adverse effect on elastin within
the body. As mentioned in section 2.2.1 a copper deficiency disrupts the formation
of elastin fibres. Also mentioned previously (in sections 2.2.4 and 2.2.5) the build
up of lipids on elastin fibres in the artery walls is believed to be the mechanism
behind atherosclerosis. While the mechanism behind this build-up is not entirely
clear, the problems associated with atherosclerotic plaques are well known. Work
has been performed looking at the arterial wall during ageing [84] and examining
the role elastin plays in these changes. Further discussion of arterial build-ups is
beyond the scope of this work and more information can be found in review article
[92] and references there in.
Two diseases caused by mutations in the elastin gene are supravalvular aortic
stenosis (SVAS) and Williams’ Syndrome [5], which are related. Both are very serious with SVAS possibly leading to heart failure if left untreated and Williams’
Syndrobe causing mental retardation and connective tissue abnormalities. Examination of the mutated tropoelastin caused by SVAS was carried out by Wachi et
al. [55] it was found that the mutated elastin did not self-associate as well as the
unmutated form.
Pseudoxanthoma elasticum (PXE) (mentioned in section 2.2.1) is a genetic disease
that is characterised by a progressive mineralisation of the elastic fibres [44].
39
2.3
2.3.1
Langmuir Monolayers
Development of the Langmuir Methodology
The behaviour of so-called “contaminated” surfaces has been examined for over
a century [103]. The core of the Langmuir trough methodology is that the surface
tension of a clean liquid surface does not alter when the surface area is changed, such
behaviour was described as “normal”. “Anomalous” or “contaminated” surfaces
have surface tensions which change with surface area. The Langmuir trough set up
allows the surface area to be altered while the subphase flows under the barriers so
pressure in the subphase is not altered. The lack of change in the surface tension of
a clean surface with area occurs because the molecular relaxation processes in such
a film are very fast; thus any alteration in the properties of the interface remain the
same as those of the surface at rest [104].
It was in 1917 that Langmuir’s theory [105] explained that oil films on water were
a single molecule in thickness. This meant that calculations on the size of the oil
molecules could be made. Insoluble films on the surface of a liquid subphase are
called Langmuir monolayers.
Langmuir Trough Methodology
A Langmuir trough, as illustrated in figure 2.9, is essentially a container for the
subphase which has hydrophobic barriers that can sweep across the subphase surface.
Anything on the surface is trapped by the barriers while the subphase passes under
them. As the barriers are moved the area available to molecules on the surface is
altered and thus the surface pressure changes. The barriers do not go to the bottom
of the trough so the pressure of the subphase does alter as the barriers are moved.
Most commonly water is used as the subphase, as it provides an ideally smooth
40
Figure 2.9: A side view of a Langmuir trough, showing the barriers (labelled A), the subphase (in
pale blue) and the monolayer (in dark blue). It can clearly be seen that the monolayer is trapped
between the barriers while the subphase is free to flow under the barriers.
surface. Fatty acids or lipids are most often the studied molecules (see review paper
[106] and references therein for further information). Analysis of surface pressurearea isotherms of lipids spread on a Langmuir trough has been a key technique in
the study of plasma membranes. More recently this technique has also been applied
to polymers including proteins [107]. Examples of recent work in this area have
been carried out by Cejudo Fernández et al. [108], which looks at mixed monolayers
containing β-lactoglobulin and monoglycerides and Cicuta [109], which looks at βcasein and β-lactoglobulin. Further details on analysing a surface pressure-area
isotherm are given in section 2.3.2.
Insoluble surface films represent the most basic model system (confining molecules
to two dimensions reduces the complexity of inter and intra bonding that can occur)
and quantitative measurements of their viscous and elastic response when subjected
to sheer deformation are possible with modern instruments [110]. The study of
biological elastic materials is an expanding area of research; thus it is natural that
the Langmuir trough methodology is applied to this area of research.
Elastin Films
The behaviour of elastin in a bulk solution has been investigated under different
conditions. Velebný et al. [71] examines the effect of ionic detergents with which
41
α-elastin forms complexes and work on coacervation, ionic interactions and lipid
elastin interactions have all ready been discussed in sections 2.2.3, 2.2.4 and 2.2.5
respectively. However, the behaviour of elastin films, essentially a two-dimensional
system has not been studied.
In this work, in order to better understand the intra- and intermolecular interactions that are important to the biophysical properties and physiological functions
of the protein, the behaviour of α-elastin at an air-liquid interface was studied. The
point of such measurements is to find the surface pressure, Π, as the surface area,
A, is altered.
As described in section 2.2.4 above there is currently much interest in the effects
of the Ca2+ ion on elastin.
2.3.2
Reading a Π-A Isotherm
Surface Pressure
In a Langmuir trough experiment the surface tension is measured as a function of
the available surface area. For an insoluble monolayer the surface concentration is
inversely related to the area A, Γ = M /A where Γ is the surface concentration and
M is the mass of material on the surface. The surface pressure, Π, is defined as the
resultant drop in surface tension as the surface concentration increases; Π = γ0 - γ,
where γ0 is the surface tension of a clean surface and γ is the surface tension when
a monolayer is present.
A Sample Isotherm
The Π-A graph shown in figure 2.10 and taken from a review by Kaganer et al. [106]
is a generalised case of a saturated fatty acid. A fatty acid, a molecule consisting
42
of a single long hydrocarbon tail and a (negatively) charged carboxyl head, is the
simplest type of amphiphile used for this type of experiment. The graph in figure 2.10
shows the various phases through which a monolayer can transition. The horizontal
sections in the isotherm represent regions of phase coexistence while the kink is a
continuous phase transition.
At very low pressures (area per
molecule >200 Å2 ) the molecules of the
monolayer are widely spread on the surface and have relatively little interaction
with each other. Molecule-surface interactions are dominant in this region.
This is known as the gaseous phase.
As the pressure increases, the monolayer changes into what has traditionally been called the Liquid Expanded
Figure 2.10:
An idealised fatty acid Π-A
(LE) phase (sometimes called the liq- isotherm taken from reference [106] which shows
uid disordered phase). In this phase the how the phase of a monolayer changes with presmolecules are much closer together but sure. Inset images indicate the orientation of the
still conformationally disordered. Fur-
molecules in each phase.
ther pressure increase takes the monolayer into the Liquid Condensed (LC) phase
(also called liquid ordered). In this phase the molecules align with their hydrophilic
heads in the subphase and their hydrophobic tails extending into the air. The transition between these two phases is of first order and regions of the more dense LC
phase form in the less dense LE phase. This will be shown visually with lipids in
chapter 6. At even higher surface pressures there is a final transition to the solid
phase. Despite the“liquid” / “solid” terminology, the liquid condensed phase does
not possess any more translational order than the solid phase. In both phases the
43
hydrophobic tails are aligned parallel to each other. In the liquid expanded phase
the tails are tilted relative the the subphase surface; while in the liquid condensed
phase they are at 90◦ to it. This orientation difference gives rise to the nomenclature
shown on figure 2.10.
The Π-A isotherm detailed above in figure 2.10 is, as noted, characteristic of
simpler amphiphiles such as fatty acid and lipids. α-elastin is a far more complicated
molecule with multiple hydrophobic and hydrophilic regions, thus the arrangement
of the protein on the subphase surface is much more complicated than the above
simplification. Figure 2.11, taken from [111], shows some possible arrangements for
a protein (β-casein) on a subphase surface and how those arrangements can change
with protein concentration. At low concentrations the protein is able to lay flat
on the subphase surface. As the surface concentration increases hydrophilic regions
of the protein are forced into the subphase. It is also possible, depending on the
solubility of the protein, for some of the molecules to be forced entirely into the
subphase. These orientations and the transitions between them allow for greater
complexity in the case of a protein monolayer.
As previously noted real systems are
often a lot more complicated than this
idealised graph and the phase changes
are not as easy to distinguish. A more
detailed description of the structure and
phase transitions in monolayers can be Figure 2.11: A schematic showing possible arfound in the review paper by Kaganer et
rangements of a protein (β-casein) on an aqueous
surface. At low surface concentrations the pro-
al. [106]. Using fluorescence microscopy tein is able to lie flat on the subphase surface, as
it is possible to image the region of co- the concentration of the protein increases first a
existence between the liquid expanded tail and then loops of the protein are forced into
and liquid condensed phases [112, 113]
the subphase. Image taken from [111].
44
(see figure 6.20 later in this work). Brewster Angle Microscopy can also be used to
distinguish monolayer phases [114].
2.3.3
Monolayer Phase Calculations
The physical properties of a polymer solution depend on the solvent, the concentration of the solute and the temperature. Generally, there are two types of solvent
defined and three ranges of concentration [115].
Solvent Considerations
The properties of a polymer in solution depend on the interaction energy between
the solvent molecules and the molecules of the polymer. For polymers solvents are
grouped into two categories: good solvents and poor solvents.
Good solvents have a strong attractive energy with polymers. When PolymerSolvent interactions are more favourable than Polymer-Polymer interactions, it is
energetically favourable for the polymer to dissolve over a wide range of temperatures. In a good solvent, the net interaction between polymer segments is negative.
This means that the excluded volume is large and positive; the polymer coil takes
up a larger volume as it is swollen by its interaction with the solvent [115].
Poor solvents are less able to accommodate polymer molecules. In a poor solvent
the Solvent-Solvent interactions are more favourable than Polymer-Solvent interactions. As a result of this a poor solvent is likely to allow a polymer dissolved in it
to precipitate if the concentration is increased or the temperature is changed.
The behaviour of a polymer in solution is temperature dependant. In some cases
a system may be in poor solvent conditions at low temperatures and good solvent
conditions at high temperatures; the temperature of transition between good and
poor solvent conditions is labelled Θ. At T = Θ the solvent is described as a Θ45
solvent [116]. In a Θ-solvent the polymer coils behave as ideal chains and assume
their random walk coil dimensions [115].
Concentration Phases of a Monolayer
The response of a monolayer depends on the concentration of the material on the
surface. Three concentration regimes in a good solvent are normally defined [115]:
dilute, semi-dilute and concentrated, although a fourth even more concentrated
phase, melt, does exist. In order to discuss the concentration of molecules on a
surface two parameters can be defined: the (dimensionless) surface area fraction
occupied by monomers (Φ); and the surface concentration, mass per unit area, (Γ);
these values are related to each other Φ = (Γ a2 ) / w; where a is the monomer size
and w is the mass of a monomer. Both of these concepts are useful but they are
used in different ways, Γ is the experimentally controllable value, while Φ is a useful
variable for theory. In general, w is known but a2 is not [117].
A dilute solution is one in which the polymer molecules are widely spaced, and so
Polymer-Polymer interactions are very limited. Each molecule on average occupies
a sphere of radius Rg . For an isolated polymer chain the radius of gyration is defined
as Rg ' aN ν ; where N is the number of monomers per chain and ν is the Flory
exponent. The physical properties of a solution are expressible as a power series with
respect to the concentration of the polymer. A monolayer of this concentration is in
the gaseous phase described in section 2.3.2 as in this form the monolayer behaves as
an ideal gas with Π = (Γ / m)RT ; where m is the molecular weight equal to N w, T
is temperature and R is the gas constant. This is the ideal gas law in two dimensions
[115]. A polymer chain in a good solvent can be modelled as a self avoiding walk
[118], and theoretical predictions indicate that ν = 0.75 in a two-dimensional system
in a good solvent [119]. Measurements made by on poly-(ter-butyl acrylate) [120]
with N ranging from 102 to 104 shows good agreement with this prediction.
46
The critical concentration, Γ∗ , occurs when the polymer chains are forced to
overlap. In a monolayer this concentration marks the transition to the semi-dilute
regime [118]. Similarly the area fraction at this point is the overlap packing fraction,
Φ∗ . If it is assumed that all available surface area A is covered by polymer molecules,
and each molecule has area Rg2 , then it follows that
Γ∗ =
M
A
(2.1)
where Γ∗ is the critical surface concentration and M is the mass of polymer on the
surface. The mass of polymer on the surface is the product of the monomer mass,
the number of monomers per chain and the number of polymer chains on the surface.
The number of chains on a completely covered surface is A / Rg2 . Substituting in to
this equation for Rg (using the above definition for the Flory exponent), and then
substituting back into equation (2.1) gives the N dependence of Γ∗ , Γ∗ = KN (1−2ν) ,
where K is equal to monomer mass / monomer size which is w/a2 . Thus, given the
above relation between Φ∗ and Γ∗ , it is seen that Φ∗ = N (1−2ν) [118].
The semi-dilute regime is characterised by large strongly correlated fluctuations in
polymer segment density which corresponds to the liquid phases described in section
2.3.2. Polymer monolayers in the semi-dilute regime are fluid and as a consequence
have low shear moduli (defined in section 2.3.5) and viscosity (defined in section
2.4.1 for a bulk system and in section 2.4.4 for a surface). In the semi-dilute regime,
Π’s dependence on Γ (or Φ) is stronger than linear.
The separation of polymer chains on a surface can be defined by a characteristic
length ξ. If ξ is large relative to Rg then polymer chains does not interact with
each other and thus the self avoiding walk model applies. For smaller values of ξ
the chain can be viewed as a series of ‘blobs’ each of which follows an ideal random
walk. At Γ∗ , ξ ≈ Rg and as the surface density increases, ξ reduces further [118].
47
If the polymer density is even higher then the fluctuations in density even out
and become small. This is the concentrated regime. The dividing point in surface
concentration between the concentrated and semi-dilute regimes is Γ∗∗ , at Γ∗∗ ξ
becomes of order a, the monomer size. At Γ∗∗ the surface is not completely covered
with polymer due to rearrangements of the polymer, thus Φ∗∗ is somewhere between
0.2 - 0.35. Above Φ∗∗ the pressure increases slowly with concentration [117, 118].
The points at which a monolayer changes regime can be observed from the elastic
moduli (discussed in detail in section 2.3.5 below).
A final transition can be seen at Φ∗∗∗ which is when the polymer becomes close
packed, at this point the monolayer becomes a two-dimensional polymer melt [117,
118]. Work by Daoud et al [121] examined the relationship between polymer concentration and ξ in bulk polystyrene solutions. It was seen that ξ decreases with
increasing concentration.
This classification into regimes is largely conceptual as the crossover between
regions is not sharp in a real system and the crossover concentrations are sometimes
hard to identify. However, work done by Monroy et al. [122, 123] using atactic
(randomly orientated constituents) polyvinylacetate with known molecular weight
and chain length, showed that Γ∗ and Γ∗∗ can be measured by measuring the surface
pressure-concentration dependence which shows clear gradient variations. It was
seen [122] that Γ∗ and Γ∗∗ are invariant with temperature (1 to 25 ◦ C), however the
value of the dilational modulus does reduce with temperature. Work carried out
by Spigone et al. [124], using PVAc, showed that Γ∗∗ is dependant on the speed
of compression applied on the monolayer but not on the molecular weight of the
polymer in the monolayer.
Π is proportional to Γ in the dilute regime (ideal gas behaviour) and proportional
to Γ2ν/(2ν−1) in the semi-dilute regime. Further work by the same group [120] also
looked at behaviour when Γ >> Γ∗∗ .
48
Flory Methodology
As mentioned above, an isolated polymer in a (dilute) solution occupies a sphere
of radius Rg ' aN ν . This equation becomes an equality in the condition that a
is equal to the Kuhn length (b). Kuhn’s model assumes that each monomer unit
behaves as a spherical particle of radius b, and that each particle is connected to its
neighbours via links of negligible frictional resistance For a very flexible polymer a
and b are likely to be similar [116]. The equilibrium properties of polymer films in
the semi-dilute regime are given by scaling laws with exponents related to ν. The
surface pressure scales as:
kB T
Πeq ' 2
Rg
Γ
Γ∗
yeq
(2.2)
where yeq = 2ν / (2ν - 1), this is the 2D relation. In a good solvent, the theoretical
value of ν means that yeq = 3.
For experimental reasons we calculate Λ, the area per microgram of applied
elastin, Λ = (Γ∗ )−1 . This means that equation (2.2) becomes:
kB T
Πeq ' 2
Rg
Λ
Λ∗
−yeq
(2.3)
Equations (2.2) and (2.3) are valid only in the semi-dilute regime [118]. The value
of the Flory exponent defines what type of conditions the chain is under: 0.5 is poor
solvent conditions, 0.57 is θ solvent conditions; 0.75 is good solvent conditions, 1 is
extended chain conditions for a two dimensional system [111, 125].
In the present work, the Flory exponents were calculated using equation (2.3) by
fitting a power law to a plot of Π against Λ. Generally the fit was only good below
Π = 5 mN/m. An example of this fitting is shown in figure 5.31. Measurements of
the Flory exponent have been carried out on a variety of polymers [107, 126, 118,
123, 127], these will be discussed in greater detail in section 5.4.5.
49
2.3.4
Surface Rheology
Introduction
Rheology is the study of flow and deformation of a material; surface rheology is
the study of flow or deformation of or on surfaces. Three-dimensional rheological
concepts may be applied to two-dimensional rheology with extra considerations. Due
to there being a surface in the system, the system has surface tension; therefore,
any stress tensor applied to the system must take this into account. A review of
surface rheology by Van den Tempel [104] highlights that the interface, being a twodimensional system can not have concepts generalised from three-dimensions easily
applied to it. The interface can not be an autonomous system as it only exists at the
boundary of a bulk phase; this coupling can be described by the tangential stress
boundary condition as described below.
Surface pressure in a Monolayer
In this work, rheological methods are applied to monolayers on the surface of liquids.
As discussed above these layers flow and have concentration related phases. Many
biological and chemical reactions occur at fluid interfaces; surfactants tend to adsorb
at interfaces forming monolayers. These monolayers are classified as either Langmuir
monolayers, where the molecule is insoluble in the neighbouring fluid, or Gibbs
monolayers, where the molecules are adsorbed from a bulk solution. In either case,
the surface pressure is defined in the same manner described above in section 2.3.2,
that is, as the reduction in surface tension caused by the addition of the molecules
to the surface. This can also be seen as a decrease in surface free energy per unit
area resulting from the adsorption of the monolayer. It is therefore possible to define
surface pressure as a thermodynamic quantity capable of describing the equilibrium
state of the adsorbed film [120].
50
If a small change in the monolayer’s area, δA(t), is considered, then there is a
change in the adsorption state of the film, which results in a change in the film’s
surface pressure.
δΠ(t) = Π(t) − Π0 =
∂Π
δA = −ε(t)u(t)
∂A
(2.4)
is the time dependant dilational modulus, which will be
where ε(t) = -A0 ∂Π
∂A
discussed in more detail in section 2.3.5 below, and u(t) = δA/A0 is the fractional
area change. Equation (2.4) is a generalised time-dependant response function, a
surface stress (δΠ) occurs as a consequence of an applied compression strain, u(t).
In the limit of constant elasticity modulus, equation (2.4) is equivalent to Hooke’s
law for a pure elastic 2D body [120].
The Surface Stress Tensor
Considering a flat surface located at z = 0 the general form of the surface stress
tensor is γ = γij (x, y), where i and j are either x or y. A force balance for a single
surface element is then calculated by considering that forces due to neighbouring
surface elements are balanced by forces due to liquid flow above (a) and below (b)
the element [104]. In the x-direction;
∂γxx ∂γyx
a
b
+
= (τzx
− τzx
)z=0
∂x
∂y
(2.5)
where τij is the shearing stress in the liquid, defined as the difference between the
equilibrium surface pressure and the instantaneous surface pressure, in non tensor
form τ = Πeq - Π(t). Naturally, a similar equation can be constructed in the y
direction. Under conditions of uniform dilation σyx = σxy = 0; and σxx = σyy = σ.
Thus equation (2.5) becomes;
∂vz
∂vz
∂γ
a ∂vx
b ∂vx
= −η
+
+η
+
∂x
∂z
∂x z=−0
∂z
∂x z=+0
51
(2.6)
where η is the viscosity (to be discussed later in section 2.4) and v is the velocity
of flow. Although equation (2.6) is usable, it can be simplified further. Under
conditions of uniform dilation in systems where surface resistance to shear is orders
of magnitude less than the resistance to area change, the second term on the left
hand side of equation (2.5) can be neglected. If this equation is applied to a gasliquid interface then it can be assumed that η a << η b . If the motion of the liquid
can be described as vx = vx (z), vy = 0, vz << vx ; that is, the motion is effectively
in the x-direction only. These conditions reduce equation (2.6) to
dγ
dvx
= ηb
|z=0
dx
dz
2.3.5
(2.7)
Elastic Moduli
The response of an isotropic, i.e. the same in all directions, two-dimensional film to
deformation is characterised by two elastic moduli. Changes in area are controlled
by the compression modulus ε, which is also called the dilational modulus. Changes
in shape and constant area are controlled by the shear modulus G.
Dilational Modulus
If the monolayer is elastic then the mechanical reaction to a quasi-static compression is proportional to the equilibrium dilational modulus (εeq ). The equilibrium
dilational modulus can be measured from the slope of the Π-A isotherm:
εeq =
∂γ
∂Πeq
=A
∂(lnA)
∂A
(2.8)
ε is also the inverse of the compressibility (β) of the film.
If the compression is not quasi-static then there may be a frictional resistance
to the compression. This resistance is described by the compression (dilational)
52
viscosity (ηd ), which is defined as:
ηd = A
Π − Πeq
d
A
dt
(2.9)
Equations (2.8) and (2.9) respectively define the real and imaginary parts of the
complex dynamic compression modulus.
Using equation (2.7) and the middle term of equation (2.8) the following equation
can be obtained,
dγ
∂γ d2 %
=
dx
∂(lnA) dx2
(2.10)
where % is the displacement of a surface particle in the x-direction.
It is also possible to define the (surface) dilational viscosity as [104]
ηd = ∆γ ÷
d(ln A)
dt
(2.11)
When a monolayer is being examined a peak in the dilational modulus, that is
a minimum in compressibility, indicates some form of phase transition. When the
molecules making up the monolayer move into a new arrangement the monolayer
becomes more compressible again. Spigone et al. locate the transition between
the semi-dilute and concentrated regimes in this manner in a poly(vinyl acetate)
monolayer [124].
Shear Modulus
As mentioned above the shear modulus describes a film’s response to shape deformation at constant area. A simple deformation in which area is conserved is shown
in figure 2.12. The initial shape has area A and side length L0 and is deformed under
the affects of τ1 and τ2 which are the values of the tension in orthogonal directions
and in the plane of the surface. The resultant shape still has area A but now has
side length L.
53
The in plane extension is defined as λ = L/L0 .
The surface shear (τs ) is defined as
τs = τ1 − τ2 = G
λ2 − λ−2
2
(2.12)
where G is the surface shear modulus and
(λ2 − λ−2 )/2 is the shear strain.
Equation (2.12) gives the relationship in an
elastic monolayer where surface shear is related
to the magnitude (squared) of the extension.
The shear modulus is non-zero in films which
possess some form of long-range order; such
films exhibit elastic, sometimes referred to as
Hookean, behaviour when equation (2.12) apFigure 2.12: The initial shape, with
side length L0 and area A, is deformed
plies. In films that are more ‘fluid’, that is, have
less order, the response is not conservative, and
by tensions τ1 and τ2 , the resultant
are related to the rate of deformation. In this
shape still has area A but the side
case the surface shear is defined as
length is now L.
τs = 2ηs
∂ ln λ
∂t
(2.13)
where ηs is the shear viscosity. This is the case in purely viscous cases. Most real
systems exhibit both elastic and viscous behaviour, which will be discussed in more
detail below [128, 129].
It has been noted that depending on the system studied the shear modulus can
be six orders of magnitude smaller than the dilational modulus [104], thus many
discussions neglect the shear modulus. The shear modulus is sometimes called the
transverse modulus [130].
54
Complex Moduli
In a standard Langmuir trough set up with non-quasi-static compression and expansion rates, the surface pressure rise is due to a combination of shear and dilational
deformations. The two components can be recovered by looking at the responses
parallel and perpendicular to the deformation applied to the film. It has also been
shown [131] in a system where the deformation is uniaxial, so the monolayer’s area
and shape are changed at a well defined rate, that anisotropy in surface pressure can
be detected via surface tension measurements performed with a pair of perpendicular
sensors. This effect can be used to measure the shear modulus of a monolayer.
For a sinusoidal area change
A(t)/A0 = 1 + δA(t)/A0 = 1 + ∆A/A0 cos ωt
(2.14)
where ω is the frequency of the applied area oscillation, also called the strain frequency, ∆A is the area change due to the oscillation, A0 is the equilibrium area and
t is time. If the response is linear then Π will oscillate with a phase shift relative to
the applied oscillation, which can be described as:
Π(t)/Π0 = 1 + ∆Π/Π0 cos(ωt + ϕ)
(2.15)
where Π0 is the equilibrium pressure and ϕ is the phase difference.
If the monolayer has a purely elastic response then the pressure oscillation is in
phase with the area oscillation; if the monolayer has a purely viscous response then
the area and pressure oscillations will be 90◦ out of phase. The frequency dependent
complex dilational modulus can be defined as:
ε∗ (ω) = ε0 (ω) + iε00 (ω) = ε0 (ω) + iωηd (ω)
(2.16)
Here the real part of the equation is the elastic storage component ε0 = |ε∗ | cos ϕ;
while the imaginary part is the dissipative component ε00 = |ε∗ | sin ϕ. Cicuta and
55
Hopkinson [132] examined the behaviour of PVAc, β-latoglobulin and β-casein and
observed that ε0 scaled with Γy , which is the same scaling as Π as shown in equation
(2.2) above. However, ε00 scaled as Γ2y .
Equation (2.16) offers a way to calculate the dilational viscosity from the complex
dilational modulus and thus obtain the frequency variance of the viscosity, this will
be discussed in more detail below in section 2.4.4.
The shear elastic modulus G is defined as the ratio between the change in shear
stress in response to a change in shear strain. The shear elastic viscosity (ηs ) is the
ratio between shear stress and the rate of shear. In a similar way to equation (2.16),
an applied oscillation can be used to measure the complex shear modulus:
G∗ (ω) = G0 (ω) + iG00 (ω) = G0 (ω) + iωηs (ω)
0
(2.17)
00
where G is the storage component of the shear modulus and G is the dissipative
component. A monolayer that behaves in a liquid-like manner is expected to have
0
00
G = 0 and a finite value for G ; similarly a monolayer displaying a finite shear
elastic modulus is said to be solid-like [124].
As detailed by Cicuta and Terentjev [107] it is possible to determine ε∗ and G∗
from the time dependant stress response Π(t). This is possible because under a
uniaxial compression in a Langmuir film both compression and shear deformation
are exerted. It was shown that for deformations of the form described in equation
(2.14) the pressure response can be expressed as:
Πk − Π0 = δΠk (t) =
∆A 0
[(ε + G0 ) cos ωt + ω(ε00 + G00 ) sin ωt]
A0
(2.18)
Π⊥ − Π0 = δΠ⊥ (t) =
∆A 0
[(ε − G0 ) cos ωt + ω(ε00 − G00 ) sin ωt]
A0
(2.19)
where Πk is the surface pressure measured parallel to barriers and Π⊥ is the surface
pressure measured perpendicular to the barriers. This methodology is discussed
56
further in section 5.3.1 which describes how equations (2.18) and (2.19) are used to
measure rheological properties of a monolayer.
