CHARACTERIZATION AND MOLECULAR MOTION IN
UNILAMELLAR LIPID VESICLES

By

Heather Pillman

A DISSERTATION
Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
CHEMISTRY
2011

ABSTRACT
CHARACTERIZATION OF MOLECULAR MOTION AND
ORGANI\ZATION IN UNILAMELLAR LIPID VESICLES

By

Heather Pillman

Membrane proteins in biological cells are involved in nearly every function of cell
life.

For these reasons, there is great interest in incorporating chemically selective

biomolecules, such as transmembrane proteins and enzymes, into membrane-based
biosensors. Proteins must be reconstituted into environments that mimic the behavior of
natural plasma membranes in order for them to retain their biological specificity and
activity. Synthetic lipid bilayers are compositionally simple systems that approximate the
dynamic fluid nature of the cell membrane, making them attractive systems for the
incorporation of biomolecules.
As an initial step toward understanding bilayer systems, we have focused our
research on investigating the dynamic behavior and fluidity of phospholipid vesicles
lacking incorporated proteins. We have interrogated specific regions of our lipid vesicles
using fluorescent ―reporter‖ molecules to determine the dynamic behavior and molecularscale organization of unilamellar lipid vesicles using time-resolved fluorescence
spectroscopy techniques.

We have used the polar chromophore nitro-2-1,3-

benzoxadiazol-4-yl (NBD) tethered to the headgroups of phosphoethanolamine (PE)

lipids to study the dependence of headgroup functionality on inter-lipid hydrogen
bonding interactions. We gauged these interactions using a time-correlated single photon
counting (TCSPC) instrument to measure rotational diffusion dynamics in lipid vesicles
and fluorescence recovery after pattern photobleaching (FRAPP) to determine
translational diffusion dynamics in supported lipid bilayers. We have also used the
untethered chromophores pyrene and perylene, imbedded in the acyl chain regions of
phosphocholine (PC) lipid vesicles to determine the local viscosity and phase of the lipid
bilayers as a function of bulk ethanol concentration. We have observed a structural
perturbation in the acyl chain regions of the bilayer at ethanol concentrations of ca. 0.6
M, which points to the influence of controlled amounts of impurities on lipid order and
phase. Lastly, we have used imbedded perylene to study the effects of excess excitation
energy dissipation on the molecular-scale organization of phospholipid vesicles, and
cholesterol-containing phospholipid vesicles. In both vesicle systems, we concluded that
the order of the local environment of the chromophore determines the efficiency of
energy transfer, and the extent of local heating. Taken collectively, the results of these
experiments have provided valuable insight into the chemical and structural factors that
influence lipid bilayer fluidity.

ACKNOWLEDGEMENTS

I first and foremost have to thank my advisor, Dr. Gary J. Blanchard, for helping
me at every stage of my graduate career. For giving me advice on everything from my
research projects to job offers, and for never letting me give up or get too overwhelmed
when the laser was uncooperative. In short, thank you for always being inspiring, patient,
and positive-thinking in all things chemistry related, and teaching me everything I know
about research and lasers. I am so grateful for the opportunity I’ve had to work with you
and with the Blanchard group. Joining your group was the best decision I have ever
made.
I would also like to thank all of the teachers and professors I have had the great
pleasure of working with and learning from over the years. Thank you to my fifth-grade
teacher, Mrs. Burnett, for helping me to discover my love for science at an early age. I
also would not be here today if not for Mrs. Laura Lutz—not only my high school
chemistry teacher, but my favorite teacher—who taught me on my very first day of class
that fire alarms were fun, and that chemistry was my destiny. Thank you also to all of my
professors at MSU for encouraging me to work hard and to always strive to be a better
scientist, especially the members of my Guidance Committee. In particular, I would like
to thank Doctors Greg Swain, Gavin Reid, David Weliky, Merlin Bruening, Dana
Spence, Kathy Severin, and Robert Ofoli.

iv

I also owe a huge debt of gratitude to the scientists who’ve been my mentors over
the years, in particular Dr. Dwella Moton Nelson and Dr. Saadia Eltayeb. You, more
than anyone else, helped me find my way on the path to making chemistry my career.
In addition, I must also thank every member of the Blanchard Group (past and
present) for being both my co-workers and friends, supporting me in innumerable ways,
and keeping me sane when the road was very rough. In particular, Monika Dominska,
Benjamin Oberts, Margaretta Dimos, Douglas Gornowich, Christine Hay, and Iwan
Setiawan. I would not have succeeded in graduate school without all of you. I will miss
you.
Most importantly, I thank God for giving me the strength and courage to achieve
my dreams, and I can never thank my family and friends enough for their unflinching
confidence in me, and their constant outpourings of encouragement, love, and prayers.
Thank you to my wonderful extended family: aunts, uncles, cousins, and grandparents.
In particular, thank you to my very best friend, Joshua Nasrollahi, for always being in my
corner and refusing to ever let me give up. I love you with all my heart. Special thanks
to Nicholas and Andrew, for being great brothers and for giving me some serious
competition over the years. And lastly, immeasurable thanks to my parents, John and
Annemarie Pillman, for always believing in me and never doubting for a second that I
would succeed. Without your love and support, I would not be where I am today. I love
you both more than I can say, and I dedicate this dissertation to you.

v

TABLE OF CONTENTS

List of Tables ...................................................................................................................viii

List of Figures ................................................................................................................... ix

Chapter 1: Introduction ...................................................................................................... 1

Chapter 2: Rotational and Translational Dynamics of Headgroup-Tethered
Chromophores in Lipid Vesicles ...................................................................................... 17
Introduction ................................................................................................................. 17
Experimental Section .................................................................................................. 23
Results and Discussion ............................................................................................... 27
Conclusions ................................................................................................................. 41

Chapter 3: Investigating Membrane Fluidity of Phospholipid Bilayers Using Steady-State
Fluorescence Measurements ............................................................................................. 42
Introduction .............................................................................................................. 42
Experimental Section ............................................................................................... 49
Results and Discussion ............................................................................................ 51
Conclusions .............................................................................................................. 59