The work of Lucassen and Van den Temple [133] was an early observation of surface behaviour using an oscillatory barrier technique and assuming that the shear
modulus of the surface (aqueous decanoic acid of varying concentrations) is negligible. Their value for the limiting adsorption at high concentration compares well
with the value obtained from equilibrium measurements showing that dynamic and
quasi-static experiments on the same system do yield similar results if shear is treated
correctly. Knowing the dilational modulus as a function of frequency allows predictions of the importance of the Marangoni Effect, that is mass-transfer along an
interface induced by a surface tension gradient in the interface, in that system. The
0
00
work of Spigone et al. [124] looks at the frequency dependence of G and G , and
also shows the pressure variation in these components of the shear modulus.
Viscoelasticity
In classical elasticity, stress is proportional to strain in the linear regime (Hooke’s
Law) but independent of the rate of strain; this applies to the idealisation of a
perfectly elastic solid. In this case, the stress and strain observed in the material
would be completely in phase. Conversely, in hydrodynamics, stress is proportional
to the rate of strain but independent of the amount of strain (Newton’s Law), this
applies to a perfectly viscous liquid. In the ideal case the stress and strain would
be 90 degrees out of phase. A real material will deviate from these ideal cases.
There are two main types of deviation. Firstly, the strain (in a solid) or rate of
strain (in a liquid) may be related to the stress in a more complicated manner
than simple proportionality, which occurs when the proportional limit is passed in
a solid. Secondly, the stress may depend on both the strain and the rate of strain
simultaneously, this is known as a time anomaly. A material that exhibits a time
57
anomaly is showing characteristics of both liquids and solids, and this behaviour is
called viscoelastic [134].
There is usually a phase lag between the stress and strain in a surface film as
it is a real rather than ideal material. This lag is designated δ and can be used to
describe the response of a monolayer. It is defined as
00
tan δ =
G (ω)
G0 (ω)
(2.20)
If δ is less than 10◦ then the monolayer’s response is elastic; if δ is greater than
75◦ then the monolayer’s response is viscous; a value of δ between these extremes
indicates that the monolayer is viscoelastic [128].
Yoo and Yu [130] looked at the temperature variation of the viscoelastic properties of PVAc and Poly(n-butyl methacrylate) (PnBMA) films. It was observed
that when the shear elastic modulus was accounted for there was a difference in its
temperature dependence between the two materials. As temperature increased the
maximum possible shear modulus for PVAc decreased, becoming negligible, while
that for PnBMA increases. They suggest that the latter result is anomalous and
that the reason for this is a slow transverse relaxation process in the PnBMA which
may be related to orientation differences between the two polymers. Millar et al.
[135] compared the dilational elasticity and the dilational viscosity of a lipid (DPPC)
monolayer as as function of surface pressure and observed that they showed a minimum at the same pressure.
Calculations
Calculation of εeq was done using a computer programme written by James Thompson [136] that smoothed the data and then calculated the differential of the smoothed
data. This was necessary to eliminate the fluctuations in the pressure. Both compressions and relaxations were analysed in this way.
58
2.4
2.4.1
Viscometry
Introduction to Viscosity
Viscosity arises due to the frictional forces Ff between
adjacent elements of fluid which are moving with different velocities. In order to qualitatively define viscosity we consider two neighbouring fluid volume elements as shown in Figure 2.13. The frictional force
between the two elements is proportional to the velocity gradient du/dx and dA (the area of contact between the two volume elements). This gives an equa-
Figure 2.13: Schematic show- tion where the constant of proportionality, η, is the
ing how the Newtonian viscos- viscosity;
ity is defined. dA is the area of
Ff = η(du/dx)dA
(2.21)
contact between the two volume
elements, dx is the distance be-
Equation (2.21) is the Newtonian definition of vis-
tween the two elements and du is
cosity. In the terms of equation (2.21) the larger the
the velocity of the upper element
velocity gradient du/dx the greater the frictional force.
relative to the lower element (reproduced from [137]).
If a liquid flows along a tube then the wall of the
tube is stationary. As attractive forces exist between
the wall of the tube and the liquid, the liquid volume elements which are adjacent
to the wall have vanishing velocities (no-slip boundary condition). The volume
elements which are next to these are able to reach a slightly higher velocity, and
so on. Thus, the maximum velocity occurs in the centre of the tube and there is a
velocity gradient across the tube’s diameter.
59
2.4.2
System Considerations for Viscosity Measurement
Newtonian and non-Newtonian Fluids
The viscosity as defined by equation (2.21) is a constant of proportionality and so
independent of flow velocity. This is not always true. In some liquids the presence
of a flow orientates the constituent molecules and so alters the forces between adjoining elements. This orientation effect occurs most frequently in liquids with very
asymmetric or easy to deform molecules or when there is a high velocity gradient.
In this latter case molecules align to reduce their resistance to flow. Liquids which
show this effect are said be non-Newtonian while those that do not, i.e. those that
obey equation (2.21), are said to be Newtonian. Examples of non-Newtonian fluids
include molasses, blood plasma, latex paint and cornflour in water.
Newtonian flow can be observed in all macromolecular solutions when the velocity gradient is small. The thermal motion of the molecules in the solution then
ensures random orientations of the molecules. Even when a fluid is observed to be
non-Newtonian the Newtonian viscosity can be evaluated by extrapolating to zero
velocity gradient.
A clean surface of a Newtonian fluid has no rheology as the surface pressure is
not affected by movement within or of the surface [104].
Types of Flow
When a fluid flows along a tube with laminar flow (Reynolds number less than 2000),
it is as if the tube were made up of many very fine parallel tubes set in the direction
of flow as the fluid in one ‘tube’ always stays in that ‘tube’. When the fluid velocity
is high or there are obstacles to the flow then the flow pattern is disturbed. Large
masses of fluid move as a unit with with both linear and rotational velocity, and so
vortices form. This disturbed flow is said to be turbulent.
60
Viscometry experiments with large scale flow generally requires that the flow be
laminar. The velocity of one volume element relative to its neighbour must be known
and as the flow patterns are much more complex in turbulent flow it is much easier
to do these measurements on laminar flow systems.
Types of Viscosity
For a Newtonian fluid it is possible to define two related types of viscosity. First there
is dynamic viscosity (also called absolute viscosity) which is what has been defined
above in section 2.4.1. The second form of viscosity is the kinematic viscosity (v)
which is obtained by dividing the absolute viscosity by the fluid’s density; v = η/ρ;
where v is the kinematic viscosity, η is the dynamic viscosity as defined above and ρ
is the density of the fluid. Due to the measurement methodology used in this work
(described below in section 2.4.4) the kinematic viscosity becomes a more useful
value. Kinematic viscosity has SI units of m2 s−1 compared to dynamic viscosity
which has SI units of kg m−1 s−1 . Surface viscosity can also be defined and is detailed
below in section 2.4.4.
For a solution the viscosity is dependant on the concentration of that solution.
If η0 is the viscosity of the pure solvent and η is the viscosity of the solution,
then the relative viscosity of the solution is defined by the ratio η/η0 . The specific
viscosity of the solution (ηsp ) is defined as (η - η0 )/η0 . It can be supposed that
the more particles of solute in the solution per unit volume the greater the energy
dissipation. Therefore, the reduced specific viscosity of a solution is defined as (η η0 )/(η0 ρ1 ) where ρ1 is the mass concentration of the solute. The reduced specific
viscosity is a characteristic of the solute in that particular solvent, however, it fails
at high concentrations of solute particles due to the interactions between them. At
infinite dilution (an ideal state) there is no interaction between the solute molecules,
at this point the intrinsic viscosity ([η]) is defined. The intrinsic viscosity is the
61
limiting value of the reduced specific viscosity. The intrinsic viscosity depends on
the molecular weight, volume and shape of the solute particles [116].
The intrinsic viscosity of a macromolecular solution can be determined by measuring the viscosity of a range of solutions of different concentrations and then plotting
the reduced specific viscosity against the mass concentration. At low concentrations
the resulting plot of ηsp /ρ1 is linear in ρ1 and the relationship can be written as
ηsp /ρ1 = [η] + kρ1 [η]2
(2.22)
k is the dimensionless Huggins constant which is a quantitative measure of the
intermolecular interactions [138].
If the solute particles are highly asymmetric then viscosity of the solution will
depend on the rate of shear. It is found [116] that the ratio of intrinsic viscosity
at finite shear rate to the intrinsic viscosity obtained by extrapolating to zero shear
rate depends on the rotational friction coefficients of the particle and the shear rate.
In compressible or non-Newtonian fluids other types of viscosity are defined but
these are not relevant to this work.
2.4.3
Poiseulle’s Law
Consider a uniform tube with a radius a and length l through which liquid is flowing
due to uniform excess pressure P , as shown in the left half of figure 2.14. It is
assumed that P is small enough that the flow is laminar. At the moment P is first
applied the liquid accelerates, the frictional force opposes the acceleration and a
steady state is reached. It is this steady state that is being analysed. As explained
above the flow velocity is zero at the tube wall and a maximum at the centre.
Symmetry considerations show that the velocity must only be a function of r the
distance from the centre of the tube. As liquids are largely incompressible, the flow
62
velocity must be the same at all points along l otherwise there would be density
alterations within the liquid.
A cylindrical volume element (as shown on
the right side of figure 2.14) of width dr and
radius r within the tube, is effectively isolated
from the rest of the liquid as the flow in the tube
is laminar. This volume element is adjacent to
two other elements: one inside with a contact
area of 2πrl; and one outside with a contact area
2π(r + dr)l. The fluid element outside is mov-
Figure 2.14: Schematic showing a
ing slower than this element, while the element uniform tube of radius a and length l.
inside is moving faster. The frictional forces at P is the uniform excess pressure causing
these contact areas are acting in opposite direc- fluid flow in the tube. r is the distance
tions. At steady state, the sum of the frictional of a volume element from the centre of
forces must equal the applied force. The applied
the tube and dr is the width of the volume element. Image reproduced from
force is equal to the product of the pressure and [137].
the cross-sectional area, while the fictional force is given by equation (2.21), so:
du
du
2πP rdr = −2π(r + dr)lη
+ 2πrlη
(2.23)
dx r=r+dr
dx r=r
where η is the viscosity of the liquid.
From equation (2.23) by using the Taylor expansion and integrating [137] it can
be shown that
u=
P 2
(a − r2 )
4lη
(2.24)
which shows that the velocity profile in a capillary tube is parabolic in form.
Equation (2.24) gives a value for the liquid’s linear flow velocity but a more
experimentally convenient property to calculate is the volume rate of flow. The
volume crossing any cross-section per second is different for each volume element,
63
but in each case is equal to the velocity multiplied by the cross-sectional area, thus
equalling 2πur dr. Substituting this value into equation (2.24) and then integrating
over all volume elements gives:
Z
πP
U=
2lη
a
(a2 − r2 )rdr =
r=0
πP a4
8lη
(2.25)
where U is the rate of flow in cubic centimetres per second. Equation (2.25) is
known as Poiseuille’s Law.
The Capillary Viscometer
This method used to measure bulk viscosity involves measuring the time required
for a known volume to flow through a capillary under gravity. The experimental
method is described in section 4.2 while this section covers the mathematics required
to analyse this data. A capillary viscometer is a U-shaped tube in which flow in
one arm is restricted by a capillary. Liquid is placed in the capillary and displaced
from a measured equilibrium; the time taken for it to flow back to equilibrium is
measured. A schematic of a capillary viscometer is shown in figure 4.1 in section
4.2.
The time taken for the liquid to flow is t. The fixed initial and final heights are
h1 and h2 respectively, the volume rate of flow varies as h changes and P also alters
from ρgh1 to ρgh2 (where ρ is the density of the liquid and g is the acceleration due
to gravity). A volume dv flows through the tube in a time dt when the height of the
liquid in the tube is h. The volume is the volume flow rate times the time taken,
so dv = U dt. Using Poiseuille’s Law (equation (2.25)) gives a value for U , the time
taken for the liquid level to fall from h1 to h2 is:
Z
h2
t=
h1
dv
8ηl
=
U
πgρa4
64
Z
h2
h1
dv
h
(2.26)
The integral on the right hand side of equation (2.26) is a constant of the viscometer, as are l and a. These constants may be determined by running a liquid of
known viscosity and density through the viscometer. Also this gives a simple way
to measure the kinematic viscosity of a liquid.
Πga4
K=
8l
Z
h2
h1
dv
η
=
h
ρt
(2.27)
As defined above, in section 2.4.2, η/ρ is the kinematic viscosity. As will be
detailed in section 4.2 equation (2.27) allows the calculation of viscosity from a flow
time.
A capillary viscometer may also be used to calculate the reduced specific viscosity
(ηsp ).
η − η0
t − t0
ηsp
=
=
+
ρ1
η0 ρ1
t0 ρ1
1
t
− ν1part
ρ0
t0
(2.28)
t0 is the flow time of the pure solvent while t is the flow time of a solution of
solute density ρ1 . ρ0 is the density of the pure solvent and ν1part is the partial specific
volume of the macromolecular solute. The term in 1/ρ0 - ν1part is a correction for
the effect of the solute on the density of the solution [116]. This can be assumed be
to be negligible for sufficiently dilute solutions.
2.4.4
Surface Viscosity
As mentioned above in section 2.3.5 there are two forms of viscosity that can be
derived from the complex elastic moduli, ηd , the compression viscosity, and ηs ,
the shear viscosity. The compression viscosity multiplied by oscillation frequency
(ω) is the imaginary part of the complex dilational modulus, which is also called
the compression modulus (see equation (2.16) above). Similarly the shear viscosity
multiplied by the frequency of oscillation is the imaginary part of the of the complex
shear modulus (see equation (2.17) above).
65
Oscillating Ring Viscometer
The initial tensiometer using a ring in the surface of a liquid to probe its surface
tension was made by du Noüy in 1925 [139]. This measurement method involved a
static ring and a simple measurement with applied masses to pull it out of contact
with the surface, and so was merely a formalisation of the measurement methodology
used earlier, for example in [103]. Further analysis of the du Noüy method was
carried out by Lukenheimer and Wantke [140] who concluded that it was a very
reliable method for surface tension measurements.
Surface monolayer viscosity can be measured by applying steady angular velocity
to a du Noüy ring. An analysis of the mathematics behind a ring with steady angular
velocity in knife edge contact with as semi-infinite plane liquid surface was carried
out by Goodrich et al. in 1971 [141]. Details of the ring rheometer used in this work
are given in section 4.3, a schematic of the rheometer is shown in figure 4.2 and a
schematic of the oscillating ring is shown in figure 4.3. Lucero Caro et al. have used
an oscillating ring rheometer to examine the properties of monolayers of DOPC
and DPPC at different pHs [114], and DPPC monolayers penetrated by β-casein
[142]. Spigone et al. examined the effects of chain length on the surface rheological
properties of poly(vinyl acetate) [124]. This methodology is seen to provide a good
way to probe the surface shear rheology of a monolayer.
Measuring Surface Dilational Viscosity
By inducing an oscillation in a surface the complex elastic moduli can be measured and thus the surface viscositic properties can be calculated.
Monroy et
al. [143] looked at the dilational viscoelasticity of three cationic surfactants of
differing chain lengths (dodecyl-trimethylammonium bromide (DTAB), tetradecyltrimethylammonium bromide (TTAB) and hexadecyl-trimethylammonium bromide
66
(CTAB)) at a range of surface concentrations. It was observed that the shortest
chain surfactant behaved insolubly at low concentrations, that is, the real part of
the dilational modulus increases with concentration and the imaginary part increases
because of the increased amount of surfactant needed to replenish a perturbed surfaces. At high concentrations diffusion occurs fast enough to remove concentration
gradients produced by the surface perturbation. This leads to a reduction in both
the real and imaginary parts of the dilational modulus. Thus, both go through a
maximum at an intermediate concentration. Although the real part of the dilational
modulus of the longer chain surfactants behaves similarly to that of the DTAB, the
imaginary components do not. TTAB has a dilational viscosity that decreases with
increasing concentration becoming negative and going through a minimum before
becoming positive again. CTAB’s imaginary component goes and stays negative.
The negativity of these components is a failure of the diffusion model but a better
model is not suggested. Monroy et al. also note that the dilational modulus of the
CTAB monolayer does not vary much over the frequency range 1 to 400 kHz.
Later work done by Monroy et al. [120] looked at the dilational viscosity of PVAc
over a ten decade frequency range up to 105 Hz. The viscosity was found to reduce by
seven orders of magnitude over this range. Aksenenko et al. [144] looked at mixed
adsorption layers made up of protein β-lactoglobulin and a non ionic surfactant
(alkyl dimethyl phosphine oxide) and obtained the frequency dependence of the
dilational modulus and the phase angle (the phase difference between stress and
strain). Miller et al. [135] observed the surface shear viscosity - surface pressure
relation for two lipids (DPPC and DMPE), observing that the viscosity was low
at low surface pressure, but above 20 mN/m the surface shear viscosity started
increasing. DPPC formed the more viscous monolayer.
67
2.5
2.5.1
Behaviour of Polymers in Solution
Introduction
A solution is usually defined as a homogeneous mixture, that is the mixing is on
the molecular level in which the solvent and solute are free [145]. In the case of
a polymer in solution the solute molecules are not as free as the rest of the chain
limits their possible conformations.
2.5.2
Thermodynamics of a Polymer Solution
Analysis of an ideal solution assumes that the molecules of solvent and solute are
the same size, when the solute is a polymer then this is not the case and the above
analysis can not be assumed to apply. Also, in the case of a chain molecule, the
shape it takes in solution is relevant to the properties of the solution.
Background to the Flory-Huggins Theory
In the first half of the twentieth century there was a surge of interest in the behaviour
of liquid polymers. A simple relationship between viscosity and molecular weight
was theorised by Dunstan [146], but experimental work by Albert a few decades
later [147] showed this relationship to be incorrect. Further experimental work by
Flory [148] built upon this earlier work and provided an exact relationship between
viscosity and chain length for a liquid polymer. The work of Flory [148] and Kauzmann and Eyring [149] showed that in the absence of a solvent polymers flow only
as a result of the random movement of segments, that is a collection of 20 to 30
atoms.
It wasn’t until the work of Powell et al. [150] that work was done on polymer
solutions. This paper looks at the behaviour of polymers when mixed with smaller
68
molecules and leads to analysis of the osmotic pressure which has a large increase
with concentration for long molecules, the vapour pressure above a concentrated
solution and the volume and entropy changes related to swelling of the polymer
molecules. It is shown probabilistically that in solution polymer particles do not
have a chance to move by segments as the surrounding space is occupied by the
solvent molecules.
Various experiments have been carried out investigating how the properties of a
polymer solution change with the length of the polymer [151]. This work by Gee
and Treloar uses a system of rubber (polymer) in benzene (solvent) and goes on
to measure and calculate various thermodynamic properties of the system. Further
work in this area lead to the Flory-Huggins theory, which is discussed in greater
detail below.
The Flory-Huggins Theory
The Flory-Huggins solution theory is based off work done by Flory in 1942 [152]
which builds of the works of Huggins in 1941 [153] and 1942 [154]. This theory
describes the properties of a solution of long chain molecules starting from the
entropy of mixing caused by adding the molecule to the solvent.
The model considers a hypothetical solution containing N2 spherical molecules
(the solvent) and N1 chain molecules (the solute) each of which consist of x submolecules each of which is equivalent in size and shape to a single molecule of type
2 [145, 154]. It is assumed that there is no volume change on mixing and the heat of
mixing is also zero. Statistically the solution is treated as if it were a solid solution
which has N2 + xN1 sites each other which can take either a solvent molecule or a
solute submolecule. The solute molecules are added (hypothetically) one at a time
then the solvent molecules are added, the different places a molecule or submolecule
can be added to are counted and thus the total number of configurations can be
69
found.
The first submolecule of the first solute molecule can be placed in any of the N2
+ xN1 sites, the second submolecules of the same solute molecule has z possible
locations, where z is the co ordination number of the molecule (number of nearest
neighbours). The third submolecule has y possible locations, where y = z - 1 if the
molecule is perfectly flexible but in real molecules y is less than this. All further
submolecules of the first solute molecule have the same number of alternatives, y.
The value of y can be adjusted to take account of ‘blocking’, that is the possibility
that one or more of the possible sites for a given submolecule may all ready be
occupied by other submolecules of the same polymer molecule.
The first submolecule of the second solute molecule can be placed in any of N2
+ xN1 - x sites. The second submolecule has z sites available to it, if it is adjacent
to no submolecules of polymer molecule one. It has z - 1 sites available if it is
adjacent to one all ready occupied site, z - 2 sites available if it is adjacent to two
occupied sites and so on. The average number of sites available for the second
submolecule is z(1 - f2 ), where f2 is the chance that a nearest-neighbour site is all
ready occupied. Further submolecules of solute molecule two each have an average
of y(1 - f2 ) alternate sites.
In general, for ith solute molecule the first submolecule has N2 + xN1 - (i - 1)x
alternative sites. The second submolecule has on average z(1 - fi ) alternatives and
each of the other x - 2 submolecules has y(1 - fi ) alternatives.
Individual number of possible placements are calculated for each molecule in the
chain. By Multiplying these numbers together, gives ν, the total number of possible
conformations of the polymer molecule in the lattice is obtained. This calculation
is carried out for each of the N1 polymer molecules. To calculate the number of
distinguishable configurations of solute molecules the total number of configurations
must be divided by N1 !, as exchanging one polymer molecule for another in exactly
70
the same position makes no difference to the configuration. After the addition of all
the solute molecules, the remaining sites are filled with the solvent molecules. As
all of these molecules are indistinguishable from each other there is no increase in
the number of different configurations at this point.
71
Chapter 3
General Methodology
3.1
Introduction
This chapter describes the methodologies which are common to all of the experiments described in Chapters 4, 5 and 6. Methodologies specific to each group
of experiments are detailed in the relevant chapters. First the procedure for the
extraction and purification of α-elastin will be detailed in section 3.2. The prior
characterisation of α-elastin is detailed in section 3.2.1. The preparation of other
materials used is described in section 3.3; this section also details equipment common to all the experiements. The Langmuir Troughs used in the experiments are
briefly described in section 3.4, more details on each trough are included in the detailed methodologies given later in this thesis. This section (3.4) also includes the
preparation carried out on the trough and the subphase surface. The cleaning of
the surface and the general procedure for cleaning the trough is described in section
3.4.1. The procedure for spreading a monolayer on the, now clean, surface is laid
out in section 3.4.2.
72
3.2
Elastin Preparation
α-elastin was prepared by acid hydrolysis of porcine (Sus domestica - domestic pig)
aorta following the procedure of Partridge et al. [16], detailed further below. It was
shown to have similar coacervation characteristics to that of Partridge et al.. A 0.05
mg/ml solution of α elastin in water was made using nanopure water produced by
a Millipore Direct Q UV-3 system. The water had a resistivity of 18.2 MΩ/cm.
Fresh aortas were obtained from the local abattoir. They were separated from the
surrounding tissue and then finely sliced. The pieces were boiled in 0.1 M NaOH for
45 minutes. After this digestion process the aorta was chilled to 4◦ C and washed in
repeated changes of water until it reached pH 7. It was then oven-dried and finely
ground. The sodium hydroxide digestion removes most other components from the
aorta leaving just the elastin. The ground aorta is then boiled in 0.25 M oxalic
(ethandioic) acid for an hour before centrifuging at 3000 rpm for 10 minutes. The
supernate is then carefully poured off and preserved. Further oxalic acid is added
to the centrifuge residue to wash it and the supernate is then added to that already
obtained. The boiling - centrifuging - washing cycle is then repeated a further five
times. The result of this process is a transparent yellow solution, which is then
dialysed against distilled water until it it is free of remaining oxalate. At this point
it will reach pH 7. The elastin is then freeze dried for storage.
The α-elastin used in this work was extracted and purified for me using the above
methodology by Ellen Green and Dick Ellis.
Solutions of elastin were prepared with water and stored in glass containers at
5◦ C.
73
3.2.1
α-Elastin Characterisation
Coacervation was observed under a light microscope in a glass cell consisting of a
microscope slide and a glass cover slip separated by a parafilm gasket. The temperature of coacervation was seen to be consistent with literature.
Titration Experiments
The charge characterisation of elastin was performed using rings of porcine aorta
digested for 72 hours in 88% formic (methanoic) acid at 45◦ C. Two titrations were
carried out, one in which the pH was lowered with 500 µL - 8 mL additions of 0.01
M HCl, and one in which the pH was raised with 50 µL - 1 mL additions of 0.01
M NaOH. Each titration began with a single elastin ring in 40 mL deionised water.
The pH of the solution surrounding the elastin was monitored continuously. After
the titration the elastin ring was dried in an oven so the dry weight of the elastin
could be recorded. This experiment was repeated in 0.5 M NaCl and 0.5 M CaCl2
solutions.
In order to calculate the number of moles of hydrogen bound per unit dry weight
of elastin (Hb ), the following equation is used;
Hb =
V ([acid] − [base] + [OH − ] − [H + ])
W (1 + [OH − ])
(3.1)
where V is the total volume of solution present around the elastin ring, [acid]
and [base] are the molar concentrationsof acid or base. [OH − ] is the concentration
of OH− ions in the solution; similarly [H + ] is the concentration of H+ ions in the
solution . W is the dry weight of the elastin ring.
The titration curves for elastin in water and in the salt solutions, which are shown
in figure 3.1, show a similar trend in their shape. Each has a steep downward sloping
section between pH 2 and 3; they then level off into a horizontal section between
74
pH 4 and 8.5; and then there is another steep downward section from pH 9 to 11.
Bendall [66] presented titration data
from solutions of α and β elastin. The
range of concentrations used was 8 to 20
mg/ml, which is comparable to the solutions used to examine the bulk viscosity in
chapter 4. Data from the solutions is consistent with the data presented in figure 3.1
in terms of the maximum and minimum Figure 3.1:
amounts of bound H+ . The solution of αelastin had a defined point of zero bound
Amount of bound hydrogen
against pH from titrations of elastin rings in
water, 0.05 M NaCl solution and 0.05 M CaCl2
solution.
H+ , and in both cases this was at pH 5.
This is a contrast to the flatter dependence presented in figure 3.1 which indicates
that the amount of hydrogen bound is very low between pH 4 and pH 8.4. The
elastin rings used in the data presented above were not completely digested; so, it
is indicated that the bindings which occur at pHs 4-5 and 5-8.4 are the result of
groups produced by the final stages of digestion from elastin to α-elastin.
The curve from elastin in 0.05 M NaCl is offset to higher pHs from the curves
from 0.05 M CaCl2 and water. This indicates that the Na+ ion affects the pH at
which hydrogen dissociation occurs. It is possible that this difference is related to
the difference in the binding between Na+ and Ca2+ and elastin [27]. In the pH
range 2.00 to 6.00 the disassociation of hydrogen has been attributed to the loss of a
hydrogen ion from a side chain carboxyl group [137, 155]. Thus the number of moles
of carboxyl groups unmasked by the titration was determined by the change in Hb
when the pH was increased in this range. Similarly, the number of moles of imidazole
groups can be determined from the change in Hb when the pH was increased from
6.00 to 8.25; and the number of moles of basic groups can be determined from the
75
Table 3.1: Results of calculations on the charged groups present in elastin in various solvents.
Calculations were made on titration data, H+ dissociates from carboxyl groups between pH 2.5
and 6; from imidazole groups between pH 6 and 8.25 and from basic groups between pH 8.25 and
10.5.
Carboxyl
Imidazole
Basic
Groups
Groups
Groups
Solution
µmoles/g
µmoles/g
µmoles/g
Content
dry weight
dry weight
dry weight
Water
691
16
35
0.05 M NaCl
588
18
180
0.05 M CaCl2
254
12
450
change in Hb when the pH was increased from 8.25 to 11.50. The titration data
does not cover this complete pH range thus carboxyl groups are determined from
the pH range 2.50 to 6.00 and basic groups are determined from the pH range 8.25
to 10.50.
The data in table 3.1 suggests that increasing the amount of Cl− present in the
solution decreases the amount of hydrogen dissociating from the carboxyl groups
and increases the dissociation from basic groups. The amount of cationic charge
present in the solution might also affect the unmasking of these groups since the
decrease in the unmasking of carboxyl groups and the increase in the unmasking of
the basic groups follows the increase in cationic charge in the solution.
3.3
Other Materials and Equipment
The subphase for most of the measurements was also nanopure water. Where ionic
subphases were used, a 0.1 M solution of salt (either NaCl or CaCl2 ) was prepared
76
with nanopure water. This concentration was chosen to mimic the physiological
concentration of Na+ [86]. The physiological concentration of Ca2+ is much lower but
as detailed in section 2.2.4 it is believed that calcium ions interact with elastin. We
chose the same concentration in order to study these interactions. pH was adjusted
with small additions of concentrated sodium hydroxide or hydrochloric acid. The
pH of the trough was measured with a Hanna Instruments pH210 Microprocessor
pH meter.
3.4
Langmuir Troughs
In this work three different Langmuir Troughs were used to suit the different requirements for each particular type of experiment. The first trough was made of PTFE
and was used in all of the quasi-static experiments that did not involve the use of a
microscope. This trough is illustrated in figure 5.1 in section 5.2. The second trough,
also PTFE, was used for the dynamic measurements and is illustrated in figure 5.4
in section 5.3. The third, shallow trough was used for the fluorescence microscopy
experiments which had an intergral window and PTFE rim. This trough is shown in
figure 6.1 in section 6.2.2. The quasi-static trough and the microscope trough were
used with a Kibron MicroTrough system (Kibron Inc, Helsinki, Finland), using a
metal alloy (resistant to all acids and bases) wire probe. The dynamic measurements
trough was used with a PS4 Nima Technology (UK) microbalance sensor, which uses
filter paper pressure probes.
When the troughs were filled with the subphase, the surface of the subphase rose
above the level of the top of the trough. This ensured that sweeping the barriers
together moved impurities on the surface together as well so they could be removed.
In general, surface tension allowed the subphase to stand 5 mm or so proud of the
trough lip. More detailed information on the methodologies used will be discussed
77
in sections 5.2 (Quasi-static methodology), 5.3 (Dynamic methodology) and 6.2
(Microscopy methodology).
3.4.1
Cleaning of the Trough
The normal cleaning procedure between experiments for the troughs and barriers
was thorough rising with water, rinsing with ethanol, and then rinsing with more
water. Between experiments the quasi-static trough was stored under ethanol to
prevent contamination. The other troughs were stored under cover. Troughs were
thoroughly rinsed with water prior to filling with subphase.
In preparation for an experiment the subphase surface was cleaned by repeated
compression and aspiration until it produced a negligible (< 1 mN/m) pressure
rise over a full compression. To ensure that the suction tip was not a source of
contamination for the surface, the metal suction tip was flamed with a gas torch
to red hot to clean it and the glass suction tip was replaced after each experiment.
Where metal alloy wire probes were used the probe was also regularly flamed to
remove contaminants, they were stored under ethanol. The filter paper probes were
replaced after each experiment.
When α-elastin had been added to the subphase the trough was cleaned with
hydrochloric acid before rinsing with water, ethanol, more water and storing.
All experiments were carried out with the trough inside a perspex enclosure to
protect the monolayer from dust and drafts. The perspex cover used for the fluorescence microscopy was tinted to cut down on background light. Over longer
experiments it was found that evaporation of the subphase could cause a problem
by lowering the level of the subphase. To counteract this paper towel soaked in
water was placed inside the enclosure to keep the level of moisture in the air inside
high and so prevent evaporation of the subphase.
78
3.4.2
Spreading a Monolayer
The solution for the monolayer was added to the subphase surface dropwise from
a microlitre syringe. With each drop separately touched to the surface, the drops
were spread over as wide an area as possible. However, the drops were not placed
too close to the barriers, trough edges, or the probe, as in these regions the water
surface is not perfectly flat due to the meniscus at the contact points, and thus the
applied α-elastin would not spread evenly. Once the monolayer had been applied it
was allowed to equilibrate for a few minutes before the start of the first compression.
Spreading of monolayers was done at maximum surface area.
79
Chapter 4
Viscometry and Surface
Rheometry
4.1
Introduction
This chapter details the experiments used to examine the bulk and surface viscosity
of α-elastin. Bulk viscosity was detailed in section 2.4.1 with measurement methods
discussed in section 2.4.3. Examining the the surface viscosity, which is described in
section 2.4.4, reduces the complexity of interactions by confining the molecules to
two dimensions; this allows further parameters to be measured, such as the radius
of gyration. The measurement methodology used to examine the surface viscosity
was also introduced in section 2.4.4.
In this chapter the methods used to measure the viscosity and elastic moduli
of α-elastin solutions under a variety of conditions are described and the results
of these experiments are presented and discussed. In section 4.2 the methodology
used to measure the bulk viscosity of the solution is detailed. The methodology
used to examine the surface viscosity is described in section 4.3. The results of the
bulk viscometry experiments are presented in section 4.4. Analysis of the intrinsic
viscosity and calculations of the radius of gyration are given in section 4.4.1.
80
This is followed by the results of the surface viscometry experiments, which are
presented in section 4.5. Firstly, the frequency and surface pressure dependences
of the shear modulus and its dissipative and storage components are presented in
section 4.5.1. Then analysis of the phase lag between the components of the shear
modulus in given in section 4.5.2. This is followed by data on the shear viscosity in
section 4.5.3. Finally, there is discussion of an observed phase transition in section
4.5.4, and fitting of the data to Eyring’s model in section 4.5.5.
4.2
Bulk Methodology
As mentioned in section 2.4.3, the bulk viscometry measurements were performed
in a glass capillary viscometer, a schematic of which is shown in figure 4.1. The viscometer was placed in a water bath (not shown) to reduce variations in temperature
during the experiment.
The viscometer was filled so the meniscus of the
solution whose viscosity was to be measured reached
the red line in arm A. A syringe was used to aspirate
liquid into arm B, the meniscus had to be above the
higher of the two blue lines indicated. The flow of the
solution back to its equilibrium position is restricted
by the capillary in arm B. The time taken for the
meniscus to pass from the higher blue line to the lower
blue line is measured.
Figure 4.1: A schematic of a
capillary viscometer
As detailed in section 2.4.3, the viscosity of the solution can then be calculated from the ratio of the time
taken for the solution to flow through to the time taken for water to flow through.
Measurements were made at room temperature and at 5◦ C at a variety of α-elastin
81
solution concentrations using three different solvents (water, 0.1 M NaCl and 0.1 M
CaCl2 ). In each case the solution was placed in a freshly cleaned viscometer and
allowed to equilibrate at 5 ◦ C. During the equilibration the open ends of the viscometer were covered to prevent contamination. Once the solution had equilibrated two
sets of ten data points were recorded. The viscometer was then transferred to the
room temperature water bath and allowed to come to a new equilibrium, again with
the open ends covered. The amount of solution was adjusted once it had warmed.
The room temperature data was collected, again two sets each of ten measurements
were taken, then the elastin solution was removed from the viscometer and it was
then thoroughly cleaned before the next concentration of elastin solution was added.
The cleaning procedure for the viscometer involved filling it with 1 M Nitric Acid
and allowing it to soak for at least an hour. Then rinsing the viscometer through
with water, then ethanol and then more water. The viscometer was then allowed to
dry before being used again.
As noted in section 2.4.2, by making a series of measurements using a particular
solvent and solute at different solution concentrations the intrinsic viscosity can be
calculated. This will be discussed in more detail in section 4.4.1 below.
4.3
Surface Methodology
The surface measurements were made with a slightly modified Camtel CIR-100
rheometer and a du Noüy ring which oscillates in the interfacial plane. The rheometer used is based on the set-up developed by Sherriff and Warburton [156]. The
schematic diagram shown in figure 4.2 shows how the rheometer is set up; image
taken from [157]. This is a stress-rheometer it maintains a set amplitude and frequency through feedback. The Du Noüy ring was made of a platinum/iridium alloy
and was hung from a virtually frictionless suspension wire. The oscillations of the
82
ring are instigated by the drive unit coil, this operates similarly to a taut band
galvanometer. The amplitude of the motion of the ring is then detected by a sensor
that reflects light off the ring. When it is working in normalized resonance mode
(>2 Hz), the feedback control system forces the system into phase resonance. Under
these conditions the input stress leads the output strain by 90 degrees. The current
required to force the phase angle to 90◦ is proportional to the shear storage modulus,
G0 , and the current which produces the required amplitude of strain is proportional
00
0
00
to the shear dissipative modulus, G . Thus G and G can be calculated.
The modifications made to the rheometer
mean that the measurement head is attached to
an optical rail and mounted so it can be positioned freely over the trough surface and lowered into the surface. An aluminium ring with
two openings for the layer to flow through (as
shown in figure 4.3) defines the shear deformation geometry, ensuring that an equal area of
Figure 4.2: A schematic of a Cam-
monolayer is probed in all experiments. The ex- tel CIR-100 Rheometer which shows the
ternal ring has a diameter of 35 mm and the Du oscillating Du Noüy ring and the mechNoüy ring has a diameter of 12 mm. The in- anism by which the amplitude oscillaner ring oscillates in the interfacial plane, in this
work the angular amplitude was set to 5 µrad
tion is monitored. Image taken from
[157].
and the frequency of the oscillation varied between 3 and 20 Hz. A mechanical
transducer measures the torque applied to the ring by the monolayer. The small
angular amplitude of the oscillation means that the strain applied to the monolayer
00
is small, a constant applied strain of 0.00697 was used, so G can be measured down
to µN/m.
83
The temperature of the monolayer was controlled
by a water bath integral to the Langmuir Trough, the
experiments were carried out at 23 ◦ C . This rheometer set-up was applied to the dynamic methodology
trough which is described in detail in section 5.3.
Overhead view
The monolayer was spread on the surface of the
of the set up used for sur-
subphase, water and 0.1 M CaCl2 were used as sub-
Figure 4.3:
face rheometry measurement. It
phases and then the pressure of the monolayer was
shows the smaller Pt/Ir Du Noüy
ring which oscillates inside the
set by adjusting the barriers and then allowing the
larger aluminium ring which sits
monolayer to equilibrate for a few minutes. A sweep
at the bottom of the trough and
of frequencies was carried out at each applied surface
ensures the same area of mono-
pressure.
layer is probed in each experiment.
The inner ring was flamed to ensure cleanliness and
allowed to cool before being hung on the transducer.
A calibration run, consisting of a full frequency sweep, was carried out on the bare
subphase before the monolayer was applied.
4.4
Bulk Viscometry Results
The graphs below, figures 4.4 and 4.5, show concentration against the flow time for
the low and room temperature data respectively taken with the α-elastin solution
made up with water only. Each point is the average from twenty measurements.
The error bars represent the range of the whole set of data.
The raw data, the flow times, can be converted to viscosity values by the use of
equation (2.27) which implies that:
ηwater
ηsample
=
ρwater twater
ρsample tsample
84
(4.1)
Figure 4.4: Concentration against flow time data taken for an α-elastin solution in water at a
variety of concentrations at a temperature between 4 and 5◦ C.
Figure 4.5: Concentration against flow time data for an α-elastin solution in water at a variety
of concentrations at room temperature (20 - 23◦ C).
In both figures 4.4 and 4.5 each data point is the mean of twenty measurements and the error bars
represent the total range of those twenty measurements.
85
Equation (4.1) can then be rearranged to give a value for the kinematic viscosity
(v) of the sample.
ηsample
ηwater tsample
=
= vsample
ρwater twater
ρsample
(4.2)
As the density and viscosity of water are constant for a given temperature, dividing the flow time of the α-elastin solution by the flow time of water at the same
temperature gives a value which is proportional to the kinematic viscosity. As the
kinematic viscosity is proportional to the dynamic viscosity, the relative flow time
must also be proportional to the dynamic viscosity.
The graphical data presented in this section is in the form of relative viscosity
as calculated from flow time. However, using values for the viscosity of water from
[158] and values for the density of water from [159], the values of the kinematic
and dynamic viscosity shown in table 4.1 were calculated for the different solutions.
When calculating the dynamic viscosity of the α-elastin solution it was assumed that
the density of the solution was equal to that of the solvent at the same temperature.
Obviously, this assumption becomes less valid the more concentrated the solution
is; however, the most concentrated solution reported here only had a concentration
of 50 mg α-elastin per mL water. The calculation of the kinematic viscosity does
not require a numeric value of the solution’s density. The density data for the
salt solutions was taken from [160] and it was assumed that the salt solutions had
viscosity equal to that of water at the same temperature and the flow time data
recorded in figures 4.6 and 4.7 supports this as a reasonable assumption.
Figures 4.6 and 4.7 show the relative dynamic viscosity of α-elastin solutions in
different solvents at 5◦ C and room temperature (23◦ C), respectively, and compares
the solutions with water at the same temperature. It can be seen that the bulk
viscosity of an α-elastin solution decreases with increasing temperature. The flow
times shown in figure 4.4 are much greater than those shown in figure 4.5. At lower
concentrations there is a factor 1.7 increase in in the flow time with the reduction
86
Table 4.1: Kinematic and dynamic viscosity for the most concentrated α-elastin solutions used as
well as values for water for comparison. The dynamic viscosity data for water was taken from [158]
and the kinematic values were calculated using water density values from [159]. It was assumed
that the α-elastin solutions had the same density as the solvent at the same temperature. Density
data for the salt solutions was taken from [160]. It was assumed that viscosity of the salt solution
was equal to that of water at the same temperature.
Temp
◦
Elastin
Salt
Dynamic
Kinematic
Conc
Present
Viscosity
Viscosity
mPa s
m2 /s x10−6
C
mg / mL
5
0
None
1.519
1.519
5
50
None
3.073
3.073
5
50
NaCl
2.977
2.987
5
50
CaCl2
3.270
3.297
20
0
None
1.002
0.935
20
50
None
1.671
1.674
23
50
NaCl
1.678
1.676
23
50
CaCl2
1.709
1.719
87
Figure 4.6: Relative viscosity against α-elastin concentration for three different solvents at 5◦ C.
Data is scaled relative to the viscosity of water at 5◦ C.
Figure 4.7: Relative viscosity against α-elastin concentration for three different solvents at room
temperature (varying between 20 and 23 ◦ C). Data is scaled relative to the viscosity of water at
20◦ C.
88
in temperature. This difference increased with elastin concentration. At 50 mg/mL
the increase was a factor 2. Figures 4.6 and 4.7 clearly show that at 5◦ C the viscosity
of an α-elastin solution is higher than at room temperature.
At low concentrations of elastin the solvent does not make much difference to
the viscosity of the resulting solution but at 50 mg/mL at both temperatures the
calcium chloride solvent made the most viscous solution. At low temperature the
calcium chloride based solution is more viscous that the other solutions beyond the
bound of experimental error. It is therefore indicated that Ca2+ ions in bulk affect
α-elastin differently to Na+ ions. The binding of Ca2+ to the elastin makes the
protein chains more rigid. This agrees with observations discussed previously in
section 2.2.4.
The sodium chloride solution is also more viscous than the water solvent; this
difference is more pronounced in more concentrated solutions at room temperature.
It is therefore suggested that Na+ also binds to α-elastin; this agrees with the
observations of Winlove et al. [27]. They suggest that while Ca2+ binds permanently
to elastin, Na+ does not; this is likely to be why the difference between NaCl and
CaCl2 are more pronounced at lower temperatures.
As table 4.1 shows, the difference in viscosity between the elastin in sodium chloride and elastin in water was much smaller than the difference in viscosity between
elastin in calcium chloride and elastin in water.
4.4.1
Intrinsic Viscosity Calculations
As described in section 2.4.2 and equation (2.22) the intrinsic viscosity can be calculated from the gradient of a reduced specific viscosity against mass concentration
plot. Figure 4.8 gives a value of [η] of 0.0081 mL/mg for α-elastin in water at 5◦ C.
The Huggins constant calculated is k = 3.05. A similar analysis of data taken at
89
Figure 4.8: The dependence of the reduced specific viscosity on the concentration of in solution.
Data taken at 5◦ C using solutions in water.
room temperature gave [η] = 0.0073 mL/mg and k = 1.88. Thus it is concluded
that with a water solvent the intrinsic viscosity of α-elastin decreases with increasing temperature, which suggests that the conformation of the elastin in solution also
changes with temperature, with the higher temperature conformation being smaller.
Monkos [138] collated data from five proteins which have similar intrinsic viscosities to α-elastin: ovine serum albumin at 5◦ C [η] = 0.00407 mL/mg; human serum
albumin at 5◦ C [η] = 0.00490 mL/mg; bovine serum albumin at 5◦ C [η] = 0.00649
mL/mg; human immunoglobin at 5◦ C [η] = 0.00959 mL/mg; and lysozyme at 5◦ C
[η] = 0.00305 mL/mg. These compare with 0.0081 mL/mg for α-elastin at the same
temperature. Ghaouar et al [161] examined cellulase at 25 ◦ C and found that [η] =
0.00325 mL/mg and increased with decreasing temperature, this is consistent with
what is seen for α-elastin. Intrinsic viscosity can take a very wide range of values depending on substance; Harding [162] produced a review article that collates
the intrinsic viscosities and molecular masses for many different proteins and and
polypeptides. Based on molecular masses the most similar to α-elastin (molecu90
lar mass 67 kDa) are Conalbumin (75.5 kDa) at 25 ◦ C [η] = 0.0035 mL/mg and
Haemoglobin (68 kDa) [η] = 0.0056 mL/mg. Thus the value measured for the intrinsic viscosity of α-elastin seems reasonable when compared with previous data.
However, similar molecular weight does not mean these molecules are comparable to
α-elastin in solution, that depends on the conformation the molecule assumes when
in solution.
Similar analyses carried out with sodium chloride and calcium chloride solvents
indicate that the intrinsic viscosity of α-elastin in these solvents also decreases
when temperature is increased. Fox and Flory [163] examined the behaviour of
polystyrenes of different molecular weights in a variety of solvents at a variety of
temperatures. It was seen that the dependence of [η] on temperature is dependant
on the solvent-polymer combination.
Radius of Gyration
Calculation of the intrinsic viscosity of a solute allows analysis of the radius of
gyration of that solute. This is shown using the Flory methodology [137, 164]. The
intrinsic viscosity can be expressed as
[η] = νk
NA vh
M
(4.3)
where νk is a constant for a given polymer and determined by the shape of the
solute molecules, NA is the Avogadro constant, M is the molecular weight and vh
is the hydrodynamic volume occupied by a molecule of solute. A flexible polymer
in solution behaves as sphere of radius R = ξRg , where Rg is the radius of gyration
and ξ is a constant for a given shape. Thus, vh = 43 πξ 3 Rg3 . In the case of a sphere,
the assumed shape of a molecule of α-elastin in solution, νk = 2.5. Substituting
these equations for νk and vh in to equation (4.3) gives,
91
[η] =
10πNA 3 3
ξ Rg
3M
(4.4)
Thus, from equation (4.4) [η] is proportional to the cube of the radius of gyration.
This analysis was initially carried out by Flory and Fox [165]. In water, the radius of
gyration of α-elastin increases by 2.2% when the temperature is decreased from 22◦ C
to 5◦ C. A similar percentage increase is observed in NaCl for the same temperature
change. The increase observed in CaCl2 is approximately 5.8%. This indicates that
α-elastin in the presence of Ca2+ ions expands more in response to a decrease in
temperature than in the presence Na+ .
At low temperatures the radius of gyration of α-elastin is seen to be smaller in
water than in either salt solution. This indicates that the presence of ions in the
solution causes the α-elastin to expand. It is known that both Ca2+ and Na+ bind to
α-elastin in similar amounts [27] and this viscometry data indicates that the binding
causes the α-elastin to increase in volume. It is presumed the binding to the elastin
is occurring at neutral and charged sites which would agree with work discussed in
section 2.2.4.
4.5
4.5.1
Surface Viscometry Results
Surface Shear Modulus
0
00
Measurements of the storage (G ) and dissipative (G ) components of the shear elastic modulus were carried out using the methodology described above in section 4.3.
Measurements were made at 2 mN/m intervals at 18 different oscillation frequencies
between 3 and 20 Hz.
Figures 4.9 and 4.10 show the storage component of the shear modulus on water
and 0.1 M calcium chloride respectively. Figures 4.11 and 4.12 show the dissipative
92
component of the shear elastic modulus on water and 0.1 M calcium chloride respectively. Finally, figures 4.13 and 4.14 show the complex shear modulus on water and
0.1 M calcium chloride respectively. These experiments were carried out at 23◦ C.
The monolayer on the water subphase was made up of 200 µL of 1 mg/mL α-elastin,
while the monolayer on the CaCl2 was made up of 50 µL of 1 mg/mL α-elastin.
0
Figure 4.10 shows a trend of increase in G with oscillation frequency; the trend
0
is not so clear on water (shown in figure 4.9). A rise in G is observed between 3
and 20 Hz for each surface pressure observed. It is unclear what the cause of the
0
peaks observed in G is. At about 18 Hz the monolayer on calcium chloride shows
0
a region with very low G , shown in figure 4.10, but its cause is unclear. Generally
speaking the monolayer behaves in a more solid-like manner at higher pressures;
this is expected due to increased intermolecular interactions due to smaller distances
between α-elastin molecules.
It is also noted from figures 4.9 and 4.10 that at surface pressures above 10 mN/m
0
the surface layer of α-elastin does not display liquid-like (G = 0, as defined in section
2.3.5) behaviour at any frequency.
Examining the low frequency part of the data (where the monolayer is closest
to equilibrium) shows that the storage component of the modulus is approximately
four times greater on the water subphase compared to the calcium chloride subphase
at the same pressure. It is believed that this may be caused by the difference in
area per molecule on the two subphases. At 10 mN/m there is 2970 Å2 /molecule
on the water and 7260 Å2 /molecule on the calcium chloride. Similarly at 18 mN/m
there is 710 Å2 /molecule, while there is 2990 Å2 /molecule on the calcium chloride.
At a given surface pressure, a molecule of α-elastin on CaCl2 has at least 2.4 times
as much area as a molecule on the water subphase; the increase in distance between
molecules leads to a reduction in the intermolecular forces between those molecules
and thus a decrease in the surface shear modulus is expected.
93
0
Figure 4.9: G against oscillation frequency at a variety of surface pressures on water. The
monolayer consisted of 200 µL of 1 mg/mL α-elastin.
0
Figure 4.10: G against oscillation frequency at a variety of surface pressures on 0.1 M CaCl2 .
The monolayer consisted of 50 µL of 1 mg/mL α-elastin.
The experiments whose results are shown in figures 4.9 and 4.10 were carried out at 23 ◦ C.
94
In terms of surface area per molecule 10 mN/m on water and 18 mN/m on calcium
0
chloride are comparable and in this case the CaCl2 subphase has a higher value of G
at all frequencies. This indicates that the presence of ions in the subphase increases
0
G at a given surface area per molecule, this is most likely related to the rise in
surface pressure observed when α-elastin is applied to an ionic subphase, this will
be discussed in more detail in section 5.4.2.
00
Figures 4.11 and 4.12 show a clear rise in G , indicating that this trend is not
reliant on the subphase. On the calcium chloride subphase (shown in figure 4.12), it
00
seems that G may be dropping at frequencies above 18 Hz; however, current data
does not allow it to be ascertained whether this is an actual peak or merely another
local maximum.
As shown by figures 4.11 and 4.12 the dissipative component of the shear modulus
of α-elastin is greater than the storage component particularly on the water subphase
and particularly at high frequencies. On the water subphase at 18 mN/m and
00
0
frequency of 19 Hz, G is nearly twice the value of G .
Spigone et al. [124] examined the shear elastic modulus of poly(vinyl acetate)
with respect to surface pressure, molecular weight, surface concentration, oscillation
0
frequency and compression speed. It was observed that G was low for all conditions
examined, showing a non-negligible value only at the lowest rate of compression;
even then values are approximately a tenth of what is here observed for α-elastin
on water even when the molecular weight was comparable (45 kDa vs 67 kDa for
the elastin). Finite solid-like behaviour is only observable when the monolayer is in
near equilibrium and able to maintain solid-like structures, it appears that α-elastin
has these conditions in the range of oscillation frequencies and surface pressures.
Comparing figures 4.9 and 4.10 with data in [124] indicates that α-elastin monolayers
0
are less solid-like (have lower G ) on calcium chloride than on water at the same
surface pressure although both are much greater than that observed for PVA (at 20
95
00
Figure 4.11: G against oscillation frequency at a variety of surface pressures on water. The
00
Figure 4.12: G against oscillation frequency at a variety of surface pressures on 0.1 M CaCl2 .
96
00
mN/m applied surface pressure and oscillation frequency 3 Hz G = 20 µN/m).
0
00
Spigone et al. [124] also observed an increase in both G and G with oscillation
frequency at ω less than 11 Hz; this fits with what is observed on α-elastin shown
in figure 4.11. Similar behaviour was seen in an α-elastin on calcium chloride.
Figures 4.13 and 4.14 show that the complex shear modulus increases with frequency on both water and calcium chloride. It is clear that these graphs show less
0
00
local fluctuations than either the G or G indicating that they cancel out when the
complex modulus is calculated. It is interesting that there is a pronounced drop
in shear modulus between 6 and 8 Hz on both subphases. It seems that as the
monolayer is particularly easy to deform at this frequency but it is unclear why.
00
In the case of G and G∗ it is clear that at a given surface pressure and frequency
the α-elastin monolayer on water has a higher resistance to shear than a comparable
0
monolayer on 0.1 M CaCl2 . This may also be the case for G but due to the variations
shown in figure 4.9 this is not completely clear. It is suggested that the electrostatic
interactions between the subphase and the α-elastin lead to a less ordered spreading
of the elastin and are responsible for the more liquid-like behaviour on the ionic
subphase.
00
As mentioned in sections 2.3.2 and 2.3.3 the surface pressure dependency of G
can be used to examine phase transitions in the monolayer. Thus, all data so far
presented is also presented as shear modulus against surface pressure at a variety
oscillation frequencies.
Cicuta et al. [110] examined the behaviour of monolayers of β-lactoglobulin at
various surface concentrations. At a surface pressure of 20 mN/m and an oscillation
0
frequency of 1 Hz it was observed that G was approximately 2200 µN/m, which is
00
comparable to data shown in figures 4.9 and 4.15. Under the same conditions G
was observed to be approximately 1100 µN/m which also agrees with data shown in
figures 4.11 and 4.17. Thus, it appears that α-elastin monolayers have similar shear
97
Figure 4.13: G∗ against oscillation frequency at a variety of surface pressures on water. The
Figure 4.14: G∗ against oscillation frequency at a variety of surface pressures on 0.1 M CaCl2 .