Chapter 4: Interrogating the Effects of Ethanol on the Organization of Lipid Vesicles
Using Time-Resolved Fluorescence Measurements ......................................................... 60
Introduction .............................................................................................................. 60
Experimental Section ............................................................................................... 65
Results and Discussion ............................................................................................ 68
Conclusions.............................................................................................................. 83

Chapter 5: Effects of Localized Energy Dissipation on the Motional Dynamics of
Unilamellar Vesicles ........................................................................................................ 84
Introduction ............................................................................................................. 84
Experimental Section .............................................................................................. 89
Results and Discussion ........................................................................................... 92
Conclusion ..............................................................................................................110

vi

Chapter 6: Consequences of Transient Heating on the Motional Dynamics of CholesterolContaining Phospholipid Vesicles ...................................................................................111
Introduction .............................................................................................................111
Experimental Section ..............................................................................................114
Results and Discussion ...........................................................................................117
Conclusions .............................................................................................................138

Chapter 7: Conclusions ....................................................................................................139

Literature Cited ................................................................................................................145

vii

LIST OF TABLES

Table 2.1. Fluorescence lifetimes (τ1 and τ2) and reorientation times (τOR1 and τOR2) of
NBD-PE in bilayer lipid vesicles, determined by TCSPC ................................................ 29

Table 2.2. Cartesian components of the rotational diffusion constant (MHz) of NBD-PE
chromophores in bilayer lipid vesicles ............................................................................. 34

Table 2.3. Mobilities and diffusion coefficients of NBD-PE in supported lipid bilayers
on fused silica. FRAPP data represent average values for three bilayers. At least five
different areas were interrogated for each bilayer. ........................................................... 38

Table 4.1. Cartesian components of the rotational diffusion constant and aspect ratio
Dz/Dx as a function of solution phase ethanol concentration at 289 K and 303 K. Errors
are ±1 for at least four individual determinations at each ethanol concentration ........... 75

Table 5.1. The rotational diffusion constant, aspect ratio Dz/Dx, and calculated local
viscosity as a function of temperature .............................................................................. 99

Table 5.2. Experimental reorientation times and calculated changes in local viscosity and
temperature associated with internal conversion of perylene ..........................................106

Table 6.1. Measured reorientation times as a function of system temperature...............129

Table 6.2. The rotational diffusion constant, aspect ratio Dmajor/Dminor, and calculated
local viscosity as a function of temperature .....................................................................131

Table 6.3. Experimental reorientation times and calculated changes in local viscosity and
temperature associated with internal conversion of perylene ..........................................136

viii

LIST OF FIGURES

Figure 1.1. Structures of (a) gel phase lipid bilayer and (b) fluid phase lipid
bilayer ................................................................................................................................. 5

Figure 1.2. Structures of chromophores used in this project and excitation and emission
spectra. (a) Free pyrene for the steady-state behavior of phospholipid vesicles, (b) free
perylene for dynamics measurements of the acyl chain region of phospholipid vesicles,
and (c) nitro-2-1,3-benzoxadiazol-4-yl tethered to the headgroup of phosphoethanolamine lipids for studying the effects of headgroup hydrogen bonding on lipid
vesicle dynamics. .............................................................................................................. 9

Figure 1.3. (a) Time-resolved emission intensities for polarizations collected parallel
( I (t ) ) and perpendicular ( I  (t ) ) to the vertically polarized excitation pulse. (b) The
functional form of the orientational anisotropy decay obtained from transients in (a).
Residuals are distributed along the y-intercept and the solid line is the fit of the
data .................................................................................................................................... 12

Figure 2.1. Excitation and emission spectra of 1e-6 M NBD-phosphatidylethanolamine
(NBD-PE) in chloroform. ................................................................................................. 19

Figure 2.2. Molecular structures of lipids used in this study: dioleoylphosphatidylcholine
(DOPC), dioleoylphosphatidylglycerol (DOPG), and the chromophore NBDphosphatidylethanolamine (NBD-PE) .............................................................................. 20

Figure 2.3. (a) Fluorescence lifetime and (b) induced orientational anisotropy data for
NBD-PE in a DOPG/DOPC mixed lipid vesicle. The residuals for a fitted twocomponent exponential decay function are shown for both panels .................................. 28

Figure 2.4. A typical FRAPP recovery curve for a DOPC/NBD-PE supported bilayer
membrane (above) along with residuals (below). Total photobleaching time was
300 ms ............................................................................................................................... 37

Figure 3.1. Structures of 14:0 DMPC, pyrene, and absorbance and emission spectra of
1e-5 M pyrene in chloroform ............................................................................................ 46
ix

Figure 3.2. Emission spectra of pyrene demonstrating the relationship between
concentration and excimer band intensity in 3e-6 M pyrene in DMPC vesicles (dotted
line) and 3e-5 M pyrene in DMPC vesicles (solid line). .................................................. 47

Figure 3.3. Pyrene excimer band intensity for the chromophore incorporated into 100 nm
diameter DMPC vesicles, as a function of ethanol concentration. Error bars are ± 1σ for
three individual determinations at each ethanol concentration. The dashed line is a linear
regression fit to the data .................................................................................................... 54

Figure 3.4. Dynamic light scattering data on vesicle size as a function of solution ethanol
concentration. Uncertainties reported are ± 1 for at least three determinations ............ 57

Figure 4.1. Structures of 14:0 DMPC and perylene, and normalized absorption and
emission spectra of perylene in DMPC vesicles ............................................................... 69
Figure 4.2. (a) Experimental data for perylene reorientation in a 100 nm diameter DMPC
vesicle, with 0.25 M ethanol concentration at 289 K. (b) Induced orientational
anisotropy decay function for the data shown in panel (a), along with the best-fit function
to the data. For this fit the function is f(t) = (0.11 ± 0.004)exp(-t/(339±24 ps)) + (0.04 ±
0.003)exp(-t/(2483±185 ps)) ............................................................................................. 73

Figure 4.3. (a) Dependence of perylene Dx and Dz on ethanol concentration for 100 nm
diameter DMPC vesicles at 289 K. (b) Dependence of perylene Dx and Dz on ethanol
concentration for 100 nm diameter DMPC vesicles at 303 K. Errors are ±1 for at least
four individual determinations at each ethanol concentration. ......................................... 76