98
0
Figure 4.15: G against surface pressure at a variety of oscillation frequencies on water. The
0
Figure 4.16: G against surface pressure at a variety of oscillation frequencies on 0.1 M CaCl2 .
99
00
Figure 4.17: G against surface pressure at a variety of oscillation frequencies on water. The
00
Figure 4.18: G against surface pressure at a variety of oscillation frequencies on 0.1 M CaCl2 .
100
Figure 4.19: G∗ against surface pressure at a variety of oscillation frequencies on water. The
Figure 4.20: G∗ against surface pressure at a variety of oscillation frequencies on 0.1 M CaCl2 .
101
elastic behaviours to β-lactoglobulin monolayers.
0
Figure 4.15 indicates that G increases with applied surface pressure. This data
indicates that there is a clear difference in the behaviour of the α-elastin monolayer
at different frequencies. In particular the behaviour at 7.5 Hz and 18.2 Hz is very
0
different from the monolayer’s behaviour at other frequencies. The overall value of G
is low at these frequencies and the trends show a maximum (14 mN/m) and minimum
(16 mN/m). These variations occur at the same pressures at both frequencies but
it is unclear why the monolayer behaves differently at these frequencies. Figure
4.16 shows similar data for α-elastin on calcium chloride. The highest frequency
presented (19.1 Hz) has a similar shape to that on water. It is interesting to note that
at low frequencies and below the transition to the concentrated regime (discussed in
more detail below in section 4.5.4) the elastic modulus of the monolayer is also very
low; this indicates that due to the relatively large distance between the molecules
the interactions between them are weak and thus no elastic/storage component of
the shear modulus. The highest two frequencies do show an elastic component, even
at low surface pressures.
00
Figure 4.17 shows a trend for G to increase with applied surface pressure at all
frequencies although there is some variation in behaviour at intermediate frequencies. It is seen that the 5 highest frequencies behave similarly as do the 3 lowest
frequencies; this further illustrates the frequency dependence of the monolayer shown
in figure 4.11. It seems likely that this difference is due to the oscillation being fast
enough that the monolayer ends up in a different state. This phenomena may well
00
be localised around the probe. Figure 4.18 clearly shows that G tends to increase
with applied surface pressure at all frequencies. The 19.1 Hz characteristic seems to
00
have a lower G than might be expected. It is suggested that this behaviour maybe
be the result of relaxations within the monolayer not being able to keep up with
the oscillation frequency, resulting in an apparent liquid-like state. As this is not
102
observed on the water subphase it is suggested that the relaxation in the monolayer
on calcium chloride must be slower than those on water. This is a clear difference
in behaviour between the monolayers on water and on calcium chloride.
Figure 4.19 shows the complex modulus of an α-elastin monolayer increases with
0
00
applied surface pressure at all oscillation frequencies. The addition of G and G
0
has evened out a lot of the variations seen in G but there is still some variation
in behaviour at the intermediate frequencies. Figure 4.20 shows similar data for a
monolayer on a calcium chloride subphase. Again, the general trend of increasing
G∗ with surface pressure is clear. In this case the variant behaviour is shown by
the two highest frequency measurements which have higher values for the complex
shear modulus than expected from the behaviour at other frequencies.
4.5.2
Phase Lag Calculations
Using equation (2.20) in section 2.3.5 it is possible to calculate the phase lag (δ)
0
00
between the storage (G ) and dissipative (G ) components of the shear modulus.
This shows whether the monolayer is viscous (δ greater than 75◦ ), elastic (δ less
than 10◦ ), or viscoelastic (δ between 10 and 75◦ ). Figure 4.21 below shows how
the phase lag between the dissipative and storage components of the shear modulus
varies with the applied oscillation frequency at several surface pressures measured
in an α-elastin monolayer on water. Similarly figure 4.22 shows how the phase lag
between the dissipative and storage components of the shear elastic modulus varies
with the applied oscillation frequency at several surface pressures measured in an
α-elastin monolayer on a 0.1 M calcium chloride solution.
Figure 4.21 shows that over the pressure range 10 to 20 mN/m and frequency
range 2 to 20 Hz an α-elastin film on water is in the viscoelastic range. At frequencies
above 8 Hz, the phase angle is ∼ 70◦ which indicates a significant viscous component.
103
Figure 4.21: δ against frequency at constant applied surface pressure on water. The monolayer
consisted of 200 µL of 1 mg/mL α-elastin.
Figure 4.22: δ against frequency at constant applied surface pressure on 0.1 M CaCl2 . The
monolayer consisted of 50 µL of 1 mg/mL α-elastin applied to 0.1 M CaCl2
104
It is only at low frequencies that there is a significant elastic component.
Similarly figure 4.22 shows that on a calcium chloride subphase the film is also in
the viscoelastic range. However, in contrast to the film on water, the film on CaCl2
shows no significant low frequency elastic component. There is also a difference
between the monolayer’s behaviour at 16 mN/m and the behaviour above and below
this pressure. It is unclear why this reduction and then increase in viscosity is
occurring.
The behaviour of an α-elastin monolayer on water is consistent across the range
of frequencies shown. When this data is plotted against surface pressure at different
frequencies, as shown in figure 4.23 the general trend is for the phase lag to decrease
with surface pressure. At no point did δ indicate that the monolayer was behaving in
a elastic manner. It seems that with the molecules being packed closer together the
intermolecular forces increase and thus the monolayer becomes more elastic. The
difference between the low frequency measurements and the measurements above 7
Hz can also be seen in figure 4.23.
When the data from the calcium chloride subphase experiments are plotted
against surface pressure at different frequencies, as shown in figure 4.24, δ is consistent at oscillation frequencies up to 17 Hz. Elastic behaviour is not observed.
For oscillation frequencies below 17 Hz there is a maximum in δ at 14 mN/m and
a minimum at 16 mN/m. At the highest frequency this is reversed suggesting that
the higher frequency is probing a different phase in the monolayer. There is a lot
of variation in δ observed at surface pressures below 12 mN/m as the monolayer is
0
00
relatively dilute at these pressures and G and G are both small, thus experimental
errors are amplified in this region.
Aksenenko et al. [144] examined the behaviour of the phase lag with frequency at
the surface of β-lactoglobulin solutions. It was observed that a 10−6 mol/L solution
behaved elastically at oscillation frequencies between 0.01 Hz and 10 Hz with the
105
Figure 4.23: δ against surface pressure at constant applied oscillation frequency on water. The
Figure 4.24: δ against frequency at constant applied oscillation frequency on 0.1 M CaCl2 . The
monolayer consisted of 50 µL of 1 mg/mL α-elastin applied to 0.1 M CaCl2
106
phase lag decreasing with frequency. This is exactly the opposite behaviour observed
from α-elastin on water; however the surface pressure of the β-lactoglobulin was less
than 5 mN/m. It is not currently known how an α-elastin monolayer behaves at
lower pressures or frequencies.
4.5.3
Shear Viscosity Results
As described in section 2.4.4 and equation (2.17) in section 2.3.5, the shear viscosity,
ηs , of a monolayer can be calculated from the imaginary/dissipative component of
the complex shear modulus and the frequency of the applied oscillations. Figure
4.25 shows a plot of the shear viscosity against frequency for a range of applied
surface pressures. This data was taken from an α-elastin monolayer made up of
200 µL of 1 mg/mL α-elastin applied to water. Figure 4.26 is a similar plot from
a monolayer consisting of 50 µL of 1 mg/mL α-elastin applied to 0.1 M CaCl2 .
The data from the monolayer on np water is presented against surface pressure for a
range of oscillation frequencies in figure 4.27; and the data from the calcium chloride
subphase is presented in this form in figure 4.28.
Figures 4.25 and 4.26 show how the surface viscosity of an α film varies with
oscillation frequency on different subphases. It can be seen that on both subphases
increasing the surface pressure of the film increases the surface viscosity of the film.
However, changing the frequency of applied oscillation did not signficantly change
the viscosity of the surface. It is also notable that the α-elastin is much more
viscous on the water subphase; at 10 mN/m ηs on water is approximately four times
ηs on CaCl2 . The difference becomes less pronounced at higher applied surface
pressures but at 18 mN/m it is still a factor two. This is the opposite to what was
observed in the bulk solutions discussed above in section 4.4. However, as noted
above in section 4.5.1, due to the different amounts of α-elastin applied the area
107
00
Figure 4.25: Shear viscosity (G / ω) against oscillation frequency at a variety of surface pressures
on np water. The monolayer consisted of 200 µL of 1 mg/mL α-elastin.
Figure 4.26:
00
Shear viscosity (G
/ ω) against oscillation frequency at a variety of surface
pressures on 0.1 M CaCl2 . The monolayer consisted of 50 µL of 1 mg/mL α-elastin.
108
00
Figure 4.27: Shear viscosity (G /ω) against surface pressure at a variety of frequencies on np
water. The monolayer consisted of 200 µL of 1 mg/mL α-elastin.
00
Figure 4.28: Shear viscosity (G /ω) against surface pressure at a variety of surface frequencies
on 0.1 M CaCl2 . The monolayer consisted of 50 µL of 1 mg/mL α-elastin.
109
per molecule at a particular pressure is not the same on the two subphases. In
fact, the area per molecule on water at 10 mN/m (2970 Å2 /molecule) is comparable
to the area per molecule on CaCl2 at 18 mN/m (2990 Å2 /molecule). Comparing
the shear viscosity under these two conditions indicates that the calcium chloride
subphase produced a monolayer around twice as viscous as the water subphase. This
behaviour is consistent with the bulk viscometry results. Whether this is a general
property of ionic subphases or unique to the Ca2+ is not known at this time, however
comparisons with data in section 4.4 suggest that NaCl would increase ηs but not
by as much as CaCl2 .
Figure 4.27 shows the same data as figure 4.25 but plotted against surface pressure
at a variety of frequencies. It is clear that the increase in shear viscosity with surface
is linear at lower surface pressures and then deviates from this at surface pressures
above 15 mN/m. At most frequencies this deviation is reducing the viscosity; it
is suggested that this is due to increasing intermolecular forces. An anomalous
behaviour was observed at 9.3 Hz but the reason for this unclear.
Figure 4.28 shows the same data as figure 4.26 but plotted against surface pressure
at a variety of frequencies. The viscosity does not vary much with frequency at
low pressure, however, oscillation frequency does matter at higher applied surface
pressures. Unlike the water subphase the relation between surface viscosity and the
applied surface pressure is not linear.
Cejudo Fernández et al. [108] examine the shear viscosity - surface pressure
relation for several different monolayers. In the range that ηs was examined for αelastin in this work, β-lactoglobulin was observed to have very similar shear viscosity
until 10 mN/m and then the viscosity increases beyond what was observed on αelastin. It was also noted that a spread film of β-lactoglobulin has a much higher
viscosity than an adsorbed film at a given pressure. This is interesting with regard
to the adsorbed α-elastin monolayers that will be discussed in chapter 6.
110
4.5.4
Phase Change Discussion
As noted previously in sections 2.3.2 and 2.3.3, in the dilute regime there is very little
interaction between monomers, even those in the same chain. When a monolayer
enters the semi-dilute regime, the monomers are forced into contact with monomers
belonging to their own and other polymer chains. This regime change is indicated by
00
an increase in G . Figure 4.29 shows an example of lower pressure data from a water
subphase. 50 µL of 1 mg/mL α-elastin formed the monolayer. The characteristic
of a single frequency of probe oscillation, 12.8 Hz, is shown. It is clear from figure
00
4.29 that there is an abrupt increase in G between 8 and 9 mN/m which indicates
a phase transition. Data from other frequencies (not shown here) show very similar
behaviour.
Flory analysis carried out on quasi-static monolayers, discussed in section 5.4.5,
indicates that this is a transition between the semi-dilute and concentrated regime.
00
This is consistent with what is shown in figure 4.29 as in the dilute regime G should
be near zero and the minimum value observed on water was around 100 mN/m which
indicates that the monolayer is not dilute.
Figure 4.18 shows data at surface pressures comparable to those
presented in figure 4.29.
It clear
that the transition to the semi-dilute
regime occurs at around 8 mN/m,
the same as on the water subphase.
It therefore appears that this regime
change is unaffected by the presence
00
of an ionic subphase. It appears that
Figure 4.29: Graph of G against surface pressure
at lower surface pressures. The frequency of oscilla-
higher frequency probing causes the
tion of the probe was 12.8 Hz. 50 µL of 1 mg/mL
were applied to water to form this monolayer.
111
transition to occur at slightly lower pressures, but, as with the calcium chloride
subphase it seems that in the 3 - 20 Hz range the probing frequency makes little
differences when the transition occurs.
Spigone et al. [124] observed with a PVA (molecular weight 45.5 kDa, comparable
00
to α-elastin) monolayer that G against surface pressure shows three distinct regions
indicating two transitions. The first transition, identified as being between the dilute
and semi-dilute regimes, occurs at 7.5 mN/m. The second transition; between the
semi-dilute and concentrated regimes, and so comparable to the transition observed
from α-elastin, occurs at around 22 mN/m.
4.5.5
Eyring’s Model
In the concentrated regime data can be fitted to Eyring’s model for viscosity [145,
166, 167], which describes motion in a polymer monolayer in terms of sections of
the polymer chain moving relative to each other. The basic three-dimensional form
of Eyring’s model is:
ηs = Φ exp
E
kB T
(4.5)
where Φ is the surface area fraction occupied by molecules, E is the energy barrier
that the molecules have to overcome to move past each other [124, 167]. In the
presence of an externally applied pressure, P , equation (4.5) is altered to become;
ηs = Φ exp
E + PV
PV
= η0 exp
kB T
kB T
(4.6)
where V is the molecular volume. In two dimensions the molecular volume and
pressure are replaced by Am the area per fragment (in a polymer Am would be be
the monomer area), and surface pressure Π. This converts equation (4.6) to;
η = η0 exp
112
ΠAm
kB T
(4.7)
Figure 4.30: Water Subphase.
Figure 4.31: 0.1 M CaCl2 Subphase.
Lines of fit to Eyring’s Model on data from an α-elastin monolayer on a water subphase (figure
4.30) and on a CaCl2 subphase (figure 4.31). All data was taken at 23 ◦ C. On both figures only
the highest and lowest frequencies are shown.
00
G is proportional to η at a particular oscillation frequency, thus, from equation
00
(4.7) plotting the logarithm of G against Π allows calculation of Am . Figure 4.30
shows this plot for a water subphase. Only the highest and lowest frequencies are
shown. Figure 4.31 shows the fitting to Eyring’s Model for data taken on a 0.1 M
CaCl2 , again only the highest and lowest frequencies are shown.
The lines of fit shown on figure 4.30 indicate that the area occupied by a molecule
of α-elastin does not change with the frequency of the measurement. Figure 4.31
indicates that the same is true on CaCl2 . On water it is revealed that Am = 48
Å2 /fragment, which indicates that the length of the moving segments is 7.0 Å(to
the nearest Å). On calcium chloride it is seen that Am = 76 Å2 /fragment, which
indicates that the length of the moving segments is 9.0 Å. These values for area are
two orders of magnitude smaller than the areas per molecule already calculated in
section 4.5.1. This would imply that the α-elastin molecules are moving in segments.
That the segments are larger on the ionic subphase agrees with the calculated area
per molecule being larger on the calcium chloride.
Spigone et al. [124] observed with a PVA monolayer (molecular weight 275.5 kDa)
113
that the value of Am varied between 40 and 80 Å2 as the compression rate was altered
between 33 and 8 mm2 /s. This gives value for the segment length of between 6 and
9 Å, this is 3-5 times the molecular bond length between monomers and comparable
to the values calculated above. It is interesting that size of the segments moving in
this regime is similar in PVA and α-elastin despite the difference in their molecular
weights.
114
Chapter 5
Elastin Monolayers
5.1
Introduction
This chapter details experiments that examined the behaviour of α-elastin when
it is confined to an air-subphase interface. This two-dimensional system allows an
examination of the intra- and inter-molecular forces in a system that is simpler than
a bulk solution due to the extra restrictions on possible molecular conformations
placed on the elastin.
In this chapter the methodologies used to examine elastin monolayers and the
results obtained from those experiments are discussed. The methodology for the
quasi-static surface pressure-area measurements is described in section 5.2, while
the methodology for the dynamic surface pressure-area experiments is described in
section 5.3. The results of the quasi-static experiments are detailed in section 5.4.
First, data from the simplest system (α-elastin on a water subphase) are presented
and examined in section 5.4.1. Then data from experiments using an ionic subphase
are detailed in section 5.4.2; these enabled the study the effects of electrostatic
interactions on α-elastin. As explained in section 2.2.4 the calcium ion has been
observed to interact with bulk elastin in a different way to other ions and so a
comparison is made of Na+ and Ca2+ cations. As also described in section 2.2.4
115
and figure 2.7, the binding of ions to elastin has been shown to vary with pH in the
bulk; thus the behaviour of α-elastin on the three subphases will be examined above
and below neutral pH. Results from pH experiments are presented in section 5.4.3.
In all cases the quasi-static surface pressure isotherms are presented together with
the dilational modulus calculated from them. In order to examine the time scale
of relaxations within an α-elastin monolayer, further experiments were carried out
holding the monolayer at constant area for a period of tens of minutes. Results from
these experiments are given in section 5.4.4. Flory analysis of this data is described
in section 5.4.5.
The results of dynamic surface pressure experiments are presented in section 5.5.
This section starts comparing α-elastin’s Π-A characteristic on water under the two
methodologies. The dynamic methodology allows the simultaneous measurement of
the shear and dilational moduli from a monolayer; these are compared in section
5.5.1, which also includes further data from a calcium chloride subphase. Lastly the
effects of temperature, an important determinant of hydrophobic interactions, are
discussed in section 5.5.2.
5.2
Quasi-Static Methodology
The quasi-static measurements were carried out using a PTFE trough on an aluminium base. It was made in-house and has maximum surface dimensions of 125
mm by 45 mm and is 9.5 mm in depth. The transducer used a platinum alloy wire
probe. The compression rate for the quasi-static measurements was a maximum of
1.55 mm2 /s because preliminary experiments indicated that at this speed a repeatable smooth isotherm was generated. The calibration for the transducer, which was
performed before every experiment, was carried out by comparing the mass of the
probe when it was freely suspended in air to the mass of the probe when the tip was
116
touching the subphase (there was no monolayer present during the calibration).
As shown in figure 5.1, the trough
has a recess in the base which allows
the use of a magnetic stirrer (Rank
Brothers Ltd. Model 300 electronic
stirrer). This was necessary in experiments where the composition of the
subphase was altered. The pH was
varied with small additions of either
Figure 5.1: A schematic diagram of a side view of
HCl or NaOH. Once the subphase pH
the trough set-up as used for the quasi-static mea-
had been set the stirrer was switched
surements.
off. The pH of the subphase was measured using a Hanna Instruments pH
210 pH meter. For all of the quasi-static measurements 100 µL of 0.05 mg/mL
α-elastin was applied as described in section 3.4.2. This gives a minimum surface
concentration of 9×10−4 µg/mm2 at maximum area.
Figures 5.2 and 5.3 show the surface pressure plotted against time and against
specific area (that is area per mass) of applied elastin (Λ, defined in section 2.3.3)
respectively. In both graphs the red line is the first compression-relaxation cycle and
the blue line is the second. The second cycle was started approximately 15 minutes
after the first had ended. These figures are useful for illustrating the experimental
protocol used and also for demonstrating the differences between the first and second
compressions. The addition of elastin to the surface causes a slight rise in pressure
which can be seen at point A in figure 5.2. The expanded monolayer is left to
equilibrate for a few minutes and then it is compressed which causes an increase in
surface pressure.
When the film, after compression, reaches minimum area it is left in this state for
117
at least 10 minutes, (region B in figure 5.2) during which time the surface pressures
reduces as the monolayer relaxes. The monolayer is then expanded at the same slow
rate. At maximum area the monolayer is again held at constant area (region C in
figure 5.2). The compression-relaxation cycle is repeated, again pausing at minimum
area. The amount of hysteresis in a compression-relaxation cycle is indicative of the
rearrangement occurring in the monolayer over the cycle. Further discussion of
figures 5.2 and 5.3 will occur in section 5.4 below.
Figure 5.2: Dependence of the surface pres-
Figure 5.3: Surface pressure - area isotherm
sure of an α-elastin monolayer on time in the
for the same two compression - relaxation cycles
course of two consecutive compression - relax-
shown in figure 5.2. In each case the compres-
ation cycles. Point A is where elastin was added
sion part of each cycle occurs at higher surface
and regions B, C, and D, are at constant area.
pressure than that for the expansion.
In this experiment 100 µL of 0.05 mg/mL α-elastin was applied to a water subphase. The compression rate was 1.55 mm2 /s.
A similar difference between first and second compressions was noted on poly(vinylacetate) by Ferenczi and Cicuta [118]. It was seen that the difference decreased with
increasing temperature between 6 ◦ C and 36 ◦ C. A loss of concentration was observed
at the lowest temperatures indicated by a reduction in maximum surface pressure on
compression-relaxation cycles after the first. This was attributed to conformations
118
with long times scales being frozen out due to jamming effects and the monolayer
ending up in a regime where only short time scale configurations are possible. As
indicated by figure 5.3 above, there is no reduction in maximum surface pressure on
subsequent compression-relaxation cycles. Thus, what is occurring on the α-elastin
must be a different effect than observed from poly(vinyl-acetate). This effect will
be discussed in greater detail below.
The difference between the first and second compressions, a measure of the degree
of rearrangement taking place in the surface layer, can be analysed by calculating
the area inside the compression-expansion curve. This was done by fitting each
separate compression or relaxation with a polynomial and then integrating over the
region of pressure increase. Finally, the calculated result from each relaxation is
subtracted from the result from the compression, the result being the area enclosed
by the curve.
5.3
Dynamic Methodology
The dynamic measurements were made in a model 610 Nima Technology (UK)
Langmuir trough with a maximum surface area 530 cm2 utilising a type PS4 Nima
Technology (UK) microbalance sensor. The probes used were a pair of Willhelmy
plates made of filter paper 1cm in width. The paper was pre-soaked in the subphase
medium and produced a reproducible contact angle of ' 0◦ . The pressure was set
to zero before the addition of the α-elastin.
Each dynamic experiment was set up by reducing the trough area until the required pressure was reached and then the area was oscillated in a sinusoidal manner
about that area by ± 2 cm2 . The compression rate used for these measurements
was 83 mm2 /s and frequency was 0.04 Hz. Several measurements were made at each
temperature, each measurement at a different pressure. Each dynamic measurement
119
consists of between 8 and 10 compression-relaxation cycles which are then averaged
over the time for one cycle. At each pressure the oscillation frequency was constant.
While early work with oscillating barriers [133] used a single moving barrier to provide the area change, this work uses two barriers moving symmetrically. The size
of the area oscillation (∆A) was fixed across all experiments. This means that the
percentage area change (∆A / A0 ) varies, ranging from 0.4 % to 2 %.
Measurements of the monolayer anisotropy were carried out using the same experimental set-up as the dynamic measurements, except that two probes were used.
One of the probes was orientated parallel to the direction of compression and the
other perpendicular to the direction of compression as shown in figure 5.4. Before
the addition of the monolayer one of the probes was set to zero. The offset value of
the other probe was noted and subtracted from the data afterwards.
As in the quasi-static case the monolayers were
spread as described in section 3.4.2 The films were
allowed to equilibrate before the area change was
started. Up to 200 µL of 1 mg/mL α-elastin solution were used in the dynamic experiments. This
Figure 5.4: A schematic diagram
gives a minimum surface concentration of 3.8×10−3
of the trought set-up as used for the
µg/mm2 at maximum area. This is approximately
dynamic measurements.
four times as much as used in the quasi-static ex-
Probe A
is perpendicular to the barriers and
probe B is parallel to them.
periments.
The temperature changes for the dynamic mea-
surements were made using an integral bath that circulates water under the trough.
The temperature was measured directly in the water bath. Measurements were made
at room temperature (23◦ C) and then at temperatures of 28◦ C, 33◦ C and 40◦ C.
120
5.3.1
Analysis of the Oscillation Experiments Data
Section 2.3.5 details the theory describing behaviour of the elastic surface layers under oscillation. Here I discuss how
these equations can be used to measure useful parameters from monolayers.
As mentioned above this measurement
system relies on having two orthogonal
probes and symmetric sinusoidal barrier oscillations. The area change is described by equation (2.14). The surface
pressure measured by probe A in figure Figure 5.5: The surface pressure as a function of
5.4 is Π⊥ as defined by equation (2.19) time when the area (change is shown by squares)
and the surface pressure measured by is altered by the movement of oscillating barriers.
probe B is Πk which is defined by equation (2.18).
The surface pressure is recorded in two orthogonal directions (red triangle and blue inverted triangle) with respect to the direction of barrier mo-
Figure 5.5 shows part of a typical tion. The frequency of oscillation was 0.04 Hz. Π0
2
surface pressure - time measurement = 4 mN/m, ∆Π = 0.1 mN/m, A0 = 468.5 cm
from an oscillation experiment on an α-
and ∆A = 2 cm2 . The monolayer consisted of
400 µL of 0.5 mg/mL α-elastin.
elastin monolayer consisting of 400 µL
of 0.5 mg/mL elastin. The frequency of oscillation was 0.04 Hz. Π0 = 4 mN/m, ∆Π
= 0.1 mN/m, A0 = 468.5 cm2 and ∆A = 2 cm2 . The area change with the oscillation
is shown by the black squares. The change in surface pressure in the two orthogonal
directions (Π⊥ and Πk ) due to the area change produced by the oscillating barriers
over time is shown by the triangles. The red triangles indicate the surface pressure
readings from the sensor parallel to the barrier motion. The blue inverted triangles
indicate the readings from the sensor perpendicular to barrier motion. In this case
121
|Π⊥ - Π0 | = |Πk - Π0 |, which indicates that there is no significant shear modulus at
this surface pressure and oscillation frequency.
From each set of data the set of times at which A = A0 are extracted as are
the set of times where Π = Π0 for each probe. The difference in time (δt) between
equilibrium area and equilibrium pressure was calculated for each orientation of
sensor and then averaged to give δtk and δt⊥ . Values of ∆A and ∆Π area were
extracted from the data by comparing the maximum and minimum value during
oscillation with A0 or Π0 and then taking the mean of the difference.
With this information the complex dilational and shear moduli can be determined
from
∆Πk
∆A
(5.1)
∆Π
⊥
∗
∗
|ε − G | = A0
∆A
As equations (5.1) allow calculation of the complex elastic moduli, the storage and
|ε∗ + G∗ | = A0
dissipative components of each can then be calculated as
0
0
0
0
|ε + G | = |ε∗ + G∗ | cos ϕ
|ε − G | = |ε∗ − G∗ | cos ϕ
00
00
00
00
(5.2)
|ε + G | = |ε∗ + G∗ | sin ϕ
|ε − G | = |ε∗ − G∗ | sin ϕ
The phase angle, ϕ, is equal to 2πωδt as in equation (2.15) and the definitions quoted
0
00
for the components of the complex moduli, G = |G∗ | cos ϕ and G = |G∗ | sin ϕ.
This analysis allows the calculation of these components from the oscillation data.
5.4
Quasi-Static Results
Each set of data recorded consisted of two compression-relaxation cycles where each
half of a cycle could be analysed separately to evaluate the Flory exponent, dila122
tional modulus and degree of hysteresis in the system. In this section compressionrelaxation isotherms recorded from a variety of subphases are presented.
5.4.1
Nanopure Water Subphase
Figures 5.2 and 5.3 above show the surface pressure - time course and the Π-A
isotherm on a nanopure water subphase. Figures 5.6 and 5.7 show the equilibrium
dilational moduli calculated from the data shown in figures 5.2 and 5.3. Figure 5.6
shows the dilational moduli calculated from the first compression-relaxation cycle,
and figure 5.7 shows the dilational moduli calculated from the second cycle. The
data from compressions are shown in red and the data from relaxations are shown
in blue. The dilational moduli were calculated using equation (2.8).
Figure 5.6:
The dilational moduli as cal-
Figure 5.7: The dilational moduli as calcu-
culated from the first compression - relaxation
lated from the second compression - relaxation
cycle shown in figures 5.2 and 5.3.
cycle shown in figures 5.2 and 5.3.
In both figures, compressions are shown in red, relaxations in blue.
As shown in the surface pressure - area isotherm shown in figure 5.3 there is a
gradual rise in surface pressure upon compression of the monolayer. Surface pres123
sures as high as 22 mN/m were reached with no sign of collapse. Whether there was
multilayer formation or a return to the bulk phase was not clear. In comparison,
poly(ethylene oxide) which is (like α-elastin) amphiphilic and soluble in water, and
had a molecular mass of 144 kDa, was seen to form stable monolayers only up to 9
mN/m [168]. Above this surface pressure desorption into the bulk was seen.
During regions B and D, when the monolayer was held at constant area, rapid
relaxation occurred over approximately two minutes with a slower drift continuing
for approximately 10 min. It is unclear whether this arises from exchange between
the surface and bulk phase or changes in conformation of the surface phase.
Subsequent relaxation of the elastin monolayer shows hysteresis indicating that
there is a rearrangement or change occurring in the monolayer during compression.
This indicates that the α-elastin is not in a true equilibrium as the changes induced
in compression are not perfectly reversible. It was observed that the α-elastin surface pressure was consistent over two cycles up to 22 mN/m on a water subphase,
indicating that whatever reorganisation is occurring within the monolayer under
compression reverses after a relaxation.
Calculating the area per molecule of this monolayer is not easy, since α-elastin
is water soluble and it is possible that the applied monolayer may desorb from the
surface. Assuming that all the applied elastin stays on the surface gives a minimum
area per molecule. The range on the x-axis in figure 5.3 is thus converted to 1100 to
10010 Å2 /molecule. Thus at a surface pressure of 15 mN/m α-elastin has a specific
area of 160 mm2 /µg, which gives a minimum molecular radius of 24 Å assuming a
circular shape for each molecule. The molecular weight is 67 kDa. In comparison
at 15 mN/m β-lactoglobulin (molecular weight = 18 kDa) has a specific area of 67
mm2 /µg [109], β-casein (molecular weight = 24 kDa) at the same surface pressure
has a specific area of 43 mm2 /µg [109] and bovine serum albumin (molecular weight
= 66.4 kDa) has a specific area of 22 mm2 /µg at 15 mN/m [169]. From these
124
estimates it is clear that each molecule of elastin is occupying a much larger surface
area at this pressure than comparable molecules. This indicates that the inter
and intra-molecular forces generated by the elastin must be higher than the other
proteins to create this surface pressure while being less densely packed.
In figure 5.2 it is notable that even though the compressions were performed
slowly (the compression rate was 1.55 mm2 /s) in an attempt to maintain quasistatic equilibrium there is still a drop in pressure when the barriers stop moving.
This is evidenced by regions B and D. This suggests that the layer has not reached
its lowest energy conformation. It can be seen in figure 5.25 (one of the experiments
showing relaxations later in this chapter) that the equilibrium pressure after the
compression is reached after approximately 3000 seconds. This shows that some
reordering or desorption is occurring in the α-elastin on longer timescales. The
conformation changes occurring in the monolayer during this relaxation are likely
responsible for the differences in the isotherm characteristics between compression
and relaxation.
Figures 5.6 and 5.7 show the equilibrium dilational modulus εeq of a typical αelastin monolayer on water. The first and second compression - relaxation cycles are
shown to have consistent dilational moduli, in terms of the maximum modulus for
each half of the cycle and the form of the dilational modulus curve. It is interesting
to observe the clear differences between compression (red) and relaxation (blue) on
the same cycle. The expanding monolayer has a higher dilational modulus than
the same monolayer under compression. It is, therefore, suggested that the relaxed
state (produced by the wait at constant area) of the α-elastin monolayer is less
compressible than the unrelaxed state. From the relaxation data it is clear that
there are reorganisation processes taking place in the monolayer on timescales longer
than the compression, this will be discussed in more detail in section 5.4.4. This
may also explain why the peak in the expanding monolayer’s dilational modulus
125
is broader as it is transiting from the relaxed state through the less relaxed state
before changing phase. Thus, the dilational moduli calculated may be equilibrium
for different arrangements of α-elastin molecules. On the compression the dilational
modulus goes through a maximum at 6-7 mN/m, however, on the relaxation the
peak is broader and at a higher applied surface pressure. This means that the
film is at its least compressible (β = βmin ) state at this point. It therefore seems
that the monolayer must go through a phase transition at this pressure to account
for the monolayer’s greater compressibility at higher surface pressures. Similarly
Monroy et al. [122] observed that the dilational modulus of a PVA monolayer went
through a maximum at 12 mN/m. This was attributed to the transition between
the semi-dilute and concentrated regimes.
The surface pressure of this transition agrees with the transition observed in
the surface viscometry results and discussed previously in section 4.5.4, indicating
that the semi dilute - concentrated transition has been observed by two different
methodologies. The shear modulus behaviour indicates the transition occurs at
around 7370 Å2 /molecule, while the quasi-static dilational modulus indicates the
transition occurs at around 3890 Å2 /molecule. It is suggested the the quasi-static
monolayer undergoes phase transition at a lower area per molecule due to the near
equilibrium state of the monolayer. In contrast, the dynamic monolayer was not
near equilibrium.
As detailed in section 2.3.2 in the dilute phase the molecules making up a monolayer are assumed to be completely on the subphase surface and not forced to interact with each other. In the semi-dilute regime the elastin molecules are still
on the surface but close enough together that interactions between them become
significant. In this regime the molecules are still on the surface but due to the intermolecular interactions are forced to reorientate. When the monolayer transitions to
the concentrated state, some of the hydrophilic regions of the elastin start enter126
ing the subphase which results in a more compressible state. Under relaxation the
broader peak is explained because as the hydrophilic regions start coming out of the
subphase the density of protein on the surface remains constant even as the area
expands. Thus, the monolayer does not start to become more compressible again
until all the protein chains are back on the surface.
5.4.2
Ionic Subphases
Figure 5.8 shows the first compression-relaxation cycle from α-elastin applied to a
0.1 M NaCl subphase. Similarly figure 5.9 shows the first compression-relaxation
cycle from α-elastin applied to a 0.1 M CaCl2 subphase. Again the compression rate
was 1.55 mm2 , these experiments were carried out at neutral pH.
Figure 5.8:
The surface pressure-area
Figure 5.9:
The surface pressure-area
isotherm of a first compression - relaxation cy-
isotherm of a first compression - relaxation cy-
cle on α-elastin on a 0.1 M NaCl subphase.
cle on α-elastin on a 0.1 M CaCl2 subphase.
In both experiments 100 µL of 0.05 mg/mL α-elastin was used to make the monolayer and the
Compression rate was 1.55 mm2 /s.
Comparing figures 5.8 and 5.9 with figure 5.3 shows that while the surface pressure
127
increases with decreasing surface area regardless of the subphase, there is a difference
in the way the surface pressure increases on the ionic subphases compared to the
water. On the ionic subphases the rate of increase of Π is more constant with
decrease in A. Calcium chloride and sodium chloride show a very similar rate of
increase, approximately 5 mN/m increase for a 200 mm2 /µg decrease in surface area.
The higher surface pressures, compared to the water subphase experiments, shown
in figures 5.8 and 5.9 agree with what was observed in section 4.5.1 where it was
seen that at a given pressure α-elastin on CaCl2 has a greater area per molecule
than a comparable monolayer on water.
Figure 5.10 shows the dilational moduli calculated from the surface pressure area
isotherm shown in figure 5.8. Similarly, figure 5.11 shows the dilational moduli
calculated from the Π-A isotherm shown in figure 5.9. In both figures 5.10 and 5.11
the dilational modulus calculated from the compression is shown in red and the
modulus from the relaxation in blue.
As noted above, a maximum surface pressure of 22 mN/m was observed on a
water subphase, slightly higher surface pressures (24 mN/m) were observed on the
salt subphases. It does not appear that the monolayer is collapsing at this pressures
since the pressure is repeatable on the same monolayer. A larger pressure increase
is observed on the initial addition of elastin to an ionic subphase, as seen in figures
5.8 and 5.9 compared to a water subphase. Overall the increase in surface pressure
was larger on the CaCl2 than on NaCl (3.6 mN/m versus 3.0 mN/m), however the
range was ± 1 mN/m for both sets of calculations. These values are significantly
higher than the initial increase in pressure for a water subphase, 1 mN/m.
The higher surface pressure increase on addition of elastin and the higher overall
pressure on the ionic subphases suggests an increase in the interactions between the
elastin molecules or between the elastin molecules and the subphase. It is clear
that the ionic subphase will alter the interactions between the subphase and the
128
Figure 5.10:
The dilational moduli calcu-
Figure 5.11:
lated from the Π-A isotherm in figure 5.8. This
lated from the Π-A isotherm in figure 5.9. This
experiment used a 0.1 M NaCl subphase.
experiment used a 0.1 M CaCl2 subphase.
In both figures 5.10 and 5.11 the red curve shows the modulus calculated from the compression
and the blue curve shows the modulus from the relaxation.
129
elastin molecules but it is also possible that the change in subphase will alter the
elastin-elastin interactions, either by changing the net charge of the elastin, which
was not seen in the bulk titration experiments shown in figure 3.1; or by changing
the α-elastin orientation on the subphase surface. However, the surface pressure area isotherms show no consistent differences between Na+ and Ca2+ that would
indicate a preferential binding of Ca2+ to α-elastin in two-dimensions.
As can be seen from figures 5.8 and
5.9 after the first relaxation of a monolayer on an ionic subphase the surface
pressure is much lower than at the start
of the first compression. Thus, second
compression relaxation cycles on ionic
subphases, as shown in figure 5.12, do
not show the initial near linear increase
in surface pressure as surface area is decreased. It is presumed that the reducFigure 5.12: Surface pressure - area isotherms tion in initial pressure occurs due to a
from the second compression - relaxation cycles more even distribution of the elastin on
of the experiments whose first compression are
the surface rather than a transition of
shown in figures 5.8 (here shown in red) and 5.9
the elastin to the bulk phase. If the lat-
(here shown in blue).
ter were the case then it would be ex-
pected that the second compression-relaxation cycle run on an ionic subphase would
have a lower maximum surface pressure, but this is not observed. It is believed that
ionic interactions between the subphase and the α-elastin have inhibited the elastin’s
initial ability to spread on the surface. Elastin has charged regions and these interact
with the ions in the subphase. It is suggested that this extra interaction changes the
surface conformation of the α-elastin. It is seen that compressing and relaxing the
130
monolayer lowers the surface pressure at maximum area after the cycle compared to
before the cycle. This suggests that the motion of the barriers and the compression
of the monolayer induces different arrangement of elastin molecules.
Comparing figures 5.10 and 5.11 with figure 5.6 shows how the dilational modulus
of the monolayer has altered with the change in subphase. The relaxation phase
shows that while the location of the peak in dilational modulus is similar across the
three subphases (9 ± 1 mN/m) suggesting that the surface pressure at the point of
phase transition is not been altered by the ionic subphase. However at the peak, the
monolayers on the ionic subphases only had dilational modulus of 15 ± 2 mN/m.
This compares to a peak value on water of 21 ± 2 mN/m. α-elastin monolayers on
ionic subphases are less compressible than monolayers on water. As noted above,
the ionic subphase monolayers had a larger area per molecule at a given surface
pressure which either indicates a greater distance between molecules leading to a
reduction in inter-molecular forces, or a difference surface conformation.
Thus, the decrease in the compressibility of the monolayer must be caused by
subphase - elastin interactions. Supporting the supposition that after the first compression the α-elastin monolayer arrangement has changed in some way is that the
dilational modulus at the peak (phase transition) on the second relaxation is 19
± 2 mN/m. This value is consistent with the value observed on water. Together
with the form of the isotherms shown in figure 5.12 this suggests that an α-elastin
monolayer on an ionic subphase does behave in a similar manner to a monolayer on
water, however, it takes the monolayer on the ionic subphase a while to reach this
state.
The dilational modulus calculated from the first compression on an ionic subphase
shown in figures 5.10 and 5.11 is obviously different from the comparable modulus
on a water subphase. For a Ca2+ subphase, the dilational modulus shown in 5.11 is
most obviously different from the water subphase characteristic, as it does not have a
131
peak. Similar behaviour was observed in some experiments on the sodium chloride
subphase so this is not a characteristic specific to the Ca2+ ion. It is suggested
that the initial spreading of the α-elastin surface is not consistent because α-elastin
can become arrested into a metastable state on the subphase surface resulting in
the incompressible initial state. When the barriers start to move the monolayer
becomes more compressible indicating that the protein has assumed a more compact
conformation. It is likely that this monolayer is in the concentrated regime due to the
surface are per molecule, however it appears that the monolayer was in this regime
right from application. What caused this highly incompressible state is not clear
but it was not a permanent change in the monolayer since the second compression
- relaxation cycle shows characteristics similar to monolayers on a water subphase.
Due to the initial pressure rise on the addition of the elastin the monolayers on
ionic subphases show a much larger hysteresis on the first compression-relaxation
cycle that the monolayers on water. It is suggested that this difference is a result of the elastin molecules assuming a different conformation when applied to an
ionic subphase compared to the water subphase. It is suggested that electrostatic
interactions between the ions and the elastin cause the elastin to assume a larger
conformation when initially spread and then when the monolayer is compressed the
α-elastin molecules assume a different, smaller conformation and do not revert to
the first conformation even after the monolayer is relaxed.
5.4.3
Effect of pH
Changes to the pH of the subphase were made prior to the addition of the α-elastin
monolayer. Figures 5.13 and 5.14 show the behaviour of an α-elastin monolayer at
pH 9 and pH 3.5 respectively.
Figure 5.15 presents the dilational moduli calculated from the isotherm in fig-
132
Figure 5.13: A Π-A isotherm measured on a
water subphase at pH 9.
water subphase at pH 3.5.
ure 5.13. Figure 5.16 below shows the dilational moduli calculated from the Π-A
isotherm shown in figure 5.14. In both figures, the dilational modulus from the
compression is shown in red and the modulus from the relaxation is shown in blue.
It was seen that the highest surface pressure obtained on an alkaline subphase
was 18 mN/m which is lower than on the water or salt subphases. Similarly, the
monolayer on the acid subphase had an maximum observed surface pressure 18
mN/m. This surface pressure was reproducible and the monolayer showed no signs
of collapse. pH 3.5 is below the isoelectric point [16] thus α-elastin is positively
charged at this pH. Figure 3.1 in chapter 3 indicates that at pH 3.5 α-elastin on
water has 50 µmole bound H+ per gramme of dry elastin which is consistent with
the protein being positively charged. pH 9 is above the isoionic point [27] and the
isoelectric point thus the elastin is negatively charged, figure 3.1 indicates that at
this pH α-elastin on water has -50 µmole bound H+ per gramme of dry elastin
which is consistent with this. A molecule of α-elastin which is charged will repel
other similarly charged molecules and other parts of the same molecule, this would
133
Figure 5.15:
Figure 5.16:
lated from the isotherm in figure 5.13. Sub-
lated from the isotherm in figure 5.14. Sub-
phase pH = 9.0.
phase pH = 3.5.
In both figures 5.16 and 5.15 the red curve shows the modulus from the compression and the blue
curve that from the relaxation.
134
suggest that at higher and lower pHs elastin should assume a larger conformation
and thus produce a higher surface pressure. This is not what is observed. It is
therefore suggested that the lower surface pressures seen on the acidic and alkaline
subphase may be due to parts of the elastin penetrating the surface of the subphase
reducing inter and intra-molecular repulsions.
Figure 5.15 shows the dilational moduli calculated from an α-elastin monolayer
at pH 9.0. Comparing it with figure 5.6 shows the changes induced by the raising of
the pH. It is clear that there is a difference between compression and relaxation on
an alkaline subphase. Under compression the monolayer is more compressible than
a comparable monolayer on a water subphase, the maximum value of εeq calculated
for the alkalinised monolayer are two thirds that seen on the monolayer on water.
The peak in the compression modulus is very broad and so it is difficult to pinpoint
where the phase transition occurs. The relaxation dilational modulus shows that
the monolayer is less compressible after relaxing at minimum area, the peak value
is double that of the compression monolayer and comparable to that shown by the
monolayer on water. This suggests that the phase transition occurs very gradually
on the alkaline monolayer as the pressure increases and that the phase transition
appears to continue at constant area leading to the much more incompressible state
seen in the relaxation.
Figure 5.16 shows the dilational moduli calculated from an α-elastin monolayer
at pH 3.5. Comparing it with figure 5.6 shows the changes induced by the lowering of the pH. The difference between compression and relaxation on an acidic
subphase is even greater than on an alkaline subphase. Under compression the dilational modulus is approximately two-thirds that of the monolayer on water, very
similar to that of the monolayer on the alkaline subphase. The monolayer is clearly
less compressible after the time relaxing at constant area as shown by an almost
threefold increase in dilational modulus. The highest value of the relaxation dila135
tional modulus peak was observed on the acidic subphase, which would indicate that
under acidic conditions, and so being positively charged, α-elastin monolayers form
the stiffest structure. However, the range of values obtained from these monolayers
was also the greatest suggesting a lack of consistency in the monolayer formation,
On acidic subphase monolayers the peak value εqs = 24 ± 6 mN/m, on alkaline
subphases εqs = 21 ± 4 mN/m, and on water εqs = 22 ± 2 mN/m, which indicates
no statistical difference between the monolayers on different subphases.
The narrow relaxation peak and broad compression peak seen in figures 5.15
and 5.16 are opposite to what is observed on neutral pH subphases. This indicates
that the transition in relaxation is occurring much more quickly. It is concluded that
altering the pH changes the initial spreading of α-elastin making the monolayer more
compressible. Then, when the monolayer is able to relax at fixed area it becomes
less compressible and more like the monolayers observed on water. Interestingly
this difference between compression and relaxation is not simply a function of initial
inhibited spreading as the second compression - relaxation cycle shows very similar
behaviour with the relaxation showing a clear peak in dilational modulus and overall
high values than the compression which is relatively constant with applied surface
pressure.
Figures 5.17 and 5.18 show data comparable to that shown in 5.13 and 5.14 but
recorded on a 0.1 M NaCl subphase. Figures 5.19 and 5.20 show the dilational
moduli calculated from the Π-A isotherms shown in figures 5.17 and 5.18 respectively. Likewise, figures 5.21 and 5.22 show comparable data recorded on a 0.1 M
CaCl2 subphase, and figures 5.23 and 5.24 show the respective dilational moduli.
The compression rate for the alkaline experiments was 0.92 mm2 /s.
Comparing the isotherms in figures 5.21, 5.22, 5.17 and 5.18 with isotherms taken
on the same subphase with unaltered pH it is seen that the form of isotherm is
relatively consistent across the two subphases and two pHs. It is notable that in all
136
0.1 M NaCl subphase at pH 9. The compression
0.1 M NaCl subphase at pH 3.5. The compres-
2
rate was 0.92 mm /s.
sion rate was 1.55 mm2 /s.
lated from the isotherm shown in figure 5.17.
In both figures 5.23 and 5.24 the compression is shown in red and the relaxation in blue.
137
Figure 5.22: A Π-A isotherm measured on
0.1 M CaCl2 subphase at pH 9. The compres-
a 0.1 M CaCl2 subphase at pH 3.5. The com-
2
sion rate was 0.92 mm /s.
pression rate was 1.55 mm2 /s
In both figures 5.23 and 5.24 the compression is shown in red and the relaxation in blue.
138
cases there is a slight initial pressure rise on the addition of α-elastin to the subphase.
The alkaline sodium chloride subphase produced an average surface pressure increase
of 1.1 mN/m with a range of 1.6 on addition of elastin, this is comparable to water.
This suggests that changing the amount of H+ bound to the elastin allows it to
spread freely on the subphase surface. It is therefore indicated that electrostatic
interactions between the elastin and an ionic subphase govern the elastin’s initial
spreading. The acidic sodium chloride subphase produced a 7.5 mN/m range of
results of initial pressure rise. The alkaline calcium chloride subphase produced an
average Π rise of 4 mN/m, which is consistent with the neutral CaCl2 , and the range
was 3.1 mN/m. Lastly, for the acidic CaCl2 subphase, the average increase in Π is
8 mN/m with a range of 4.5 mN/m. Data from all the acidic subphases (the acidic
water initial increase in Π on application of α-elastin had a range of 3.7 mN/m)
showed a wide variation in surface pressure increase when α-elastin was applied to
the surface, which suggests that elastin does not spread consistently at pH 3.5.
Despite the differences in initial spreading, surface pressure rises on compression
were consistent on first and second compressions indicating this is not affected by
the initial state. The compression dilational moduli shown in figures 5.19, 5.20, 5.23,
and 5.24 indicate that like the monolayer on neutral CaCl2 shown in figure 5.11, the
initial state of the monolayer is very incompressible. The monolayer then becomes
more compressible as surface pressure increases, it is suggested that the monolayer
is undergoing a transition during the compression.
The relaxations on the sodium chloride and on the alkaline calcium chloride all
show a peak in the dilation modulus indicating a phase transition. This transition
occurs at a similar surface pressure to other monolayers studied and is thus speculated to be the same transition. These peaks are also consistent in value of the
dilational modulus indicating that the compressibility of the monolayer is unaffected
by these subphases. The monolayer on acidified calcium chloride, dilational modulus
139
data shown in figure 5.24, shows different behaviour to the other monolayers. The
compressibility to the monolayer is greater under relaxation than any of the other
monolayers and the peak indicating phase transition is not clear. It is suggested
that there is a peak but its is very broad, the drop in dilational modulus above 20
mN/m indicates this. Despite the inconsistencies seen in the initial spreading of the
monolayer, the dilational moduli of subsequent compression - relaxation cycles do
show a clear peak indicating that even on this subphase the α-elastin does find an
equilibrium surface state similar to its state on water.
5.4.4
Relaxation Experiments
In order to study these changes an αelastin monolayer was compressed on
a water subphase, then the area was
held constant in the compressed state
for an hour and the pressure drop due
to the film relaxation was observed.
Figure 5.25: A Π-t dependence measured on wa- The experiment was then repeated on
ter, showing the gradual pressure relaxation seen at
0.1 M CaCl2 and 0.1 M NaCl.
constant area (region A).
The expansion experiments were
done quasi-statically to eliminate the short relaxation time processes. Figure 5.25
shows a single compression - expansion cycle plotted as Π against time. The monolayer consisted of 100 µL of 0.05 mg / mL α-elastin. The compression rate was
1.55 mm2 /s. This data is directly comparable to that shown in figures 5.2 and 5.3.
Figures 5.26 and 5.27 show similar data obtained from monolayers on a calcium
chloride subphase and a sodium chloride subphase respectively.
It was observed (figure 5.25) that the pressure initially dropped quickly indicating
140
Figure 5.26: A single compression-relaxation
Figure 5.27: A single compression-relaxation
of an α-elastin monolayer carried out on 0.1 M
of an α-elastin monolayer carried out on 0.1 M
calcium chloride subphase.
sodium chloride subphase.
In both experiments the monolayer consisted of 100 µL of 0.05 mg/mL α-elastin monolayer, and
the compression rate was 1.55 mm2 /s.
a fast rearrangement of the protein. The pressure then levelled off indicating that
an equilibrium conformation had been reached. On the calcium chloride subphase
the pressure drop was much slower and continued throughout the entire time that
the monolayer was held at minimum area, it does not seem that an equilibrium
situation is reached. As the elastin film behaves in very similar manner on NaCl it
seems that this slowing down of the protein rearrangement is an ionic effect rather
than a specific effect of calcium
In figures 5.25, 5.26 and 5.27 the monolayer is still relaxing after an hour at
minimum area. Similarly after the monolayer has been expanded to maximum area,
the surface pressure continues to rise over the same timescale. It is clear that these
monolayer did not reach equilibrium in the time (5000s) available.
Examining the dilational moduli calculated from 5.25, 5.26 and 5.27 shows no
141
Figure 5.28:
Water sub-
phase.
Figure 5.29:
CaCl2 sub-
Figure 5.30:
NaCl sub-
phase.
phase.
Exponential fits, shown in red, to the surface pressure decays at constant area shown in figures
5.25, 5.26 and 5.27 above.
significant differences from the monolayers which did not have an hour to relax,
indicating that changes that affect the compressibility of the monolayer occur very
early on in the monolayer’s relaxation.
The rate of decay of the surface pressure can be examined by fitting an exponential
of the form Π(t) = c exp (−λt)+Π∞ ; where λ is the decay constant, t is the time and
v is the vertical offset of the surface pressure, and c is a constant; to the relaxations
shown in figures 5.25, 5.26 and 5.27. In each experiment Π∞ was determined from
the data. The time constant of the decay, τ , is defined as τ = 1/λ.
From these exponential fits the time constant was calculated and was found to be
(1.0 ± 0.3)×10−3 s for all three subphases. Calcium chloride showed the most consistent surface pressure decay with an average time constant of (1.32 ± 0.01)×10−3
s which is interesting given the inconsistent spreading seen on the ionic subphases.
This value for τ indicates that variations in initial spreading of a monolayer are
erased by compressing the monolayer. There was no statistical difference between
the three subphases in this series of experiments which suggests that the relaxation
processes are not limited by electrostatic interactions between the subphase and the
142
monolayer.