Figure 5.1. Structures of 14:0 DMPC and perylene, and normalized excitation and
emission spectra of perylene in DMPC vesicles ............................................................... 96

Figure 5.2. (a) Experimental data for perylene reorientation in 100 nm diameter DMPC
vesicles at 292 K, excited at 435 nm (S1 ← S0). (b) Induced orientational anisotropy
de-cay function for the data shown in panel (a), along with the best-fit function to the
data .................................................................................................................................... 98

x

Figure 5.3. (a) Experimental data for perylene reorientation in 100 nm diameter DMPC
vesicles at 292 K, and excited at 260 nm (S2 ← S0). (b) Induced orientational anisotropy decay function for the data shown in panel (a), along with the best-fit function to
the data. ............................................................................................................................101
Figure 5.4. Dependence of reorientation time on temperature for 100 nm diameter
DMPC vesicles, for excitation of the S1 ← S0 transition (○) and the S2 ← S0 transition
(●). Errors are ±1 for at least four individual determinations at each
temperature ......................................................................................................................104

Figure 6.1. Structures of 14:0 DMPC, perylene, cholesterol, and normalized excitation
and emission spectra of perylene in DMPC vesicles .......................................................119

Figure 6.2. (a) Experimental data for perylene reorientation in 100 nm diameter DMPC
vesicles at 270.5 K, excited at 430 nm (S1 ← S0). (b) Induced orientational anisotropy
decay function for the data shown in panel (a), along with the best-fit function to
the data .............................................................................................................................122

Figure 6.3. (a) Experimental data for perylene reorientation in 100 nm diameter DMPC
vesicles at 273 K, excited at 260 nm (S2 ← S0). (b) Induced orientational anisotropy
decay function for the data shown in panel (a), along with the best-fit function to
the data .............................................................................................................................123

Figure 6.4. (a) Experimental data for perylene reorientation in 100 nm diameter DMPC
vesicles at 283 K, excited at 260 nm (S2 ← S0). (b) Induced orientational anisotropy
decay function for the data shown in panel (a), along with the best-fit function to
the data .............................................................................................................................124

xi

CHAPTER 1
INTRODUCTION
The development of novel, chemically robust sensing devices that mimic the
behavior and functionality of biological cells are in increasingly high demand in a variety
of markets. These biologically-based sensors have various applications in areas such as
2

medical diagnostics (i.e. blood glucose monitors) and the detection of biological warfare
3-4

agents and toxins

.

Biosensors, which borrow design elements from biological

components, consist of three main parts: an analyte of interest, a bioreceptor, and a
transducer. A bioreceptor is a biologically-derived molecule, such as an enzyme or a
transmembrane protein that is chemically selective for a specific analyte. The interaction
of the analyte with the bioreceptor results in the production of a chemical signal, which
the transducer then converts into a new type of signal by electrochemical or optical
means, for example. This new signal is detected with the aid of electronics, allowing for
interpretation by the user. In this way, an analyte of interest can be measured and
quantified.
Biological cell membranes are the regulators of cell life; they mediate the influx
of nutrients and the release of waste, and are extremely complex and dynamic in nature.
Cell membrane structure and fluidity play critical roles in regulating the life of the
cell.

5-6

The physical properties of the cell membrane are vital to cell function, including

protein activity and membrane permeability.

A biological cell membrane typically

consists of approximately 100 – 500 different chemical constituents, depending on the

1

location and function of the cell in the organism. The cell membrane is composed of
various lipids, including phospholipids, glycerophospholipids, and sphingolipids, which
make up a bilayer leaflet structure in which other components are imbedded. Some other
membrane constituents are carbohydrates, cholesterol, sphingomyelin, and peripheral and
transmembrane proteins. All of these constituents work cooperatively to regulate the cell
and, for this reason, the cell membrane is often touted as being the ideal platform for
biosensor technology.

7-9

Proposed membrane-based biosensors consist of a phospholipid bilayer with
imbedded transmembrane proteins fused directly to a solid substrate, such as gold or
quartz, which would act as a chemical or optical transducer.

10-12

Then using

electrochemical means, for example, the flow of charged analytes could be detected and
measured.

However, there are several issues that need to be addressed before a

functional membrane-based biosensor could be constructed.

The first obstacle to

overcome is the insertion of a functional transmembrane protein.

Transmembrane

proteins lose their biological activity when they are removed from the fluid environment
of the plasma membrane of the cell; therefore, the surface of the transducer must be
modified in such a way that it will support a fluid environment similar to that of a
biological membrane, in order to promote protein reconstitution.

In order for a

transmembrane protein to function properly in a lipid bilayer, both ends of the protein
must have access to the surrounding aqueous environment. If a lipid bilayer is in direct
contact with a solid substrate, there will be very little space for a protein to imbed, and if
a protein does imbed, it will most likely denature and become biologically inactive.

2

Several methods have been proposed for rectifying this issue; one promising approach is
achieved by the formation of a hydrophilic spacer between the substrate surface and the
lipid bilayer. The hydrophilic spacer is prepared by conjugating lipids with a short chain
polymer, such as polyethylene glycol (PEG), terminated with a thiol group, and fusing
the functionalized lipids to a gold substrate, thus tethering the bottom lipid leaflet to the
13-16

substrate.

This method supports the formation of an aqueous layer between the

substrate and the bottom lipid leaflet, which helps the bilayer to maintain some of its fluid
properties and protein bioactivity by shielding the protein from direct contact with the
substrate.
It was the goal of this project to investigate the molecular motion and
organizational properties of simple lipid bilayer systems under a variety of conditions to
provide a foundation for the development of membrane-based biosensors.

For our

experiments we considered two lipid bilayer systems: the planar lipid bilayer and the
lipid vesicle or liposome. As was outlined above, planar lipid bilayers or solid-supported
lipid bilayers appear to be the ideal platform for the development of biomimetic sensors,
and can be prepared in a step-wise manner, such as the transfer of individual lipid
monolayers from an air-water interface to a solid substrate using Langmuir—Blodgett
18

(lower leaflet) and Langmuir—Schaefer

17

(upper leaflet) techniques. These methods of

lipid bilayer deposition produce chemically robust and stable solid-supported bilayer
systems. However, the incorporation of transmembrane proteins into solid-supported
lipid bilayers prepared using the Langmuir—Blodgett and Langmuir—Schaefer
techniques presents a challenge because a requirement of monolayer deposition is that the
3

proteins are located at the air-water interface, which causes them to denature. Another
method for the preparation of solid-supported lipid bilayers that is more conducive to the
incorporation of transmembrane proteins is the adsorption and fusion of unilamellar lipid
vesicles onto the substrate surface.

19-21

Phospholipids are a major component in biological cell membranes and are
amphipathic molecules that consist of a polar headgroup region and two nonpolar acyl
chains. In aqueous solution phospholipids aggregate and self-assemble into spherical
lipid vesicle structures, where the polar headgroups make up the interior and exterior
surface area of the vesicle in direct contact with the aqueous environment, and the acyl
chain moieties are directed toward one another creating a hydrophobic inner leaflet
region.

4

Lipid vesicles not only make excellent model cell membrane systems, but they

are relatively simple to prepare and manipulate, and their structure promotes the
incorporation of transmembrane proteins. For these reasons, lipid vesicle fusion is an
extremely appealing method for the formation of solid-supported lipid bilayers, and
ultimately for the development of membrane-based biosensors.

However, before

attempting to incorporate functional transmembrane proteins into lipid vesicles, it is
necessary to first investigate the behavior and dynamic properties of simple lipid vesicle
systems.
The interactions between neighboring lipids, as well as the interactions between
lipids and non-lipid constituents influence the dynamic properties of the bilayer
22

membrane.

One way a biological cell can regulate itself is by manipulating the fluidity

of the membrane.

5

In this way, the cell is able to control the permeability of the
4

membrane, and selectively activate proteins.

Membrane fluidity is a concept that

describes a number of intricate molecular motions, and can be thought of as a measure of
the viscous drag effects between individual membrane constituents.

Some of the

molecular motions that comprise membrane fluidity are lipid rotational, translational, and
lateral diffusion, as well as protrusion in and out of the bilayer, and the formation of
trans-gauche isomers or kinks.

23

The concept of membrane fluidity is also related to the phase of the lipids. All
lipids have a characteristic gel-to-fluid phase transition temperature, Tm. The Tm can be
thought of as the lipid melting point. At temperatures below Tm, lipids exist in a highly
ordered ―gel‖ phase characterized by fully extended acyl chains, tilted at an angle slightly
away from the bilayer normal. In gel phase lipid bilayers, the close-packed structure of
the bilayer, strong van der Waals interactions between acyl chain moieties, and hydrogen
bonding interactions between the polar headgroups and the surrounding aqueous medium
all contribute to the bilayer exhibiting very little motional freedom. At temperatures
above Tm, lipid acyl chains become disordered as the structure of the hydrocarbon tails
becomes increasingly disrupted by the formation of trans-gauche conformers.

24

These

kinks in the formerly trans-configured lipid acyl chains disrupt both the lipid order and
the inter-lipid interactions, resulting in an increase in the motional freedom of the lipids
and the formation of the ―fluid‖ phase. The gel-to-fluid phase transition is demonstrated
in Figure 1.1. The phase and fluidity of the lipid bilayer are influenced by a number of
factors, including the length of the acyl chains, the existence of any unsaturations in the

5

(a)

(b)

Figure 1.1. Structures of (a) gel phase lipid bilayer and (b) fluid phase lipid bilayer.

6

chains, the presence of non-lipid constituents in the bilayer, and the temperature of the
system.

25

The effect of lipid phase on the fluidity and dynamic properties of the

membrane must be better understood in compositionally simple systems if there is to be
hope of constructing a membrane-based biosensor.
Fluorescence techniques are especially appealing in biological applications
because they are non-destructive and highly sensitive.

26

To study the effects various

chemical ―impurities‖ and temperature-induced perturbations have on the dynamic
behavior of lipid vesicle systems, we utilized a fluorescent molecular probe or
chromophore, which acts as a reporter molecule, to investigate these effects. In order to
study the consequences of various chemical and physical perturbations on the fluidity of
the lipid bilayer, we need to exploit the fluorescent properties of the chromophore.
Fluorescence is the emission of energy in the form of photons from an electronically
excited species.

27

Fluorescent molecules are first excited by the selective absorption of

photons at a specific wavelength. The excited molecules can then emit photons at a
longer wavelength than that required for excitation, and by this radiative relaxation
mechanism, return to the ground state.
electronic state, Kasha’s rule

28

Although molecules may be excited to an

states that emission occurs only from the transition from

the lowest vibrational level of the lowest energy excited state to the ground state. The
average amount of time electronically excited molecules remain in the excited state
before emission is the fluorescence lifetime. The interaction of excited state molecules
with their local environment can affect the molecules’ fluorescence spectrum and

7

lifetime, so by monitoring the optical response of a population of fluorescent probes, we
29-33

are able to derive information about the local medium of the chromophores.

Chromophores provide detailed information about the structure, behavior, and
polarity of their local environments, and can be chosen to interrogate a selected region of
the bilayer. For example, a small nonpolar probe preferentially resides in the acyl chain
region of the bilayer while a polar chromophore will associate with the headgroup region
of the bilayer, or the aqueous surroundings. By selectively exciting the chromophore
sequestered in its local environment, and measuring the resulting fluorescence signal, one
is able to collect both steady-state and dynamic information on the probe, and acquire
information about the surrounding molecules as well.

Time-domain fluorescence

measurements are typically in the form of fluorescence lifetime or anisotropy,

26

and are

related to the rotational and translational diffusion of the chromophores in their
environment, as well as the motion of the molecules surrounding the probe. These data
can then be used to elucidate information about the membrane permeability, fluidity,
local viscosity, and lipid phase.
To begin investigation of vesicle systems, a chromophore was initially chosen to
interrogate either the acyl chain region of the bilayer or the lipid headgroup region,
selectively. The aromatic hydrocarbon pyrene (Fig. 1.2a) was chosen to interrogate the
acyl chain region of the lipid bilayer. Due to pyrene’s nonpolar, planar structure, it
preferentially imbeds in the nonpolar region of the bilayer.