5.4.5
Flory Analysis of Quasi-Static Data
As described in section 2.3.2, the Flory exponent ν may be calculated by fitting a
power law to a plot of Π against A then
applying equation (2.3). Figure 5.31 shows
the fitting of a power law to one half of a
compression relaxation cycle. This is the
first compression shown in figure 5.3. For
first compressions on water the Flory exponent was remarkably consistent with a
mean value of 0.671 ± 0.002, which indi-
Figure 5.31: Flory analysis of the data from cates that the monolayer is between good
figure 5.3 For this set of data the Flory Ex-
and Θ solvent conditions. All the data ob-
ponent ν = 0.673, indicating that the monolayer is between good and Θ solvent condi-
tained on a neutral water subphase indi-
tions. This result was typical of those ob-
cated that the film remained in these con-
tained on water (first compression).
ditions over two compression-relaxation cycles. Reproducible values for ν indicate
that the monolayer must be in the semi-dilute regime. Figure 5.31 shows that the
semi-dilute regime holds to a surface concentration of 400 mm2 /µg. If it is assumed
that all the applied α-elastin is still on the surface at this point this corresponds to
an area of 4450 Å2 per molecule. In section 4.5.4 the phase transition from the semidilute to concentrated regime is observed, the value for the surface concentration at
8 mN/m in figure 4.29 is 663 mm2 /µg applied elastin, which corresponds to 7370 Å2
per molecule. This is not consistent with the quasi-static data discussed here and it
is suggested that the difference is due to the difference in compression speeds. The
143
compression speed used to obtain the different pressures in the experiment discussed
in figure 4.29 was 50 mm2 /s. Thus, the surface pressure measured from a monolayer
is not only dependent on the surface concentration of elastin. The phase transition
occurs at a lower surface concentration when the compression is faster.
Due to the large pressure rise when the monolayer was applied to an ionic subphase fitting to the first compression was impossible, however, examining the second
compression-relaxation cycle was possible. Monolayers on a sodium chloride subphase showed that over the whole cycle they were also between Θ and good solvent
conditions conditions, with νN aCl = 0.70 ± 0.04. α-elastin monolayers on calcium
chloride subphase were also found to be close to good solvent conditions, νCaCl2 =
0.7 ± 0.2. It is suggested that this larger error is the result of inconsistent spreading
and unclear partitioning between bulk and surface. There was no significant difference between the two counterions, which is consistent with our bulk viscometry
data, despite the established views, as discussed in section 2.2.4, that calcium has
specific interactions with elastin.
These numbers defining the initial phase of the monolayers allow the phase transition observed in the shear modulus results in section 4.5.1 and in the dilational
moduli results discussed above to be firmly labelled as the semi-dilute to concentrated phase transition.
Further Flory analysis carried out on pH varied monolayers gave a variety of
inconsistent results suggesting that equation (2.3) is being applied outside its realm
of validity, likely due to uneven spreading. From observations of the dilational
modulus it is indicated that under all conditions α-elastin monolayers are in the
semi-dilute regime.
A monolayer of poly(methyl methacrylate) on water was found to have a Flory
exponent of 0.55 ± 0.02 [126] and 0.53 [127] (both poor solvent conditions), atactic
polyvinylacetate on water was found to have a a Flory exponent of 0.78 ± 0.03
144
[118, 123, 170], atactic poly(methylacrylate) on water was found to have a Flory
exponent of 0.78 ± 0.01 (extended chain conditions) [127]. β-lactoglobulin was found
to have a Flory exponent of 0.61 (mid way between Θ and good solvent conditions)
at neutral pH [107], both β-lactoglobulin and β-casein had previously been shown to
have Flory exponents which changed with pH [132]. β-casein had a Flory exponent
of 0.6 at pH 7.2 [132]. Poly(ethylene oxide), which as previously noted is amphiphilic
and water soluble, had ν = 0.78 ± 0.01 [170]. It is unsurprising that poly(methyl
methacrylate) was in worse solvent conditions than poly(methylacrylate) since the
extra methyl group per monomer would tend to make the polymer less soluble. βlactoglobulin is a globular protein while β-casein has a random coil structure, thus
the Flory exponent can not be directly related to the structure of the protein.
There is no significant difference between a monolayer’s behaviour in the presence
of Na+ and its behaviour in the presence of Ca2+ . This is consistent with the bulk
viscometry data discussed in section 4.4, despite suggestions that the calcium ion
has specific interactions with elastin [79, 80].
Other Calculations based on ν
Many other properties of a monolayer can be described in terms of ν or y, where as
detailed in section 2.3.3, y = (2ν/(2ν − 1)). The relationship between the surface
pressure and dilational modulus can also be used to determine the Flory exponent.
Equation (2.2) indicated that Π = kΓy , where k is constant for a given monolayer
at a given temperature. Taking the natural logarithm of this equation and then
differentiating gives,
dΠ
= yΠ
d ln Γ
(5.3)
Equation (2.8) gives a definition for the dilational modulus, in the case where the
monolayer is insoluble in the subphase the surface concentration is proportional to
145
the inverse of the surface area, and thus the dilational modulus can be written as,
ε=
dΠ
d ln Γ
(5.4)
The left hand side of equation (5.3) is
therefore equal to ε, which means that y =
ε/Π. This indicates that there is a linear
relationship between the surface pressure
and the dilational modulus in the semidilute regime [125]. Benjamins et al. [171]
showed that this relationship holds from 0
- 6 mN/m for ovalbumin and bovine serum
albumin on both oil/water and air/water
Figure 5.32: A linear fit to the low pressure
interfaces. This range is consistent with
region of a surface pressure against dilational
the range over which the Flory fitting was
modulus plot. In this instance y = 3.60.
done on α-elastin.
The work of Aguié-
Béghin et al. [125] goes into further details
discussing a multiblock model to discuss the behaviour of polymers at a gas-liquid
interface and defines many properties in a variety of concentration phases.
Figure 5.32 shows a plot of Π against ε from a first compression of an elastin
monolayer on water. The gradient was 3.60. On average the range over which the
fit was good was from 0 - 4 mN/m.
5.5
Dynamic Measurements
Figure 5.33 shows the two simultaneous isotherms measured parallel and perpendicular to barrier motion from monolayer made of 200 µL of 1 mg/mL α-elastin. The
compression rate was 50 mm2 /s.
146
The orientation of the probes did not affect the values of the pressure measured,
which indicates the monolayer has no significant shear modulus, and is thus isotropic.
Like the quasi-static ionic subphases (discussed in section 5.4.2 above) the monolayer
showed a near linear increase in surface pressure, approximately 5 mN/m for a 65
mm2 reduction in area, which is faster than that seen in the quasi-static experiments.
Unlike the monolayers on ionic subphases the dynamic data shows a linear decrease
with increasing surface area.
A higher pressure is reached in this
measurement than was reached in the
quasi-static measurements detailed in
section 5.4.
It can be seen from fig-
ure 5.33 that the slope of the isotherm
changes at 75 mm2 / µg elastin applied.
It is suggested that this large rise in surface pressure with only a small change
in surface area indicates that the monolayer is in an incompressible phase. As
Figure 5.33: A graph comparing measurements
noted above in section 5.4.5 the semi-
of surface pressure made by the probe A (black)
dilute to concentrated phase change oc-
which was perpendicular to the trough’s long edge
curs at a lower surface concentration in
with probe B (pale blue) which was parallel to the
dynamic experiments. It therefore ap-
long edge.
pears that this gradient changed occurs
as a result of a phase change in the monolayer (the new phase being more rigid).
As the semi-dilute to concentrated transition has been located, this higher surface
concentration transition is concentrated to melt.
147
5.5.1
Dynamic Elastic Moduli
Figure 5.34 shows how the dynamic elastic moduli vary with surface pressure. It is
clear that as indicated by the lack of difference between parallel and perpendicular
measurements in the isotherm in figure 5.33 the shear modulus is much smaller than
the dilational modulus. The dilational modulus is an order of magnitude greater
than the shear modulus, and varies with applied surface pressure. The smaller
shear modulus and its lack of variation with surface pressure indicates that the
monolayer is behaving as an isotropic fluid and possesses no long range order. The
shear modulus data shown in figure 5.34 is comparable to the data from the ring
rheometer experiments shown in figure 4.13.
Figures 5.35 and 5.36 show a comparison between the quasi-static dilational moduli (compression and relaxation) and the dynamic dilational modulus on water and calcium chloride subphases.
Figure 5.35 shows the same
data as figure 5.6 and the dynamic dilational moduli calculated from a single
α-elastin monolayer at different surface
pressures (black symbols). Figure 5.36 Figure 5.34: Data from a series of dynamic exshows the same data as figure 5.11 (com- periments at 23◦ C using α-elastin monolayers on
pression in red and relaxation in blue) water showing how the dilational modulus (solid
and the dynamic dilational moduli cal-
symbols) and shear modulus (open symbols) vary
with surface pressure. The pairs of points (one
culated from a single α-elastin mono-
light blue and one dark blue) were measured si-
layer at different surface pressures (in multaneously by the parallel and perpendicular
black).
sensors.
148
On water the dilational modulus peaks at an applied surface pressure of 10 mN/m
which is comparable to the quasi-static data. The value of the dynamic modulus is
slightly higher, this probably reflects the lack of relaxation in the dynamic measurements. Figure 5.35 also suggests that the peak in dilational modulus at around 10
mN/m is a local maximum with further increases in dilational modulus at surface
pressures greater than 20 mN/m. The minimum in dilational modulus at 20 mN/m,
which indicates a maximum in compressibility (β = βmax ), is suggested to be the
end of the phase transition between the semi-dilute and concentrated phases which
began at the maximum in dilational modulus.
Figure 5.35: A comparison between the dy-
Figure 5.36: A comparison between the dy-
namic (black) and quasi-static (red and blue)
namic (black) and quasi-static (red and blue)
dilational moduli of an α-elastin monolayer on
dilational moduli of an α-elastin monolayer on
a water subphase plotted against surface pres-
a 0.1 M CaCl2 subphase plotted against surface
sure. The quasi-static data was previously pre-
pressure. The quasi-static data was previously
sented in figure 5.6.
presented in figure 5.11.
On calcium chloride the value of the dynamic dilational modulus at its peak
is higher that the quasi-static modulus at the same pressure by a factor of 1.5.
This suggests that on the ionic subphase the α-elastin is stiffer when probed by the
149
dynamic methodology. The peak is not significantly different from the characteristic
of the monolayer on water.
The dynamic data shown in figures 5.34 and 5.35 can be compared with data
for β-lactoglobulin and β-casein presented by Cicuta [109]. Both of these proteins
are amphiphillic and water soluble. These experiments were carried out with a fast
compression and then examining the monolayer as the surface pressure decreased.
β-casein is observed to go through a peak in dilational modulus (a lower peak value
than α-elastin) at a surface pressure between 10 and 12 mN/m, it was also seen to
have a negligible shear modulus. The the β-lactoglobulin showed a linear increase
with surface pressure, it was also observed to have a finite shear modulus. It was
concluded that the β-lactoglobulin was in a sterically jammed state while the βcasein remained fluid. Thus α-elastin behaves more like the β-casein, this is not
surprising as casein is closer in structure and molecular weight.
Dynamic and static dilational moduli are presented for PVA and PEO by Sauer et
al. [170]. On PVA the maximum in dilational modulus occurred at the same surface
pressure in dynamic and static measurements, however the dynamic measurements
gave a value that was 10 mN/m higher. PEO showed consistent values for both
the surface pressure and the height of the peak in dilational modulus under the two
measurement methods. This is consistent with what is seen on α-elastin on water.
5.5.2
Temperature Effects
In order to examine the role of hydrophobic interactions in the properties of the
monolayer the compression modulus was investigated as a function of temperature,
over the range known to produce coacervation in the bulk. Figure 5.37 shows four
sets of data which illustrate how the dynamic dilational modulus of an α elastin
film varies with applied surface pressure and temperature. The pairs of points were
150
Figure 5.37: Data from dynamic measurements showing how the dilational modulus changes
with pressure and temperature. The pairs of points (one black and one grey) were measured
simultaneously from the parallel and perpendicular sensors.
measured simultaneously with parallel and perpendicular sensors.
Figure 5.37 shows how the dilational modulus of α-elastin is affected by temperature. The shear modulus (data not shown) did not change significantly with
temperature indicating that the monolayer remains fluid at higher temperatures.
Figure 5.38 shows the value of the peak dilational modulus at the four different
temperatures and the value of the shear modulus at the surface pressure at which
the dilational modulus peaked. The dilational modulus is shown by the diamonds
and the shear modulus by the triangles.
151
From figure 5.38 it is clear that the
peak value of the dilational modulus
does not change significantly with temperature which shows the compressibility of the monolayer is not affected
by temperature. However, figure 5.37
shows that the form of the dilational
modulus peak varies with temperature.
The peak narrows as the temperature
Figure 5.38: Maximum dilational modulus (di-
increases suggesting that the phase tran-
amonds) and the shear modulus at the same pressure (triangles) at the four different temperatures.
sition occurs in a smaller surface pressure interval at higher temperatures.
The phase transition occurs at slightly lower pressure at higher temperatures, the
reduction is 2 mN/m over the 17◦ C rise in temperature. It is suggested that hydrophobic interactions between the elastin molecules are inducing these changes.
Data from a poly(vinylacetate) monolayer [118] at 6◦ C, 20◦ C and 36◦ C also clearly
shows these two effects. As seen in figure 5.38 the shear modulus did not change
significantly with temperature, and remained small compared to the dialational modulus showing the monolayer remained fluid. There were no clear features observed
in the shear modulus which indicates the fluid nature of an α-elastin monolayer over
these temperatures.
It is noteworthy that data was obtainable at 33◦ C and 40◦ C as in bulk solutions αelastin coacervates (as discussed in section 2.2.3) at around 30◦ C. The coacervation
of α-elastin from a monolayer is therefore not a clear phenomenon.
As described in section 4.5.1 the shear elastic modulus of a monolayer is affected
by the frequency of the oscillations applied to the surface. Data presented by Lucassen and Van den Tempel [133] showed this using a monolayer of decanoic acid
152
and varying the frequency of the oscillating barrier between 0.01 and 1 Hz that the
dilational modulus of the monolayer increases with oscillation frequency. As barrier
motion in these α-elastin experiments was at a constant frequency, it would be interesting to investigate the frequency dependence and compare it with the oscillating
ring results discussed in chapter 4.
153
Chapter 6
Elastin-Lipid Interactions
6.1
Introduction
This chapter details experiments that examine the interactions between α-elastin
and phospolipids. These are both biologically important and poorly understood.
As already discussed in section 2.2.5, the binding between lipids and elastin in
vivo is believed to be important to the early stages of atherosclerosis. As previously
discussed, the interaction between lipids and elastin may be important in discussions
about the elastin receptor, secretion of tropoelastin and the early stages of fibril
formation.
In this chapter I describe the methodologies used to examine lipid - elastin interactions. Surface pressure - area measurements utilised the method set out in section
5.2 in the previous chapter. The methodology used in the fluorescence microscopy
experiments is detailed in section 6.2. This section includes details of the specific
materials used in section 6.2.1, then goes on to describe the trough used in these
experiments in section 6.2.2 and the microscopy set-up in section 6.2.3. The results of these experiments are then detailed in section 6.3. Firstly, data involving
the addition of α-elastin to a subphase with no lipid present and the adsorption of
the elastin to the surface is given in section 6.3.1. Then, the surface pressure area
154
isotherms showing lipid-elastin interactions are presented and discussed in section
6.3.2. An examination of the surface pressure - area characteristics of the flurorescently labelled lipids with and without α-elastin in the subphase are discussed in
section 6.3.3. Lastly the fluorescence microscopy results are presented in section
6.3.4.
6.2
Microscropy Methodology
In the experiments detailed in this chapter a lipid monolayer was created on the
surface of a nanopure water subphase, then a small amount of elastin solution was
injected into the subphase and allowed to penetrate into the lipid monolayer.
In order that the monolayer could be imaged the lipid applied to the surface was a
mixture of fluorescently labelled and normal lipid of the same type. The fluorescent
label used was Nitro-benzoxadiazole (NBD). As this fluorescent label is relatively
large and attached to one of the lipid’s tails and the lipids are saturated ) the labelled
lipids are not able to pack as closely as unlabelled lipids. This means that denser
domains in the monolayer have less fluorescence as the labelled lipids are excluded.
Therefore the liquid condensed and liquid expanded phases (as discussed in section
2.3.2) can be visually differentiated throughout their coexistence regime. Optical
fluorescence microscopy provides a convenient method of making direct observations
of the phase state of a lipid monolayer.
6.2.1
Materials
Lipids used in this work, both the normal lipids and their fluorescently labelled counterparts, were purchased from Avanti Polar Lipids (Alabaster, AL, USA). The lipids
used were 1,2-Dipalmitoyl-sn-Glycero-3-Phosphocholine (DPPC), 1-Palmitoyl-2-[12-
155
[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]dodecanoyl]-sn-Glycero-3-Phosphocholine (referred to as NBD-PC), 1,2-Dipalmitoyl-sn-Glycero-3-[Phospho-L-serine] (DPPS) and
1-Palmitoyl-2-[12-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]dodecanoyl]-sn-Glycero-3Phosphoserine (referred to as NBD-PS). All lipids were supplied dissolved in chloroform and were further diluted as needed with the same solvent. Where mixtures
of non-fluorescent and fluorescent lipid were used, they were mixed in a 50:1 molar ratio. All lipid were stored in glass containers at -20◦ C, mixtures containing
fluorescently labelled lipids were stored wrapped in foil to prevent bleaching.
In the body phosphatidylcholine (PC) and phosphatidylserine (PS) make up the
majority of the phospolipid content of cell membranes [172]. Thus, examining the
interaction between them and α-elastin is a useful model for the transfer of tropoelastin through the cell membrane prior to the assembly of elastin fibres in the extracellular space. PC and PS have also been extensively studied and their surface
pressure - area characteristics are well known.
6.2.2
Microscope Trough
The trough used for the microscope work was a commercial trough from Kibron
Inc. Finland. It has a resin base and polytetrafluroethyleme (PTFE) rims, see
figure 6.1. The trough’s dimensions are 59 mm wide by 230 mm long by 0.5 mm
deep. The reason that this trough is so much shallower than the trough used for the
elastin monolayer studies is so that Π-A measurements and fluorescence microscopy
can be carried out simultaneously. The 20 mm diameter quartz window in the
base of the trough allows the light from the microscope (described below) to reach
the monolayer being examined. The barriers used were the same as used with the
quasi-static trough.
156
Figure 6.1: Schematic diagram of the trough used for the fluorescence imaging experiments.
6.2.3
Fluorescence Microscopy
The microscope with which the fluorescence imaging was performed was an inverted
Lecia DM IL Microscope. Images were taken using either the 20x or the 40x magnification objective in combination with a x10 magnification eyepiece; thus the overall
magnification was either x200 or x400. Images were observed via a Watec WAT902B (CCD) digital camera with output directly to a display. Videos were recorded
via a Lite-On LVW-5045 HDD + DVD recorder. Still images were then extracted
from the videos using Microsoft Windows MediaPlayer and Adobe Photoshop.
6.3
6.3.1
Elastin Adsorption Results
α-elastin in Subphase
Figures 6.2 and 6.3 show data which was taken on the quasi-static trough described
in section 5.2 and figure 5.1 and shows the effect of adding elastin to the subphase
when no lipid monolayer was present. In this experiment a clean water surface was
prepared and the surface pressure set to zero, then 2 µL of 100 mg/mL α-elastin
was injected into the subphase which was stirred (something that was not possible
157
Figure 6.2:
Surface pressure - time depen-
Figure 6.3:
Surface pressure - area depen-
dence of adsorbing α-elastin, showing the for-
dence of adsorbing α-elastin, during formation
mation and two compression-relaxation cycles
and two compression-relaxation cycles of a sur-
of a surface layer.
face layer.
2 µL of 100 mg/mL α-elastin was injected into the water subphase while the monolayer was at
maximum area. The subphase was stirred. The compression rate was 22 mm2 /s.
158
in the microscope trough). A total of 0.2 mg α-elastin was injected. Assuming a 5
mm high meniscus (this gives an upper limit for the volume, therefore lower limit for
concentration) gives a bulk concentration of 0.0025 mg/mL. The injection of elastin
was done with the trough at maximum area. The trough was left at maximum
area throughout region A (see figure 6.2). The area change in the course of each
compression or relaxation was 3397.5 mm2 and the compression rate was 22 mm2 /s.
In figures 6.2 and 6.3, the rise in surface pressure, which was observed before
the barriers were moved, shows the rate at which elastin is able to migrate to the
surface; the timescale of the pressure rise being dictated by the rate of transport
of elastin to the surface. In the case of figure 6.2, a surface pressure increase of 25
mN/m was observed in 3000 s, with the first 17 mN/m coming in the first minute
after the addition of the elastin. However, as the subphase was stirred throughout
this experiment this pressure rise was not spontaneous. The adsorption of a protein
from the bulk to an air-water interface is not unique to α-elastin. For example,
Kawaguchi et al. [168] noted that poly(ethylene oxide) (PEO), an amphiphilic
water soluble polymer, in dilute solution is able to spontaneously adsorb at the
solution/air interface, as well as form stable monolayers.
For the adsorped monolayer, the maximum surface pressure was greater than
that observed on an applied monolayer. This is largely due to the initial pressure
increase as over a full compression increase in Π was 18 mN/m, comparable with
the observed on the applied monolayers as shown in figures 5.2 and 5.3. As the
monolayer was formed by adsorption from the bulk the distribution of the α-elastin
between the subphase and the surface is not clear, and the elastin is clearly able to
transition between the two phases. However, as the second compression - relaxation
cycle is shown to have a very similar Π-A isotherm to the first cycle (figure 6.3) it
is assumed that the amount of α-elastin on the surface at the start of the second
cycle is the same as at the start of the first cycle.
159
Figures 6.4, 6.5 and 6.6 show the α-elastin adsorption behaviour for an unstirred
subphase. 40 µL of 1 mg/mL α-elastin were injected into the subphase at Π = 0
and t = 200 s. As can be seen from figure 6.4 there was no observable rise in surface
pressure over the first 6000 s after the addition of the elastin. The addition to the
subphase was made with the barriers in a semi-compressed position. Point A of
figure 6.4 indicates when the barriers were expanded back to maximum area, the
compression rate was 27 mm2 /s, it was when the barriers started moving that the
surface pressure started to increase. When the barriers stopped the surface pressure
was 12.4 mN/m and continued to increase while the barriers were stationary. Region
B was at constant area. Then, at 8000 s, the barriers were moved together for
the first time; the first of two compression - relaxation cycles is shown. Figure
6.5 shows the surface pressure isotherm and figure 6.6 shows the dilational moduli
calculated from this compression - relaxation cycle. The modulus calculated from
the compression is shown in red and the modulus from the relaxation is shown
in blue. In figures 6.4, 6.5 and 6.6 if the same assumption as above is used, it is
calculated that after the addition of elastin the bulk concentration is 0.00054 mg/mL
.
Figure 6.5, like figure 6.3, indicates that there is a significant difference in the
Π-A characteristics of an adsorped monolayer compared to an applied monolayer.
The compression and relaxation values of the dilational modulus, shown in figure
6.6, are similar in most of their overlap range (17 - 20 mN/m). At approximately 22
mN/m it is observed that while during compression the modulus is going through a
minimum, the relaxation dilational modulus has a minimum and a maximum very
close together. This is attributed to the non-equilibrium state of the monolayer
during compression.
The minimum in dilational modulus is at similar surface pressure to the minimum
observed from an applied α-elastin film under dynamic compression (shown in figure
160
Figure 6.4:
course.
The Π - time
Figure 6.5:
The Π-A
Figure 6.6:
The dilational
moduli.
isotherm.
The first compression-relaxation cycle from an adsorped α-elastin monolayer. 40 µL of 1 mg/mL
α-elastin was injected into an unstirred water subphase at Π = 0 and t = 200 s. The compression
rate was 27 mm2 /s. In figure 6.6 the modulus from compression is shown in red and the modulus
calculated from the relaxation is shown in blue.
5.35). However, despite this similarity, it is unlikely that the adsorped and applied
monolayers are similar in structure due to them not being in equilibrium.
In systems where α-elastin was injected into an unstirred water subphase there
was very little variation in pressure over a time period of up to 8000 seconds. However, in these systems, when there was any barrier movement a surface pressure
increase of 15 ± 2 mN/m was observed. In these cases, there was an initial rise in
Π and then the surface pressure is seen to be constant as the surface area increases
indicating that the rate of transfer of α-elastin to the surface is balanced by the
expansion of the surface. When the expansion of the surface finishes the surface
pressure increases again: it was seen that the surface pressure continued to rise for
at least 1400 seconds. It was seen that movement the subphase, either from stirring
or from barrier movement, was needed to start the adsorption. It is presumed that
it is a lower energy state for the elastin to be at the water’s surface with its hydrophobic regions outside the water, this in agreement with observations by Davis
161
et al. [173] discussed below.
Once the surface pressure has started to increase it must be assumed that the
whole subphase surface, including that outside the barriers, is covered with an even
film of elastin. Comparing figure 6.3 with the applied elastin monolayer Π-A shown
in figure 5.3 suggests that the α-elastin surface concentration at equilibrium before barrier movement must be greater than 0.007 µg / mm2 . It is interesting to
note the adsorbed α-elastin films do not show signs of collapse even at surface pressures greater than 40 mN/m. This comparison also shows that the rate of pressure
decrease at maximum area was much faster for the adsorbed monolayer than the
applied monolayer. It is suggested that part of this rapid decline in pressure is due
to the non-equilibrium nature of the compression; likewise, the large rise in pressure
observed when the relaxation has finished.
The energy required to transfer the injected elastin to the water surface can not
come purely from interactions between the hydrophobic regions of the α-elastin and
the surrounding water; as spontaneous adsorption was not observed. This makes
sense because in order to be soluble (as described in section 2.2.3) α-elastin cannot
have many exposed hydrophobic regions. As noted above, once a surface pressure
increase has started it continues even after the triggering motion has ceased. In the
experiment shown in figures 6.4 - 6.6 a bulk α-elastin concentration of 0.0004 mg/mL
was used, 6.25 times less than used in the stirred subphase experiments shown in
figures 6.2 and 6.3 and used in the PC monolayer penetration experiments. This is
most likely why a lower surface pressure was reached in this experiment than in the
experiment shown in figures 6.2 and 6.3.
Carrera Sánchez et al [174] observed that β-casein injected into an aqueous subphase with no monolayer present produced no observable pressure rise after 24 hours
(similar observations were made by the same group with β-lactoglobulin in bulk solution at similar concentrations [108]). The bulk concentration of the β-casein [174]
162
was two orders of magnitude smaller than that was used in figures 6.2 and 6.3 and
an order of magnitude smaller than that used in figures 6.4 - 6.6. The concentration
of the β-lactoglobulin [108] was even lower.
However, later work by Lucero Caro et al. [142] (the same group responsible
for the β-casein work discussed above [174]) again using β-casein at even lower
concentrations (approximately 1000 times less concentrated that the subphase in
figures 6.2 and 6.3) showed a surface pressure rise of 15 mN/m in around 3 hours.
In this experiment energy was supplied to the system due to vertical movements
of a ring rheometer in the interface. Thus, it seems that the protein’s adsorption
characteristics change drastically with the amount of energy available in the system.
This suggests that the adsorption of β-casein at surface is an effect of stirring. Data
shown above for α-elastin indicates that its adsorption is also an affected by stirring
of the subphase.
Surface pressure readings for a range of solution concentrations have been observed for β-casein [173]. It was noted that the equilibrium surface pressure increases
with concentration to a point above which increasing the bulk concentration of the
protein had no affect on the observed surface pressure. This experiment was also
carried out with apo-HDL2 (a delipidated high density lipoprotein) and Fraction
III (the lipoprotein’s major component peptide); all three had a maximum solution surface pressure of 23 mN/m. Davis et al [173] attribute this behaviour to the
hydrophobicity of the proteins and note that all three are surface active and have
tertiary structures that are free from severe conformational constraints. Thus it is
implied that α-elastin has no major conformational constraints. The work of Davis
et al[173] differs from this work as they note that even in unstirred solution the
equilibrium value of surface pressure is obtained quickly. This does not seem to be
the case for α-elastin solutions of the concentrations used in this work. The concentrations of α-elastin used in this work are below the saturation value for β-casein;
163
at the lowest concentration used in this work, a β-casein solution showed a pressure
rise of 1 mN/m [173].