29-33

Pyrene is also useful as

a molecular probe because of its sensitivity to solvent polarity and its ability to form
34

excimers.

An excimer, (MM)*, is formed when an electronically excited molecule of
8

pyrene (M*) collides and reacts with a ground state molecule of pyrene (M). The process
can be thought of as dynamic fluorescence quenching of the pyrene monomers.

35

When

a pyrene excimer is formed, its characteristic emission band is broad, structureless, and
red shifted from the emission bands of pyrene monomer. The formation of excimers is
concentration- dependent, therefore the higher the concentration of pyrene monomer in
solution prior to excitation, the higher the probability of a collision between an excited
state and ground state monomer during the lifetime of the excited state.

27

Because of

the concentration dependence of excimer formation, the intensity of the excimer emission
band scales with pyrene concentration. Excimer formation is also a diffusion-limited
process, meaning the temperature and viscosity of the medium affect the intensity of the
excimer emission band centered at ca. 472 nm. We exploited pyrene’s ability to form
excimers to study the effects of ethanol perturbation on phospholipid vesicle systems, and
we discuss these experiments in Chapter 3.
To interrogate the rotational diffusion dynamics of our lipid vesicle systems, we
used time-resolved fluorescence spectroscopy techniques to measure the fluorescence
anisotropy of either headgroup-tethered nitro-2-1,3-benzoxadiazol-4-yl (NBD, Figure
1.2c) or the polycyclic aromatic hydrocarbon perylene (Figure 1.2b), which selectively
localizes in the nonpolar aliphatic chain region of the bilayer. Chromophores with
absorption transition moments oriented in a specific direction will preferentially absorb
light that is polarized along this same axis. So, if a short, intense pulse of vertically
polarized light is directed into a sample of randomly oriented chromophores, the
molecules with transition moments oriented nominally parallel to the incident pulse will

9

normalized intensity (a.u.)

(a)

1.0

absorbance
emission

0.8
0.6

excimer band

0.4
0.2
0.0

(b)

normalized intensity (a.u.)

300

1.0

350
400
450
wavelength (nm)

500

excitation
emission

0.8
0.6
0.4
0.2
0.0

350 400 450 500 550 600
wavelength (nm)

(c)

normalized intensity (a.u.)

NO2
1.0
0.8

excitation

emission

0.6
0.4
0.2
0.0
250 300 350 400 450 500 550 600
wavelength (nm)

Figure 1.2. Structures of chromophores used in this project and excitation and
emission spectra. (a) Free pyrene for the steady-state behavior of phospholipid
vesicles, (b) free perylene for dynamics measurements of the acyl chain region of
phospholipid vesicles, and (c) nitro-2-1,3-benzoxadiazol-4-yl tethered to the
headgroup of phosphoethanolamine lipids for studying the effects of headgroup
hydrogen bonding on lipid vesicle dynamics.
10

be preferentially excited.

This non-random population of excited molecules then

produces a fluorescence emission transient with polarization determined by the
distribution of emission transition dipole moments. The fluorescence emission from the
excited chromophores decays over the lifetime of the excited state as the polarizations of
the transition dipole moments re-randomize or depolarize. The total intensity of light, I,
emitted from a random distribution of chromophore molecules is the sum of its Cartesian
components:

I  Ix  I y  Iz

(1.1)

For a vertically (z-axis) polarized excitation pulse propagating along the x-axis, the
components of the fluorescence intensity oriented parallel and perpendicular to the
incident beam are defined as I z  I

and I x  I y  I  .

Therefore, observing the

polarized transients, Iy and Ix, along the z-axis as a function of time yields a total intensity
defined by Equation 1.2.

I (t )  I (t )  2 I  (t )

(1.2)

For fluorescence anisotropy measurements, the polarized emission transients from the
sample, I (t ) and I  (t ) , are combined to produce the induced orientational anisotropy
function, R(t):

R(t ) 

I (t )  I  (t )

(1.3)

I (t )  2 I  (t )

11

where the denominator represents the total intensity of the system (Fig. 1.3, Eq. 1.2).
Chromophores that are in a completely unrestricted environment, such as a continuum
solution, are able to access every directional orientation as they re-randomize over the
lifetime of the fluorescence. So, as the unhindered excited chromophores emit and rerandomize, the anisotropy decays over time to zero, at which point the orientations of the
molecules are randomly oriented. The amount of time measured from the point when a
non-random orientational distribution of excited state chromophores first begins emitting
fluorescence, until fluorescently decaying to the ground state and re-randomizing is
called the reorientation time constant (  OR ). This value provides information about the
motion of the chromophore while re-randomizing, as well as the ―restrictiveness‖ of the
local environment of the probe.

36

If the anisotropy function is determined according to

Eq. 1.3, and plotted as a function of time (Figure 1.3b), it produces an exponential decay
described by:

R(t )  R(0) exp(t /  OR )

(1.4)

where R(0) is the initial orientational anisotropy. For one photon excitation, R(0) usually
ranges in value from -0.2 to 0.4, depending on the angle between the absorption and
emission transition moments.

The reorientation time constant for spherical solute

molecules in a continuous solvent is related to the rotational diffusion constant, D, and to
the local viscosity, η, of the surrounding environment according to the Debye-StokesEinstein equation:

 OR 

37-39

1
Vf

6 D k BTS

(1.5)
12

a

1800
1600

counts

1400
1200
1000
800
600
400
200
0
0

1000 2000 3000 4000 5000 6000 7000 8000 9000

time (ps)
0.15

b

R(t)

0.10

0.05

0.00

-0.05
0

1000 2000 3000 4000 5000 6000 7000 8000 9000

time (ps)
Figure 1.3. (a) Time-resolved emission intensities for polarizations collected parallel
( I (t ) ) and perpendicular ( I  (t ) ) to the vertically polarized excitation pulse. (b) The
functional form of the orientational anisotropy decay obtained from transients in (a).
Residuals are distributed along the y-intercept and the solid line is the fit of the data.

13

1

where V is the hydrodynamic volume of the chromophore, f is a frictional term to
compensate for the interactions between the solute and solvent, kB is the Boltzmann
constant, T is the absolute temperature, and S is a shape factor that accounts for the shape
of the chromophore. Therefore, time-resolved fluorescence anisotropy measurements of
chromophores incorporated into either the acyl chain or headgroup region of the lipid
bilayer provide detailed information on the molecular motion of the probe molecules in
their local environment.
Typically, fluorescent probes sweep out a non-spherical volume while reorienting,
and can therefore be thought of as ellipsoidal rotors. The recovered anisotropy decay
describes the shape of the volume swept out by the reorienting chromophore, and is
related to the Cartesian components of the diffusion constant (Dx, Dy, and Dz), according
to a treatment by Chuang and Eisenthal,

40

where

1
D  ( Dx  Dy  Dz )
3

(1.6)

Theoretically the time-resolved anisotropy can produce up to five exponential decays, but
typically no more than two exponentials are observed experimentally. According to
Chuang and Eisenthal, the ellipsoid shape of the reorienting probe depends on the
orientation of its absorbing and emitting transition moments. Assuming that both the
absorption and emission transition moments are aligned along the same axis, in the πplane system of the chromophore, two assumptions can be made. The first is that the
molecule is a prolate rotor rotating predominantly about its long in-plane axis, where

14

Dx > Dy = Dz, and the recovered orientational anisotropy function takes the form of a
single-exponential decay:

R(t )  0.4exp(6Dz t )

(1.7)

The second assumption is that the chromophore is behaving as an oblate rotor, with its
dominant rotation about the out-of-plane axis, perpendicular to the chromophore’s πplane, where Dz > Dx = Dy, and the anisotropy function produces a two-component
exponential decay:

R(t )  0.1exp((2Dx  4Dz )t )  0.3exp(6 Dxt )

(1.8)

Another set of empirical expressions for prolate and oblate rotors is produced if it is
assumed that the absorption and emission transition moments are polarized along the
chromophore short in-plane axis
Prolate: R(t )  0.1exp((4Dx  2Dz )t )  0.3exp(6 Dxt )

(1.9)

Oblate: R(t )  0.4exp((4Dx  2Dz )t )

(1.10)

Fitting the experimentally-obtained anisotropy decay function according to assumptions
regarding the orientation of the absorption and emission transition moment of the
chromophore provides information on the diffusion constant of the molecule based on the
assumption of a rotor shape (Eqs. 1.7 – 1.10). Then using the Debye-Stokes-Einstein
equation (Eq. 1.5), the viscosity of the local environment of the chromophore can be
determined, and conclusions can be made as to the dynamic behavior and fluidity of the
local environment.
15

In the course of this project, we have used three different chromophore molecules
to investigate the molecular motion and organization of lipid bilayer systems in the form
of unilamellar vesicles. In Chapter 2 we used a nitro-2-1,3-benzoxadiazol-4-yl (NBD)
(Figure1.2c) chromophore tethered to the headgroup of a phosphoethanolamine lipid
(18:0 NBD-PE) to investigate the hydrogen bonding interactions between the headgroup
moieties of the NBD-PE lipids and systems consisting of either phosphocholine (18:0
PC) or phosophoglycerol (18:0 PG) lipids. In Chapter 3 we discuss using the polycyclic
aromatic hydrocarbon (PAH) pyrene (Figure 1.2a) to investigate the lateral diffusion and
fluid properties of acyl chain regions of 14:0 PC lipid vesicles. In Chapter 4 we used the
PAH perylene (Figure 1.2b) to investigate the organizational and fluid properties of the
acyl chain regions of 14:0 PC lipid vesicles. In Chapters 5 and 6, we again used perylene
to study the effects of thermal energy dissipation on the motional dynamics of 14:0 PC
lipid vesicle acyl chain regions.

The conclusions of all of these experiments are

discussed in Chapter 7.

16

CHAPTER 2
Rotational and Translational Dynamics of Headgroup-Tethered
Chromophores in Lipid Vesicles

Introduction
A vital concept in the regulation of biological cell life is membrane fluidity,
which encompasses the translational, rotational and trans-leaflet diffusion of constituent
lipids,

41

and influences cellular functions such as signaling, protein activity and response

to stimuli.

42

In order to study the molecular-scale behavior of cell membranes, simple

model membrane systems composed of pure lipids can be employed. Some typical
model lipid bilayer systems which can be used to mimic the physical properties and
dynamics of cell membranes

43-44

substrate-supported lipid bilayers,
47-48

cushions.

46

are unilamellar vesicles in solution,

45

solid or

and lipid bilayers or monolayers tethered to polymer

Changes in membrane fluidity also affect the permeability of the

membrane to nutrients and waste by altering the rate of transbilayer diffusion. For
example, the rate of transbilayer diffusion of an important signaling molecule that
transmits information between mammalian cells, nitric oxide, has been shown to decrease
with decreasing membrane fluidity.

49

One way in which biological cells control fluidity is by adjusting the composition
of membrane constituents. For example, incorporation of lysophosphatidylcholine lipids
into egg-phosphatidylcholine solid-supported lipid bilayers increases the rate of
17

translational diffusion by disrupting the van der Waals interactions between hydrophobic
acyl chains, making the bilayers more fluid. On the other hand, the addition of eggphosphatidylethanolamine into solid-supported egg-phosphatidylcholine lipid bilayers
results in pronounced headgroup hydrogen bonding, which enhances van der Waals
interactions due to tighter packing and lowers the rate of translational diffusion.

50

In the present study, we have investigated the effects of phospholipid headgroup
hydrogen bonding on the translational and rotational dynamics of 1,2-dioleoyl-snglycero-3-phosphoethanolamine-N-7-nitro-2-1,3-benzoxadiazol-4-yl (18:1 NBD-PE, Fig.
2.1) in lipid bilayers comprised of 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC)
and 1,2-dioleoyl-sn-glycero-3-[phospho-rac-(1-glycerol)] (DOPG).
structures of these lipids are shown in Figure 2.2.

The molecular

Although zwitterionic

phosphatidylcholine (PC) lipids such as DOPC comprise the majority of lipids found in
biological membranes, anionic lipids such as DOPG also play significant roles in
regulating membrane function.