It would seem that it is energetically favourable for some of the α-elastin to
reach the surface and orientate so that the hydrophobic regions are out of the water
subphase which agrees with observations in [173]. This would naturally cause the
formation of a monolayer. The review paper by Dickinson [175] covers much of the
work on protein adsorption to date but focusses largely on globular proteins.
6.3.2
Lipid-Elastin Π-A Measurements
Figure 6.7 shows a typical Π-t isotherm
from a monolayer made up of 3 µL of 1
mg/mL (16:0) PC, which then had 2 µL of
100 mg/mL α-elastin injected into the subphase. This experiment was carried out in
the general Langmuir trough (described in
section 5.2 and shown in figure 5.1), thus
it was possible to stir the subphase during this experiment. The compression rate
Figure 6.7:
A surface pressure - time
isotherm showing the insertion of elastin from
was 21 mm2 /s. The lipid was applied to
the subphase into a lipid monolayer on the sur-
the subphase surface at point A (see fig-
face. 3 µL of 1 mg/mL (16:0) PC made up the
ure 6.7). At point B, after 3 compression-
monolayer and 2 µL of 100 mg/mL α-elastin
expansion cycles, the lipid monolayer was
was injected into the subphase at point B. The
at minimum area and the elastin was in-
compression rate was 21 mm2 /s.
jected. This injection was performed outside the barriers using a fine tipped microlitre syringe to avoid any disturbance to
the monolayer. The monolayer was left compressed throughout region C. Region D
164
denotes the expansion and relaxation after the addition of elastin, this was followed
by at least two compression-expansion cycles.
For the elastin insertion experiments, as shown in figure 6.7 the lipid was first
applied to the subphase surface and the monolayer characterised one or more compression - relaxation cycles. Then the elastin was added to the subphase and allowed
to equilibrate. Then the combined monolayer was characterised by two or more compression - relaxation cycles. Experiments were repeated at least three times on each
subphase.
As figure 6.7 shows, when the subphase was stirred the α-elastin was able to
insert into the PC monolayer (the same process was observed with a PS monolayer)
while that monolayer was at high surface pressures. This suggests that α-elastin
can insert into monolayers at surface pressures above 30 mN/m thus increasing the
surface pressure by approximately 5 mN/m.
It is also notable in figure 6.7 that the addition of α-elastin to the lipid monolayer prevents the surface pressure dropping to zero upon expansion. Comparing
this surface pressure increase to those in figure 5.3 suggests that at maximum area
the concentration of elastin on the surface could be greater than that obtained at
minimum area on the plain elastin applied monolayer. It is more likely that the
increase in surface pressure is not entirely due to the physical presence of the αelastin on the surface; it seems likely that disruption to the lipid’s arrangement and
interactions between the the PC and the elastin are also contributing to the increase
in surface pressure.
Mixed PC-elastin monolayers
Figure 6.8 shows the first compression-relaxation cycle of a lipid monolayer, while
figure 6.9 shows the first compression-relaxation cycle after the addition of elastin
165
to the subphase below the lipid monolayer. The elastin was added at minimum area
4500 seconds after the application of the lipid monolayer. These graphs are from the
same experiment whose pressure - time course is shown in figure 6.7. The α-elastin
was added while the PC monolayer was compressed. The monolayer was left in
this state for approximately 1 hour and then expanded. The compression relaxation
cycle recorded in figure 6.9 was taken 5000 s after the addition of the elastin. This
experiment with a PC monolayer involved the injection of a total of 0.2 mg αelastin. Making the same assumptions as above this gives a bulk concentration of
0.0025 mg/mL, the same as shown in figures 6.2 and 6.3.
Figure 6.8: The Π-A isotherm of PC mono-
Figure 6.9:
layer consisting of 3 µL of 1 mg/mL 16:0 PC.
monolayer whose Π-A characteristic is shown
The Π-A isotherm of the PC
in figure 6.8 after the addition of 2 µL of 100
mg/mL α-elastin.
In both these experiments the subphase was stirred, and the compression rate was 21 mm2 /s.
Figure 6.8 is a classic lipid Π-A isotherm and shows very little alteration in the
monolayer between compression and expansion. This is also shown by the dilational
modulus calculated from it as shown in figure 6.10 below. Figure 6.9 shows that
there have been significant changes to the monolayer with the addition of elastin.
166
The surface pressure at a given area has increased, this is expected due to the
increase in the amount of material on the surface. The hysteresis in figure 6.9 must
have been induced by the addition of elastin to the monolayer, indicating that the
mixed monolayer behaves more like an α-elastin monolayer than a PC monolayer
since there are reorientations going on in the monolayer between compression and
expansion which account for the differences between the compression and relaxation
curves. The drop in pressure when the monolayer is at maximum compression is
larger than that seen on the applied monolayers and indicates that this monolayer
is not in equilibrium, which was expected due to the compression speed.
Figure 6.10 shows the dilational moduli calculated from the data shown in figure
6.8. Figure 6.11, calculated from figure 6.9, shows results for the same monolayer
(red is from the compression and blue is from the relaxation) after the addition of 2
µL of 100 mg/mL α-elastin to the subphase. The compression rate was 21 mm2 /s.
The dilational moduli were calculated in the same way as for the quasi static data
discussed in section 5.4, however, it is not believed that the moduli calculated here
equilibrium due to the fast rate of barrier movement.
As noted above figure 6.10 indicates that, as expected, there are no changes in
a PC monolayer between compression and relaxation. Figure 6.11 shows that the
dilational moduli of the PC monolayer with elastin exhibits a very clear difference
between compression and relaxation. It is observed that the monolayer under compression is more compressible than the monolayer under relaxation. This perhaps
indicates that the relaxed state is less compressible that the non-relaxed state. The
peak in dilational modulus at 25 mN/m suggests that there is a phase transition
here. The surface pressure seen is higher than the transition as observed in the pure
α-elastin results, and no transitions are observed in the case of pure lipid monolayers. It is therefore indicated that the mixed monolayer has phase transition at
a surface pressure between its constituents; the maximum value of the dilational
167
Figure 6.10:
The dilational modulus of a
Figure 6.11: The dilational modulus of a PC
PC monolayer whose Π-A isotherm is shown in
monolayer with α-elastin, the Π-A isotherm of
figure 6.8.
this monolayer is shown in figure 6.9.
In both figures 6.10 and 6.11 the dilational modulus from a compression is shown in red and that
from the relaxation in blue. The compression rate was 21 mm2 /s.
168
modulus indicates that the mixed monolayer is less compressible than the PC or
α-elastin monolayers.
Mixed PS-elastin monolayers
The data shown in figures 6.7 - 6.11 above, was taken using a PC film. Comparable
data was obtained using PS monolyer. In these experiments 4 µL of 1 mg/mL PS
was applied to form the monolayer and then 2 µL of 100 mg/mL α-elastin was
injected into the subphase while the monolayer was compressed. Figure 6.12 shows
the surface pressure-area characteristic of the plain PS monolayer, while figure 6.13
shows the Π-A characteristic of the same monolayer after the addition of elastin.
After the elastin was injected the monolayer was left compressed for 4500 s, then
the monolayer was expanded. Figure 6.13 shows the first full compression-expansion
cycle after the addition of elastin. The cycle started 6500 s after the elastin was
added. The compression rate was 21 mm2 /s. The subphase was stirred and the αelastin concentration in the subphase was calculated to be 0.0025 mg/mL. Figures
6.14 and 6.15 show the dilational moduli calculated from the surface pressure - area
isotherms shown in figures 6.12 and 6.13 respectively. The modulus calculated from
the compression is shown in red and the modulus from the relaxation is shown in
blue.
As with the PC results, figure 6.12 shows a classic lipid Π-A isotherm which
has little hysteresis and so there is very little change in the monolayer between
compression and expansion. The drop in surface pressure observed at maximum
compression is a result of the speed of the compression meaning the monolayer
takes a while to equilibrate.
Figure 6.13 shows a Π-A isotherm from a PS monolayer after the addition of
α-elastin; as with the PC monolayer the surface pressure of the elastin penetrated
PS monolayer at a given surface area is higher than that of the plain PS monolayer
169
Figure 6.12:
The Π-A isotherm of a PS
Figure 6.13:
The Π-A isotherm of the PS
monolayer consisting of 4 µL of 1 mg/mL 16:0
monolayer whose Π-A isotherm is shown in
PS.
figure 6.12 after the addition of 2 µL of 100
mg/mL α-elastin
In both these experiments the subphase was stirred, and the compression rate was 21 mm2 /s.
170
which is due to the protein molecules added to the surface. The shape of the
isotherm has also changed, the initial increase in surface pressure is close to linear.
This indicates that the monolayer has been modified by the addition of elastin in a
different way from the PC monolayer as the PC monolayer’s surface pressure - area
isotherm did not change shape in this way. The hysteresis in figure 6.13 indicates
that the combined monolayer is behaving more like an α-elastin monolayer than a
PS monolayer.
Figure 6.14: The dilational modulus of a PS
Figure 6.15: The dilational modulus of a PS
monolayer whose Π-A characteristic is shown in
monolayer with α-elastin, the Π-A characteris-
figure 6.12.
tic of this monolayer is shown in figure 6.13.
In both figures 6.14 and 6.15 the dilational modulus from a compression is shown in red and that
from the relaxation in blue. The compression rate was 21 mm2 /s.
In the case of the PS monolayer, the compression and relaxation dilational moduli
were only comparable between 0 - 6 mN/m. In the case of the compression the
dilational modulus continued to increase with surface pressure; showing that the
monolayer is becoming less and less compressible. In the case of the relaxation, the
monolayer starts out in a relatively relaxed state at around 13 mN/m, the monolayer
then becomes less and less compressible, reaching a minimum in compressibility at
171
5 mN/m. The monolayer then becomes more compressible. The behaviour of the
compression can be explained by the increased surface density of material as the
monolayer is compressed, it is likely that the fast compression rate does not allow
time for phase transitions to occur during the barrier movement.
The changes made by the α-elastin in the PS monolayer are very different from
those made to the PC monolayer. The monolayer is more compressible with the
addition of the elastin, and the compression and relaxation are similar to each other.
The elastin penetrated PS monolayer appears more stable with pressure, possibility
this monolayer equilibrates faster; there are peaks and troughs located differently
for compression and relaxation showing that the monolayer cannot be completely
relaxed.
Discussion of Results
As figures 6.9 and 6.13 show similar behaviour in terms of initial pressure rise on
addition of α-elastin (15 - 20 mN/m at the start of the compression) and insertion
pressure, but there are large quantitative differences between the two. As indicated
by figure 2.8 in chapter 2 the choline and serine head groups are differently charged
(serine is overall negative and choline is overall positive) and thus, the differences
may be a result of charge interactions. However, as both the serine and choline
have positive and negative charges and the overall charge will vary with pH thus the
interaction is likely complicated.
Figure 6.16 compares the Π-A characteristic of a PC monolayer before (red) and
after (blue) the addition of α-elastin to the subphase. This data was previously
presented as figures 6.8 and 6.9. Similarly, figure 6.17 compares the Π-A characteristic of PS monolayer before (red) and after (blue) the addition of α-elastin to the
subphase. This data was previously presented as figures 6.12 and 6.13.
172
Figures 6.16 and 6.17 clearly show changes in the behaviour of a monolayer during
a compression - relaxation cycle. It has already been noted for both lipids that higher
surface pressures were reached after the addition of elastin. As can be seen from
figures 6.16 and 6.17 below the surface pressure - area isotherms of the monolayer
are drastically changed by the addition of elastin to the subphase. The precise shape
of these changes will be examined by dilational modulus analysis and by hysteresis
calculations. In both figures 6.16 and 6.17 the surface pressure of the monolayer
after the adsorption of α-elastin is higher at a given area, for all areas, than that
of the monolayer prior to the adsorption at all surface area. This suggests that the
elastin is not forced out of the monolayer at high pressure.
Figure 6.16: A comparison of Π-A isotherms
before (red) and after (blue) the addition of α-
elastin under a PC monolayer. This was previ-
elastin under a PS monolayer. This was previ-
ously presented as figures 6.8 and 6.9.
ously presented as figures 6.12 and 6.13.
As discussed in section 2.2.5 in vivo there are many interactions between lipid
and elastin. Tropoelastin is able to pass through the lipid membrane of a cell so it
is not surprising that α-elastin can insert into a lipid monolayer.
173
6.3.3
Π-A Characteristics of Fluorescently Labelled Lipids
When fluorescently labelled lipids were used the experimental procedure was as
described above except that video images were recorded from the monolayer. Figures
6.18 and 6.19 show surface pressure - area isotherms recorded from labelled lipid
monolayers. All lipids were in 16:0 form and mixed 50:1 unlabelled to labelled.
Figure 6.18 shows data from a monolayer made up of 3 µL 1 mg/mL PC:NBD-PC.
The red curve shows the Π-A isotherm of the lipid monolayer and the blue curve
shows the characteristic after the addition of 30 µL of 1 mg/mL α-elastin. The
‘before’ curve is the first compression-relaxation cycle on the monolayer; the ‘after’
data is from the third compression-relaxation cycle after the addition of the elastin,
the elastin had been present in the subphase for 11000 seconds at the start of this
cycle. Figure 6.19 shows data from a monolayer made up of 4 µL of 1 mg/mL 1:1
PC:PS. Again, the red curve indicates the Π-A characteristic of just the lipid and the
blue curve indicates the Π-A characteristic of the same monolayer after the addition
of 30 µL of 1 mg/mL α-elastin. The ‘before’ curve was the second compressionrelaxation cycle; and the ‘after’ data was from the third compression-relaxation
cycle after the injection of elastin, the elastin had been on the surface for 11000
seconds at the start of this cycle. In both of these experiments the compression
rate was 29 mm2 /s. This data was taken on the microscope trough and the bulk
α-elastin concentration after addition was 0.0004 mg/mL.
Figure 6.18 can be directly compared to figure 6.16. The PC:NBD-PC mixture
shows a phase transition at ' 5 mN/m which is not seen in the unlabelled lipid
and must therefore be attributed the the effect of the fluorescent label. As with the
unlabelled lipid monolayers the initial pressure rise induced by the addition of the
α-elastin is, again, 15 - 20 mN/m with either the PC:NBD-PC monolayer or the
PC:NBD-PC:PS:NBD-PS monolayer. This indicates that the addition of the labels
has not affected the ability of elastin to penetrate the monolayer.
174
from an PC:NBD-PC monolayer before (red)
from an PC:NBD-PC:PS:NBD-PS monolayer
and after (blue) the addition of α-elastin to the
subphase. The initial monolayer consisted of 3
elastin to the subphase. The initial monolayer
µL 1 mg/mL 50:1 PC:NBD-PC.
consisted of 4 µL of 1 mg/mL 1:1 PC:PS. Both
lipids were 50:1 unlabelled to labelled mixes.
In both figures 6.18 and 6.19 30 µL of 1 mg/mL α-elastin was added to the subphase. The
compression rate for both experiments was 29 mm2 /s.
175
6.3.4
Fluorescence Microscopy
Figure 6.20 shows an image of a monolayer of PC before the addition of α-elastin.
This image was taken using the x20 objective at a surface pressure of 26 mN/m.
The domain structure shows the coexistence of a liquid expanded (lighter regions)
and a liquid condensed phase (darker regions). 3 µL of 1 mg/mL 50:1 NBD-PC were
applied to the surface to form the monolayer. It can be seen that the condensed
phase domains are approximately 5 µm in diameter.
Figure 6.21 shows the same PC monolayer as in figure 6.20 but after the addition
of 30 µL of 1 mg/mL α-elastin to the unstirred subphase. Once the elastin had been
added to the subphase the PC monolayer was left at minimum area for an hour and
half. The monolayer was then relaxed. The monolayer was then compressed and
relaxed and then compressed to a pressure of 26 mN/m at which time the image
shown in figure 6.21 was taken. Making the same assumptions as above, the bulk
α-elastin concentration was 0.0004 mg/mL.
Figures 6.22 and 6.23 show the same 1:1 PC:PS (50:1 unlabelled:labelled) monolayer made up of 4 µL of 1 mg/mL lipid before (6.22) and after (6.23) the addition
of 30 µL of 1 mg/mL α elastin to the subphase. The mixed lipid monolayer does
not have the clear domain structure that PC alone has. Figure 6.22 was taken at a
pressure of 17 mN/m; while figure 6.23 was taken at a pressure of 20 mN/m. The
bulk α-elastin concentration in figure 6.23 was 0.0004 mg/mL.
Fluorescence images were not obtained from a PS monolayer; thus it was impossible to observe what affect the addition of α-elastin has on any domain structures
in the PS monolayer.
Figure 6.20 clearly shows the domain structure of PC in the region of phase
coexistence. Adding α-elastin to the subphase disrupts the domain structure as can
be see in figure 6.21. The domain structure present before the addition of elastin
176
Figure 6.20: A fluorescence microscopy im-
age of a PC (50:1 normal:NBD) monolayer. 3
age of the same PC monolayer as figure 6.20
µL of 1 mg/mL were applied to form the mono-
on the second compression-relaxation cycle af-
layer. Darker regions are the liquid condensed
ter the addition of 30 µL of 1 mg/mg α-elastin
phase while the lighter regions are the liquid
to the subphase which was not stirred.
expanded phase.
Both figures 6.20 and 6.21 were taken at x20 magnification at a surface pressure of 26 mN/m. The
white bar in each image is 50 µm in length.
177
A fluorescence microscopy
image of a (1:1) PC:PS monolayer 50:1 nor-
age of the same PC:PS monolayer as figure 6.22
mal:NBD). 4 µL of 1 mg/mL lipid was applied
on the first compression after the addition of 30
to the subphase surface to form the monolayer.
µL of 1 mg/mL α-elastin to the subphase which
Figure 6.22:
was not stirred.
Both figures 6.22 and 6.23 were taken at x40 magnification and surface pressure of 20 mN/m. The
white bar in each image is 50 µm in length.
shows a quasi-regularity; this has been destroyed by the addition of elastin to the
monolayer as the image in figure 6.21 shows.
The effect of adding α-elastin under a PC:PS monolayer is less clear. Figure 6.22
shows the monolayer before the addition of elastin. Figure 6.23 shows the same
monolayer after the addition of elastin to the subphase. Overall it appears that
there are less fluorescent regions in the after image which would correspond with
there being more material on the water’s surface and more of the monolayer being
in an LC phase which excludes the fluorescently labelled lipids.
178
Chapter 7
Conclusions and Future Work
7.1
Introduction
In this chapter I shall summarise and conclude my discussion and detail some of
the directions future work in this area could take. Section 7.2 contains a summary
of the results and discussion detailed in each of the results chapters. Section 7.2.1
summarises the results of the the bulk and surface rheometry detailed in chapter 4.
The results of the experiments on applied α-elastin monolayers, detailed in chapter
5, are summarised in section 7.2.2. Section 7.2.3 then summarises the results of the
experiments on lipid-elastin monolayers that were detailed in chapter 6.
Section 7.3 summarises the conclusions for each set of experiments. Section 7.3.1
gives the conclusions from the rheometry experiments, section 7.3.2 gives the conclusions from the α-elastin monolayer experiments, and section 7.3.3 gives the conclusions from the lipid-elastin interaction experiments. Finally, section 7.4 gives
suggestions for possible future experiments in each of these three areas. Section
7.4.1 details possible future rheology experiments, section 7.4.2 gives possible future
elastin monolayer experiments, and experiments to further investigate α-elastin’s
interactions with lipids are given in section 7.4.3.
179
7.2
7.2.1
Discussion Summary
Rheometry Results and Discussion
In chapter 4 bulk viscometry measurements using a capillary viscometer was used
to characterise the behaviour of α-elastin in water, 0.1 M NaCl and 0.1 M CaCl2 .
This led to calculations of the dynamic, kinematic and relative viscosities for each
subphase.
0
00
Surface rheometry was used to observe that G , G and G∗ were seen to increase
0
with surface pressure and all except G on water were seen to increase with oscillation
frequency. These measurements enabled the surface shear viscosity to be calculated.
It was seen that ηs increased with applied surface pressure which was expected as at
higher surface pressures the elastin molecules are forced closer together and so the
00
interactions between them increase. Examination of G showed a phase transition at
around 8 mN/m, which was attributed to a semi-dilute to concentrated transition.
7.2.2
Elastin Monlayers Results and Discussion
It was seen that α-elastin formed stable monolayers up to 22 mN/m on water under
quasi-static compression. Spreading of elastin on water was shown to be consistent
producing an average surface pressure rise of 0.8 mN/m on application. Hysteresis
was observed from α-elastin monolayers which indicates rearrangement taking place
within the monolayer.
From dilational modulus calculations a phase transition was observed between
5 and 10 mN/m. In contrast it was seen that applying α-elastin to an ionic subphase produced an average increase in surface pressure of 3.3 mN/m. The dilational
moduli from the first expansion of the monolayer showed a phase transition at a
similar surface pressure to the monolayers on water. Altering the pH of a water
180
subphase was shown to make no difference to the dilational modulus observed from
a monolayer, however, it was also seen that α-elastin does not spread consistently
at pH 3.5
Examining the relaxations of a compressed monolayer indicated that the relaxation process was occurring over longer than 5000s. Exponential fitting to the surface
pressure decay allowed calculation of the time constant of the relaxation.
Dynamic compression of α-elastin allowed a slightly higher pressure regime to be
probed and hinted at another phase transition at around 22 mN/m. The shear and
dilational moduli were analysed. The shear modulus was shown to be negligible
compared to the dilational modulus which indicates that the monolayer is fluid
in the pressure range examined. A peak was observed in the dilational modulus
which is consistent with that observed in the quasi-static experiments. Comparisons
between dynamic and quasi-static measurements showed that the dilational moduli
were consistent on water, but on calcium chloride the dynamic measurements were
shown to give a much clearer peak, which may have been due to the history of the
monolayer.
It was seen that the peak value of the dilational and shear moduli did not alter
significantly with temperature. The surface pressure at which the peak in dilational
modulus and so the phase transition occurs was seen to reduce slightly as temperature increases, a reduction of 2 mN/m for a 17◦ C rise in temperature. It is suggested
that temperature dependent hydrophobic interactions between the elastin molecules
are inducing these changes.
7.2.3
Lipid-Elastin Interactions Results and Discussion
It was seen that α-elastin will adsorp from a stirred subphase to produce a monolayer. These monolayers were shown to have significantly different surface pressure
181
- area characteristics to an applied monolayer. The surface pressures obtained from
the adsorped monolayer were much higher than those from the applied monolayers
and included a large abrupt pressure increase caused by the transition of elastin to
the surface. However, on a compression the amount by which the surface pressure
increased was consistent with the quasi-static measurements. It was also observed
that in an unstirred subphase α-elastin does not spontaneously form a monolayer.
In these experiments low bulk concentrations were used, however, the surface pressures reached after the formation of the monolayer suggested relatively high surface
concentrations. In the case where the bulk concentration was 0.00054 mg/mL comparison with quasi-static experiments suggested that the surface concentration after
the initial monolayer formation was at least 0.007 µg / mm2 .
Calculation of the dilational moduli of adsorped monolayers showed that when
adsorped monolayers were left at minimum area, which resulted in a drop in surface pressure, the dilational modulus dropped indicating that the monolayer was
becoming easier to compress.
α-elastin was seen to be able to insert into a lipid monolayer even when the
lipid monolayer was compressed to 26 mN/m. The surface pressure - area isotherm
of the resulting mixed monolayer showed hysteresis similiar to that of the applied
elastin monolayers. Dilational modulus calculations showed that holding a PCelastin monolayer at constant area resulted in a decrease in compressibility and
peak indicating a phase transition, while a PS-elastin monolayer showed an increase
in compressibility.
The use of fluorescently labelled lipids did not inhibit the α-elastin’s ability to
insert into the monolayer and comparable surface surface pressure increases were
observed. Fluorescence microscopy showed the domain structure of a NBD-PC:PC
monolayer in the coexistence regime between the liquid expanded and liquid condensed phases. It was seen that the addition of α-elastin to the monolayer signifi182
cantly disrupts the domains.
7.3
7.3.1
Conclusions
Rheometry Conclusions
It was seen that in bulk solution the Ca2+ ion affects α-elastin differently to Na+
ions. This is consistent with previous work. The intrinsic viscosity of α-elastin in
water was 0.0081 mL/mg at 5◦ C and 0.0073 mL/mg at room temperature. These
values were seen to be comparable to similar proteins. Calculation of the intrinsic
viscosity led to comparisons of the radius of gyration. It was seen that in water and
sodium chloride that RG increased by 2%, however, in calcium chloride RG was seen
to increase by 6% which indicates the affect of the calcium ion binding makes the
α-elastin expand.
It was shown that in the surface pressure range 12-20 mN/m α-elastin monolayers
on water behave viscoelastically. Viscoelastic behaviour was also seen in monolayers
00
on 0.1 M CaCl2 in the surface pressure range 14-18 mN/m. The value of G before
the phase transition allowed the identification of the phase transition as being semidilute to concentrated.
Fitting to Eyring’s model allowed calculation of the area per segment in motion.
On water Am = 48 Å2 /segment, and calcium chloride Am = 76 Å2 /segment. These
values are correspond well with conclusions based on area per molecule calculations
which indicate that at a given surface pressure an α-elastin molecule on calcium
chloride takes up at least 2.4 times are much space as a molecule on water.
183
7.3.2
Elastin Monolayers Conclusions
Flory analysis indicated that below 5 mN/m the monolayer was in the semi-dilute
regime and between good and Θ solvent conditions. This indicated that the observed
phase transition is semi-dilute to concentrated which confirms what was observed
in the surface viscometry results. On water ν = 0.671 ± 0.002. The salt solutions
were seen to provide the α-elastin with conditions closer to good solvent νN aCl =
0.70 ± 0.04. and νCaCl2 = 0.7 ± 0.2.
Calculation of the area per molecule of these monolayers suggests that each
molecule of elastin is occupying a larger surface area than molecules of comparable molecular weight, such as bovine serum albumin and β-casein, which suggests
that inter and intra-molecular forces in the elastin must be higher than the other
proteins to create this surface pressure while being less densely packed. Monolayers
on calcium chloride was not seen to produce results that were significantly different
from those of sodium chloride. This is in contrast to the bulk results described
above.
Exponential fitting to surface pressure decays on water, 0.1 M NaCl and 0.1
M CaCl2 showed that the time constant of the decay was consistent for all three
subphases, τ = 0.001 ± 0.0003 s. This indicates that the relaxation process is not
limited by electrostatic interactions between the monolayer and the subphase. This
data is consistent with the compression-relaxation data.
It was seen that α-elastin forms monolayers up to 40◦ C, which is above the
temperature at which bulk solutions of α-elastin coacervate. This indicates that
coacervation from a monolayer may follow a different mechanism than in the three
dimensional case. It is likely that confining the elastin at an interface reduces the
molecules ability to bond together to coacervate.