51

Phosphatidylglycerol (PG) lipids are believed to be

responsible for maintaining membrane lipid surface charge density, a property that not
only affects membrane permeability to ions and charged metabolites, but also the activity
of certain membrane proteins.

52

Based on the structure of PC lipid bilayers, it is theorized that hydrogen bonding
is possible between the oxygen atoms in the phosphate functionalities of neighboring
lipid headgroups, and that there are electrostatic interactions between the negatively

18

100000
emission

excitation

intensity (a.u.)

80000

60000

40000

20000

0
250

300

350

400

450

500

550

600

wavelength (nm)
Figure

2.1.

Excitation

and

emission

spectra

phosphatidylethanolamine (NBD-PE) in chloroform.

19

of

1e-6

M

NBD-

NO2
N
O
N

HO

N+

NH
HO

O

H
O

O

O

O

DOPC

+
O NH4
O P O
O

+
O Na
O P O
O

O
O P O
O

H
O

O
O

O

DOPG

O

H
O
O

NBD-PE

Figure 2.2.
Molecular structures of lipids used in this study:
dioleoylphosphatidylcholine (DOPC), dioleoylphosphatidylglycerol (DOPG), and the
Greiner et al. Figure 1
chromophore NBD-phosphatidylethanolamine (NBD-PE).

20

charged phosphate oxygen and the positively charged choline nitrogen.

53-56

PG lipids,

in contrast, are rich in headgroup functionalities that can form hydrogen bonds with other
membrane constituents,

57,58

and they are also capable of forming hydrogen bonds

between the glycerol moiety and phosphate oxygen of neighboring PG lipids.

59-60

presence of hydrogen bonds can be measured directly using infrared spectroscopy

The
61

or

indirectly using fluorescence spectroscopy to monitor the dynamics of fluorescentlytagged lipid headgroups in lipid bilayers. In this study, we investigated interactions
between NBD-PE and headgroup regions of neighboring lipids in both vesicle and solidsupported lipid bilayer systems. We measured translational diffusion coefficients and
mobile fractions using fluorescence recovery after pattern photobleaching (FRAPP), and
fluorescence lifetimes and rotational diffusion times using time-correlated single photon
counting (TCSPC).
In order to the investigate lipid headgroup-headgroup interactions we chose the
fluorescently-tagged lipid NBD-PE, which has an NBD chromophore tethered to its polar
headgroup. In addition, two bulk lipids were used to form the lipid bilayer system,
DOPC and DOPG, which have identical hydrophobic acyl chain structures, but differ in
headgroup functionality. Therefore, using NBD-PE as the molecular probe of our lipid
systems allowed us to examine molecular-scale issues related to lipid headgroup
interactions while maintaining comparatively constant acyl chain interactions. We must
note that we are able to form intact lipid vesicles comprised of 98% of the anionic lipid
DOPG, however no direct comparison can be made in supported lipid structures because
only a few concentrations of anionic lipids such as DOPG can produce fluid bilayer
21

structures.

62

Regardless of this limitation, it is possible to compare mixed lipid systems

in both structural formats, and these data provide useful insight into the dynamical
properties of the NBD-PE contained within them.

22

Experimental Section
Vesicle Preparation:

Phospholipids 1,2-dioleoyl-sn-glycero-3-phosphocholine

(DOPC), 1,2-Dioleoyl-sn-Glycero-3-[Phospho-rac-(1-glycerol)] (DOPG), and 1,2Dioleoyl-sn Glycero-3 Phosphoethanolamine-N-7-nitro-2-1,3-benzoxadiazol-4-yl (18:1
NBD-PE) were purchased from Avanti Polar Lipids Inc. (Alabaster, AL) and used
without further purification. For fluorescence lifetime and anisotropy measurements 1
mM lipid solutions were made consisting of: 1) 98 mol% DOPC and 2 mol% of NBDPE, 2) 98 mol% DOPG and 2 mol% NBD-PE, and 3) 59 mol% DOPC, 39 mol% DOPG,
and 2 mol% NBD-PE were prepared. For FRAPP measurements 1 mM lipid solutions
were made consisting of: 1) 98 mol% DOPC and 2 mol% NBD-PE, and 2) 59 mol%
DOPC, 39 mol% DOPG, and 2 mol% NBD-PE. The chloroform solvent was then
evaporated from the lipid mixture and the dry lipid film was hydrated with Tris®-HCl
buffer (Sigma-Aldrich, St. Louis, MO) containing NaCl (FisherScientific, Fair Lawn, NJ)
for 30 min. prior to extrusion. The buffer (10 mM Tris®, 150 mM NaCl, pH 8) was
prepared using water from a Milli-Q Plus water purification system (Millipore, Bedford,
MA). The mixtures were processed five times through a freeze-thaw-vortex cycle to
ensure complete mixing of the sample constituents. Each cycle consisted of immersion in
liquid nitrogen for 5 min., followed by immersion in a 60C water bath for 5 min., and
then vortexing the thawed sample for 2 min. The hydrated mixtures were extruded 11
times through a polycarbonate membrane with a nominal pore diameter of 100 nm
(Avanti Polar Lipids Inc., Alabaster, AL).
temperature.

23

All extrusions were performed at room

Time-Correlated Single Photon Counting (TCSPC): Time-resolved fluorescence
anisotropy and lifetime data were measured using a TCSPC instrument.
experimental setup has been described in detail elsewhere,