184
7.3.3
Lipid-Elastin Interactions Conclusions
In chapter 6.3.2 it was shown that α-elastin does not spontaneously form monolayers
at a water surface when added to the bulk. However, with stirring a monolayer
formed quickly, this resulted in a surface pressure rise of 15 ± 2 mN/m. It was also
shown that in the presence of a lipid monolayer and a stirred subphase α-elastin is
able to insert at surface pressure above 30 mN/m.
Using NBD labelled lipids it was seen that the label did not affect the elastin
insertion into the lipid monolayer. It was seen that the α-elastin insertion into a
PC:NBD-PC monolayer disrupted the domain structure of the monolayer.
7.4
Future Work
All of the experiments detailed in this work could be repeated with similar proteins
tropoelastin and κ-elastin and soluble lamprin, an elastic protein that is farther down
the evolutionary scale than elastin. Comparing the behaviour of these proteins with
that of α-elastin would enable the affects of the structural differences between them
to be examined.
7.4.1
Surface Rheometry
0
In order to probe the frequency dependence of G and G00 it would be useful to repeat
the experiments in chapter 4 using a closer spaced range of oscillation frequencies.
It would also be interesting to probe the low pressure region and use phase lag to
see if, like β-lactoglobulin [144], α-elastin behave elastically below 10 mN/m.
Using a variety of compression speeds and then fitting to Eyring’s model will allow
an examination of how the segment length in the concentrated regime changes with
compression speed. It has been observed in PVA [124] that the speed of compression
185
affects the segment length and it would be interesting to see if the same effect occurs
in α-elastin.
The semi-dilute to concentrated phase transition could be examined on a variety
of subphases to see if differences observed were due to the presence of ions in the
subphase or the presence of the Ca2+ ion specifically. Given the observations made
in chapter 5 it is suggested that calcium ions will not be seen to have a significantly
difference affect from sodium ions.
Further measurements could be made to examine the surface shear viscosity at
higher surface pressures and as the temperature is altered. It was seen that αelastin will form monolayers at temperatures above which bulk solutions coacervate,
therefore seeing how the temperature affects the viscosity will indicate if there are
changes occurring in the monolayer.
Applying these experimental techniques to mixed lipid-elastin monolayers would
enable more information about their behaviour to be be gathered. Comparing the
viscoelastic properties of a lipid monolayer before and after penetration by α-elastin
would give further indication of the effect of the addition of elastin and would be
comparable with the dilational modulus data discussed in chapter 6.
7.4.2
α-elastin monolayers
It would be interesting to probe the higher surface pressure regime since some preliminary experiments, and the dynamic compresions, suggested a phase transition
at around 25 mN/m. Given the phase transition observed in this work was semidilute to concentrated it is suggested that this higher pressure transition must be
concentrated to melt.
The compression of α-elastin monolayers to examine their stress relaxation was
carried using a slow compression to simplify the relaxation. Using a faster compres-
186
sion would enable examination of the faster timescale relaxations, these could then
be compared the with work of Cicuta [109] using β-casein and β-lactoglobulin.
7.4.3
Lipid-Elastin Interactions
There remains much scope for the examination of the interaction of α-elastin with
PC, PC/PS mixtures of different ratios and PS. Determination of the minimum
bulk concentration that elastin will adsorp to the surface from and the maximum
surface pressure at which adsorption can occur, would provide further information
about the ability of α-elastin to penetrate and interact with lipid monolayers. It
would also be interesting to examine α-elastin interactions with phosphophadtidyl
ethanolamine (PE) the other major phospholipid constituent of cell membranes.
187
Bibliography
[1] G.W.G Chalmers, J.M. Gosline, and M.A. Lillie. J. Exp. Biol., 202:301–314,
1999.
[2] R.P. Mecham. Methods, 45(1):32–41, 2008.
[3] A.E.E. Miao, M.and Bruce, T. Bhanji, E.C. Davis, and F.W. Keeley. Matrix
Biol., 26(2):115–124, 2007.
[4] V. Achan. Pediatr. Cardiol., 20:94–96, 1999.
[5] T. Chowdhury and W. Reardon. Pediatry Cardiology, 20:103–107, 1999.
[6] S.D Shapiro, S.K. Endiscott, M.A. Province, J.A. Pierce, and Campbell E.J.
J. Clin. Invest., 87:1828–1894, 1991.
[7] S. Ritz-Timme, I. Laumeier, and M.J. Collins. Br. J. Dermatol., 149:951–959,
2003.
[8] S.M. Partridge and H.F. Davis. Biochem. J., 61(1):21–30, 1955.
[9] D.W. Urry. Ultrastructural Pathology, 4:227–251, 1983.
[10] K.K. Kumashiro, J.P. Ho, W.P. Niemczura, and F.W. Keeley. J. Biol. Chem.,
281(33):23757–23765, 2006.
188
[11] W.F. Daamen, J.H. Veerkamp, J.C.M. van Hest, and T.H. van Kuppevelt.
Biomaterials, 28:4378–4398, 2007.
[12] D. He, M. Chung, E. Chan, T. Alleyne, K.C.H. Ha, M. Miao, R.J. Stahl, F.W.
Keeley, and J. Parkinson. Matrix Biol., 26(7):524–540, 2007.
[13] L. Debelle, A.J.P. Alix, M.-P. Jacob, J.-P. Huvenne, M. Berjot, B. Sombret,
and P. Legrand. J. Biol. Chem., 270(44):26099–26103, 1995.
[14] E. Green, R. Ellis, and P. Winlove. Biopolymers, 89(11):931–940, 2008.
[15] B.A. Cox, B.C. Starcher, and D.W. Urry. J. Bio. Chem, 249(3):997–998, 1974.
[16] S.M. Partridge, H.F. Davis, and G.S. Adair. Biochem. J., 61(1):11–21, 1955.
[17] B.A. Cox, B.C. Starcher, and D.W. Urry.
Biochim. et Biophys. Acta,
317(1):209–213, 1973.
[18] K. Miyakawa and K. Kaibara. J. Phys. Soc. Jpn., 60(5):1468–1471, 1991.
[19] K. Miyakawa, Y. Ito, and K. Kaibara. J. Phys. Soc. Jpn., 62(7):2511–2515,
1993.
[20] G. Abatangelo and D. Daga-Gordini. Biochim. et Biophys. Acta, 342:281–289,
1974.
[21] C.M. Bellingham, K.A. Woodhouse, P. Robson, S.J. Rothstein, and F.W.
Keeley. Biochim. et Biophys. Acta, 1550:6–19, 2001.
[22] F.W Keeley, C.M. Bellingham, and K.A. Woodhouse. Phil. Trans. R. Soc.
Lond. B, 357:185–189, 2002.
[23] B. Bochiccio, N. Floquet, A. Pepe, A.J.P. Alix, and A.M. Tamburro. Chem.
Eur. J., 10:3166–3176, 2004.
189
[24] B. Li, D.O.V. Alonso, and V. Daggett. J. Mol. Biol., 305:581–592, 2001.
[25] K Kaibara, K. Akinari, Y. amd Okamoto, Y. Uemura, S. Yamamoto, H. Kodama, and M. Kondo. Biopolymers, 39:189–198, 1996.
[26] L. Duca, N. Floquet, A.J.P. Alix, B. Haye, and L. Debelle. Crit. Rev. Oncol.
Hemat., 49:235–244, 2004.
[27] C.P. Winlove, K.H. Parker, A.R. Ewins, and N.E. Birchler. J. Biomech. Eng.,
114:293–300, 1992.
[28] W.F. Daamen, T. Hafmans, J.H. Veerkamp, and T.H. van Kuppevelt. Biomaterials, 22:1997–2005, 2001.
[29] R.B. Rucker, D Ford, W. Goettlich Riemann, and K Tom. Calc. Tiss. Res.,
14:317–325, 1974.
[30] A. Noma, T. Takahashi, and T. Wada. Atheroscrosis, 38:373–382, 1981.
[31] L. Debelle and A.J.P. Alix. Biochim., 81:981–994, 1999.
[32] L.B. Sandberg, N. Weissman, and D.W. Smith. Biochemistry, 8(7):2940–2945,
1969.
[33] W.R. Gray, L.B. Sandberg, and J.A. Foster. Nature, 246:461–466, 1973.
[34] M. Gacko. Cell. Mol. Biol. Lett., 5:327–348, 2000.
[35] D.W. Urry, T. Hugel, M. Seitz, H.E. Gaub, L. Sheiba, J. Dea, J. Xu, and
T. Parker. Phil. Trans. Roy. Soc. Lond. B, 357:169–184, 2002.
[36] B. Vrhovski, S. Jensen, and A.S. Weiss. Eur. J. Biochem., 250:92–98, 1997.
[37] J.P. Mackay, L.D. Muiznieks, P. Toonkool, and A.S. Weiss. J. Struc. Biol.,
150:154–162, 2005.
190
[38] B. Vrhovski and A.S. Weiss. Eur. J. Biochem., 258:1–18, 1998.
[39] B.A. Kozel, H. Wachi, E.C. Davis, and R.P. Mecham.
J. Biol. Chem.,
278(20):18491–18498, 2003.
[40] L. Debelle, A.J.P. Alix, S.M. Wei, M.-P. Jacob, J.-P. Huvenne, and M. Berjot.
Eur. J. Biochem., 258:533–539, 1998.
[41] P. Brown-Augsburger, T Broekelmann, J. Rosenbloom, and R.P. Mecham.
Biochem. J., 318:149–155, 1996.
[42] N. Floquet, M. Pepe, A. Dauchez, B. Bochicchio, A.M. Tamburro, and A.J.P.
Alix. Matrix Biol., 24:271–282, 2005.
[43] L. Debelle, S.M. Wei, M.P. Jacob, W. Hornebeck, and A.J.P. Alix. Eur. J.
Biophys., 21:321–329, 1992.
[44] D. Gheduzzi, D. Guerra, B. Bochicchio, A. Pepe, A.M. Tamburro,
D. Quaglino, S. Mithieux, A.S. Weiss, and I. Pasquali Ronchetti. Matrix
Biol., 24(1):15–25, 2005.
[45] C. Fornieri, M. Baccarani-Contri, D. Quaglino Jr., and I. Passquali Ronchetti.
J. Cell Biol., 105(3):1463–1469, 1987.
[46] E.R. Dalferes, B. Radhakrishnamurthy, H.A. Ruiz, and G.S. Berenson. Exp.
Mol. Pathol., 47(3):363–376, 1987.
[47] M.P. Jacob, W. Hornebeck, and L. Robert. Int. J. Biol. Macromol., 5(5):275–
278, 1941.
[48] U.R. Rodgers and A.S. Weiss. Pathologie Biologie, 53:390–398, 2005.
[49] S.M. Partridge, D.F. Elsden, and J. Thomas. Nature, 197:1297–1298, 1963.
191
[50] J. Thomas, D.F. Elsden, and S.M. Partridge. Nature, 200:651–652, 1963.
[51] S.R. Kakivaya and C.A.J. Hoeve. Proc. nat. Acad. Sci. USA, 72(9):3505–3507,
1975.
[52] D.W. Urry. Faraday Discussions, 61:205–212, 1976.
[53] C.M. Kielty and C.A. Shuttleworth. Microsc Res. Techniq, 38:413–427, 1997.
[54] C.M. Kielty, M.J. Sherratt, and C.A. Shuttleworth. J. Cell Sci., 115:2817–
2828, 2002.
[55] H. Wachi, F. Sato, J. Nakazawa, R. Nonaka, Z. Szabo, Z. Urban, T. Yasunaga,
I. Maeda, B.C. Starcher, R.P. Mecham, and Y. Seyama. Biochem. J., 402:63–
70, 2007.
[56] M.A. Lillie and J.M. Gosline. Biomaterials, 28:2021–2031, 2007.
[57] L. Debelle and A.M. Tamburro. J. Exp. Biol., 31:261–272, 1999.
[58] A.H. Simmons, C.A Michal, and L.W. Jelinski. Science, 271:84–87, 1996.
[59] C.A. Benassi, R. Ferroni, M. Guarneri, A. Guggi, A.M. Tamburro, R Tomatis,
and R. Rocchi. FEBS Lett., 14(5):346–348, 1971.
[60] A.M. Tamburro, A. Pepe, B. Bochicchio, D. Quaglino, and I.P. Ronchetti. J.
Biol. Chem., 280(4):2682–2690, 2005.
[61] A. Ostuni, B. Bochicchio, M.F. Armentano, F. Bisaccia, and A.M. Tamburro.
Biophys. J., 93:3640–3651, 2007.
[62] A. Pepe, B. Bochicchio, and A.M. Tamburro. Nanomedicine, 2(2):203–218,
2007.
[63] V. Villani and A.M. Tamburro. J. Mol. Struc. (Theochem), 308:141–157, 1994.
192
[64] G.S. Adair, H.F. Davis, and S.M. Partridge. Nature, 1670:605, 1951.
[65] M. Mammi, L. Gotte, and G Pezzin. Nature, 220:371–373, 1968.
[66] J.R. Bendall. Biochem. J., 61:31–32, 1955.
[67] B.C. Starcher and D.W. Urry. Bioinorg. Chem, 3:107–114, 1974.
[68] J. Stiegman, E. Peggion, and A. Coasani. J. Am. Chem. Soc., 91(7):1822–
1829, 1969.
[69] D.W. Urry, B. Starcher, and S.M. Partridge. Nature, 222:795–796, 1969.
[70] L.B. Sandberg, W.R. Gray, and C. Franzblau, editors. Elastin and Elastic
Tissues, 1977.
[71] V. Velebný, M. Ledvina, J. Wimmerová, and V. Lankašová. Connect. Tissue
Res., 47:219–226, 1980.
[72] R.J. Elliott and L.T. McGrath. Calcif. Tissue Int., 54:268–273, 1994.
[73] L. Robert. Pathologie Biologie, 50:503–511, 2002.
[74] I. Gonçalves, M.W. Lindholm, L.M. Pedro, N. Dias, J. Fernandes e Fernandes,
G.N. Fredrikson, J. Nilsson, J. Moses, and M.P.S. Ares. Stroke, 35(12):2795–
2800, 2004.
[75] G. Abatangelo, D. Daga-Gordini, G. Garbin, and R. Cortivo. Biochim. et
Biophys. Acta, 371:526–533, 1974.
[76] A.M. Tamburro, V. Guantieri, D. Daga-Gordini, and G. Abatangelo. J. Biol.
Chem., 253:2893–2894, 1978.
[77] M. Terbojevich, M. Palumbo, A. Cosani, E. Peggion, L. Gotte, and S. Garbisa.
Biopolymers, 20:2213–2223, 1981.
193
[78] M.P. Molinari Tosatti, L. Gotte, and Moret V. Calc. Tiss. Res., 6:329–334,
1971.
[79] D.W. Urry. Proc. Nat. Acad. Sci. USA, 66(4):810–814, 1971.
[80] D.W. Urry, J.R. Krivacic, and J. Haider. Biochem. Bioph. Res. Co., 43(1):6–
11, 1971.
[81] M.M. Long, D.W. Urry, and B.C. Starcher. Biochem. Biophy. Res. Co.,
56(1):14–20, 1974.
[82] B.C. Starcher and D.W. Urry. Biochem. Bioph. Res. Co., 53(1):210–216, 1973.
[83] S. Koutsopoulos, P.C. Paschalakis, and E. Dalas. Langmuir, 10:2423–2428,
1994.
[84] L. Robert. Athersclerosis, 123:169–179, 1996.
[85] W Hornebeck and S.M. Partridge. Eur. J. Biochem., 51:73–78, 1975.
[86] G.J. Tortora and S.R. Grabowski. Principles of Anatomy and Physiology, 9th
Edition. John Wiley and Sons, Inc, 2000.
[87] A. Vyavahare, M. Ogle, F.J. Schoen, and R.J. Levy. Am. J. Path., 155(3):973–
982, 1999.
[88] M.M. Long, D.W. Urry, R.S. Rapaka, and W.D. Thompson. Biophys. J.,
21(3):A80, 1978.
[89] D.W. Urry, K. Okamoto, R.D. Harris, C.F. Hendrix, and M.M. Long. Biochemistry, 15(18):4083–4089, 1976.
[90] K. Miyakawa, M. Totoki, and K. Kaibara. Biopolymers, 35:85–92, 1995.
[91] D.W. Urry, C.F. Hendrix, and M.M. Long. Calc. Tiss. Res., 21:57–65, 1976.
194
[92] L. Robert, A.M. Robert, and B. Jacotot. Athersclerosis, 140:281–295, 1998.
[93] M.A. Lillie and J.M. Gosline. Biopolymers, 64:581–592, 2001.
[94] D. Kramsch. Exp. Geront., 4:1–6, 1969.
[95] S. Hashimoto, Y. Seyama, T. Yokokura, and M. Mutai. Exp. Mol. Pathol.,
37(2):150–155, 1982.
[96] S. Hashimoto, T. Setoyama, T. Yokokura, and M. Mutai. Exp. Mol. Pathol.,
36(1):99–106, 1982.
[97] T. Fulöp Jr., N. Douziech, M.P. Jacob, M. Hauck, J. Wallach, and L. Robert.
Pathol. Biol., 49:339–348, 2001.
[98] A. Hinek, D.S. Wrenn, R.P. Mecham, and S.H. Barondes.
Science,
239(4847):1539–1541, 1988.
[99] R.P. Mecham, A. Hinek, D.S. Wrenn, and G.L. and Griffin. Biochemistry,
28(9):3716–3722, 1989.
[100] A.C. Burton. Physiol. Rev., 34(4):619–642, 1954.
[101] E.N. Lamme, R.T.J. van Leeuwen, A. Jonker, J. van Marle, and E. Middlekoop. J. Invest. Dermatol., 111:989–995, 1998.
[102] S. Neuenschwander and S.P. Hoestrup. Transpl. Immunol., 12:359–365, 2004.
[103] A. Pockels. Nature, 43:437–439, 1891.
[104] M. Van Den Tempel. J. Non-Newtonian Fluid Mech., 2:205–219, 1977.
[105] I. Langmuir. P. Natl. Acad. Sci. USA, 3:251–257, 1917.
195
[106] V.M Kaganer, H Möhwald, and P. Dutta. Rev. Mod. Phys., 71(3):779–819,
1999.
[107] P Cicuta and E.M. Terentjev. Eur. Phys. J. E, 16:147–158, 2005.
[108] M. Cejudo Fernández, C. Carrera Sánchez, M.R. Rodrı́guez Niño, and J.M.
Rodrı́guez Patino. Langmuir, 23:7178–7188, 2007.
[109] P Cicuta. J. Colloid Interf. Sci., 308:93–99, 2007.
[110] P Cicuta, E.J. Stancik, and Fuller G.G. Phys. Rev. Lett., 90(23):236101, 2003.
[111] P Circuta and I Hopkinson. J. Phys. Chem., 114(19):8659–8670, 2001.
[112] J. Ding, H.E. Warner, and J.A. Zasadzinski. Langmuir, 18:2800–2806, 2002.
[113] J. Ding,
H.E. Warner,
and J.A. Zasadzinski.
Phys. Rev. Lett.,
88(16):168102(4), 2002.
[114] A. Lucero Caro, M.R. Rodrı́guez Niño,, and J.M. Rodrı́guez Patino. Colloid
Surface A, 327:79–89, 2008.
[115] M. Doi and S.F. Edwards. The Theory of Polymer Dynamics. Oxford University Press, 1988.
[116] E.G. Richards. An introduction to the physical properties of large molecules
in solution. Cambridge University Press, 1980.
[117] E. Spigone, G.-Y. Cho, G.G. Fuller, and P. Cicuta. Langmuir, 25(13):7457–
7464, 2009.
[118] T.A.M. Ferenczi and P. Cicuta. J. Phys.: Condens. Matter, 71:S3445–S3453,
2005.
[119] R. Vilanove and F Rondelez. Phys. Rev. Lett., 45(18):1502–1505, 1980.
196
[120] F. Monroy, F. Ortega, R.G. Rubio, and M.G. Verlade. Adv. Colloid Interfac.,
134–135:175–189, 2007.
[121] M. Daoud, J.P. Cotton, B. Farnoux, G. Jannink, G. Sarma, H. Benoit, R. Duplessix, C. Picot, and P.G. de Gennes. Macromol., 8(6):804–818, 1975.
[122] F. Monroy, F. Ortega, and R.G. Rubio. Eur. Phys. J. B, 13:745–754, 2000.
[123] F. Monroy, H.M. Hilles, F. Ortega, and R.G. Rubio.
Phys. Rev. Lett.,
91(26):268302–1–268302–4, 2003.
[124] E. Spigone, G.-Y. Cho, G.G. Fuller, and P. Cicuta. Langmuir, 25(13):7457–
7464, 2009.
[125] V. Aguié-Béghin, E. Leclerc, M. Daoud, and R. Douillard. J. Colloid Interfac.
Sci., 214:143–155, 1999.
[126] A. Maestro, F. Ortega, F. Monroy, J. Kr¨’agel, and R. Miller. Langmuir,
25(13):7393–7400, 2009.
[127] R. Vilanove, Poupinet, and F Rondelez. Macromol., 21:2880–2887, 1988.
[128] M.R. Buhaenko, J.W. Goodwin, and R.M. Richardson. Thin Solid Films,
159:171–189, 1988.
[129] E. Evans and D. Needham. J. Phys. Chem, 91(16):4219–4228, 1987.
[130] K.-H. Yoo and H. Yu. Macromol., 22:4019–4026, 1989.
[131] J.T. Petkov, T.D. Gurkov, B.E. Campbell, and R.P. Borwanker. Langmuir,
16:3730–3711, 2000.
[132] P Cicuta and I Hopkinson. Europhys. Lett., 68(1):65–71, 2004.
[133] J. Lucassen and M. Van Den Tempel. Chem. Eng. Sci., 27:1283–1291, 1972.
197
[134] J.D. Perry. Viscoelastic Properties of Polymers. John Wiley and Sons, Inc,
1961.
[135] R. Miller, R. Wüstneck, J. Krägel, and Kretzschmar. Biopolymers, 35:85–92,
1995.
[136] J. Thompson. Order and interactions in model biological membranes. PhD
thesis, University of Exeter, 2006.
[137] C. Tanford. Physical Chemistry of Macromolecules. John Wiley and Sons,
Inc, 1967.
[138] K. Monkos. J. Biol. Phys., 31:219–232, 2005.
[139] P.L. du Noüy. J. Gen. Physiol., 7:625–632, 1925.
[140] K. Lukenheimer and K.D. Wantke. J. Colloid Interf. Sci., 66:579–581, 1978.
[141] F.C. Goodrich, L.H. Allen, and A.K. Chatterjee. Proc. Roy. Soc. Lond. A,
320:537–547, 1971.
[142] A. Lucero Caro, M.R. Rodrı́guez Niño,, and J.M. Rodrı́guez Patino. Colloid
Surface A, 341:134–141, 2009.
[143] F. Monroy, J. Geirmanska Kahn, and D Langevin. Colloid Surface A, 143:251–
260, 1998.
[144] E.V. Aksenenko, V.I. Kovalchuk, V.B. Fainerman, and R. Millar. Adv. Colloid
Interfac., 122:57–66, 2006.
[145] R.J. Young and P.A. Lovell. Introduction to Polymers, Second Edition. Chapman and Hall, London, 1994.
[146] A.E. Dunstan. Z. physik. Chem., 56(3):370–380, 1906.
198
[147] O. Albert. Z. physik. Chem., A182(6):421–429, 1938.
[148] P.J. Flory. J. Am. Chem. Soc., 62:1057–1070, 1940.
[149] W. Kauzmann and H. Eyring. J. Am. Chem. Soc., 62:3113–3125, 1940.
[150] R.E. Powell, C.R. Clark, and H. Eyring. J. Chem. Phys., 9:268–273, 1941.
[151] G Gee and L.R.G. Treloar. Trans. Faraday Soc., 38:147–165, 1942.
[152] P.J. Flory. J. Chem. Phys., 10(1):51–61, 1942.
[153] M.L. Huggins. J. Chem. Phys., 9(5):440, 1941.
[154] M.L. Huggins. J. Phys. Chem., 46(1):151–158, 1942.
[155] B.K. Hartman and S. Bakerman. Biochem., 5(7):2221–2229, 1966.
[156] M. Sherriff and B. Warburton. Polymer, 15:2940–2945, 1974.
[157] J.P. Coffman and C.A. Naumann. Macromol., 35:1835–1839, 2002.
[158] R.C Weast, editor. Handbook of Chemistry and Physics, 53rd Edition. The
Chemical Rubber Co., Ohio, USA, 1972.
[159] A.E. Bailey, editor. Kaye and Laby Tables of Physical and Chemical Constants,
Fifteenth Edition. Longman Group Limited, 1986.
[160] R.C Weast and S.M. Selby, editors. Handbook of Chemistry and Physics, 47th
Edition. The Chemical Rubber Co., Ohio, USA, 1967.
[161] N. Ghaour, A. Aschi, L. Belbahri, S. Trabelsi, and A. Gharbi. Physica B,
404:4246–4252, 2009.
[162] S.E. Harding. Prog. Biophys. Molec. Biol., 68(2/3):207–262, 1997.
199
[163] T.G Fox Jr. and P.J. Flory. J. Am. Chem. Soc., 73(5):1915–1920, 1951.
[164] P.J. Flory. Principles of Polymer Chemistry, Sixth Edition. Cornell University
Press, Ithaca, New York, 1967.
[165] P.J. Flory and T.G Fox Jr. J. Am. Chem. Soc., 73(5):1904–1908, 1951.
[166] H. Eyring. J. Chem. Phys., 4:283–291, 1936.
[167] R.H. Ewell and H. Eyring. J. Chem. Phys., 5:726–736, 1937.
[168] M. Kawaguchi, B.B. Sauer, and H. Yu. Macromol., 22:1735–1734, 1989.
[169] Z.-H. Xue, S.-X. Dai, B.B. Hu, and Z.-L. Du. Mater. Sci. Eng. C, 29:1998–
2002, 2009.
[170] B.B. Sauer, M. Kawaguchi, and M. Yu. Macromol., 20:2732–2739, 1987.
[171] J. Benjamins and E.H. Cagna, A. Lucassen-Reynders. Colloid Surface A,
114:245–254, 1996.
[172] C.A. Smith and E.J. Wood. Molecular and Cell Biochemistry: Biological
Molecules. Chapman and Hall, London, 1991.
[173] M.A.F. Davis, H. Hauser, R.B. Leslie, and M.C. Phillips. Biochim. et Biophys.
Acta, 317:214–218, 1973.
[174] C. Carrera Sánchez, M. Cejudo Fernández, M.R. Rodrı́guez Niño,, and J.M.
Rodrı́guez Patino. Langmuir, 22:4215–4224, 2006.
[175] E Dickinson. Colloid Surface B, 15:161–176, 1999.
200

Behaviour of α-elastin in bulk and at aqueous surfaces

Transcription

Similar documents

1 Osborne Reynolds Research Student Award: The

Langmuir monolayer miscibility of single

- ATS Journals

Privé Aesthetics

The John F. Barnes Myofascial Release Approach

Swiss Cosmetic

AEREON WILDCAT Wellhead Gas Compression Marketing Sheet