63-64

The

so only the essential

features are given here. The light source is a CW mode-locked Nd:YAG laser (Coherent
Antares 76-s, Santa Clara, CA) that produces 100 ps 1064 nm pulses at a 76 MHz
repetition rate. The third harmonic of the Nd:YAG laser output was used to excite a
cavity-dumped dye laser (Coherent 702-2, Santa Clara, CA) operated with Stilbene 420
dye (Exciton, Dayton, OH) at 460 nm. The average output power of this laser was ca. 25
mW, with 5 ps pulses at a 4 MHz repetition rate. The sample emission was detected with
a microchannel plate photomultiplier tube (MCP-PMT, Hamamatsu R3809U-50,
Bridgewater, NJ). The electronics used to temporally resolve the fluorescence transients
included a constant fraction discriminator (CFD, Tennelec 454) and a time-to-amplitude
converter/biased amplifier (TAC, Tennelec 864). The collection wavelength (545 nm, 10
nm detection bandwidth) and polarization were computer controlled using LabVIEW 7.1
code. The fluorescence lifetime data were collected at 54.7with respect to the vertical
excitation polarization, while the reorientation data were collected at polarizations
parallel (0) and perpendicular (90) to the vertically polarized incident light pulse.
Deposition of Supported Lipid Bilayers: Fused silica microscope slides (75 mm x
25 mm x 1 mm) were purchased from Technical Glass Products, Inc. (Painesville, OH).
The slides were cleaned by bath sonication (Branson 1510, Branson Ultrasonic
Corporation, Danbury, CT) in detergent solution for 20 min., rinsed with DI water, baked
at 160C for 4 h, and plasma treated (Harrick Plasma, Ithaca, NY) with oxygen under
vacuum (150 mTorr) for 10 min. immediately before bilayer deposition. Supported
24

bilayer lipid membranes were deposited by vesicle fusion in a custom-made flow cell as
described elsewhere.

65

The flow cell was initially washed with buffer, followed by a 3-

hour incubation with liposome solution and a final buffer wash to remove unadsorbed
vesicles. All experiments were performed at room temperature.
Fluorescence Recovery After Pattern Photobleaching (FRAPP): FRAPP was
used to measure the rate of translational diffusion of chromophores in supported bilayer
membranes formed on silica microscope slides.
described in detail elsewhere.

65

The experimental setup has been

A double syringe pump system (Harvard Apparatus)

was used to infuse and withdraw solutions simultaneously from a custom-made 1 mL
flow cell. The 488 nm laser line from an argon ion laser (95, Lexel Laser, Fremont, CA)
was directed through a 5x beam expander (Edmund Optics, Inc., Barrington, NJ), a filter
cube (Ex: 450-490/DM: 510/Em: 515-565), and a pair of optical flats65 that toggle
between a low intensity (20µW) monitoring beam for continuous fluorescence emission
measurements, and a high intensity (500 mW) beam for photobleaching chromophores.
Light passing through the sample was collected by an inverted microscope (Zeiss
Axiovert 135M, Carl Zeiss, Thornwood, NY) with a 32x objective lens (Carl Zeiss,
Thornwood, NY). The light intensity was measured by a side-on photomultiplier tube
(PMT, Hamamatsu R4632, Bridgewater, NJ) connected to a photon counter (SR400
Stanford Research Systems, Sunnyvale, CA) and a fast preamplifier (SR445 Stanford
Research Systems, Sunnyvale, CA). A Ronchi ruling (50 lines per inch, Edmund Optics,
Inc., Barrington, NJ) in the back image plane of the microscope created a fringe pattern
on the sample. An aperture placed in the image plane in front of the PMT restricted the

25

observation area, resulting in an illuminated area of approximately 200 µm in diameter
and an observed area of 75 µm. This was done to prevent unbleached fluorophores from
diffusing from outside the bleached (patterned) area into the observation zone during
fluorescence recovery measurements. Stripe periodicity in the sample plane was 25μm.

26

Results and Discussion
The goal of this work was to investigate the interactions between neighboring
lipid headgroups in a bilayer structure by examining the rotational and translational
diffusion behavior of a tethered chromophore in lipid vesicles and supported lipid
bilayers. DOPC and DOPG were chosen as the primary lipids in this study for two
reasons. First, these lipids have low characteristic phase transition temperatures (Tm) of
253K for DOPC

66

and 258K for DOPG,

60

ensuring that the bilayers will exist in a

disordered, fluid phase at ambient temperature. Second, the two lipids have identical acyl
chain lengths but different headgroup functionalities, so differences in chromophore
dynamics can be attributed entirely to headgroup interactions. We consider the rotational
diffusion and fluorescence lifetime data first.
Fluorescence Lifetime and Anisotropy Data.

Fluorescence lifetime and

anisotropy decay data provide information on the molecular environment(s) of the NBD
chromophore in vesicles of the following compositions: 1) 98 mol% DOPC and 2 mol%
NBD-PE, 2) 98 mol% DOPG and 2 mol% NBD-PE, and 3) 59 mol% DOPC, 39 mol%
DOPG, and 2 mol% NBD-PE. Additionally, we examined vesicles 100 nm and 2000 nm
in diameter to assess the influence of membrane curvature on rotational dynamics. For
all measurements, there were no discernable differences in either the lifetime or the
anisotropy data for the two vesicle sizes, suggesting that curvature is not an issue for
these measurements.
A two-component exponential decay was observed for the fluorescence lifetime
data for NBD-PE in each vesicle system (Figure 2.3a, Table 2.1). There are two possible
27

3500

a
a

3000

counts

2500
2000
1500
1000
500
0
0
0.40

2500

5000

7500

10000

12500

15000

time (ps)

bb

0.35
0.30

r(t)
R(t)

0.25
0.20
0.15
0.10
0.05
0.00
-0.05
0

2500

5000

7500

10000

12500

15000

time (ps)

Greiner et al. Figure 2

Figure 2.3. (a) Fluorescence lifetime and (b) induced orientational anisotropy data
for NBD-PE in a DOPG/DOPC mixed lipid vesicle. The residuals for a fitted twocomponent exponential decay function are shown for both panels.
28

Table 2.1. Fluorescence lifetimes (τ1 and τ2) and reorientation times (τOR1 and τOR2) of
NBD-PE in bilayer lipid vesicles, determined by TCSPC.

τ1 (ns)

τ2 (ns)

τOR1 (ns)

τOR2 (ns)

DOPC

1.92 ± 0.12

7.83 ± 0.21

0.65 ± 0.05

3.67 ± 0.10

DOPG

2.13 ± 0.07

7.70 ± 0.11

0.49 ± 0.20

3.20 ± 0.52

DOPC/DOPG

1.89 ± 0.19

7.48 ± 0.14

0.77 ± 0.52

3.93 ± 0.05

29

explanations for this finding.

Either the chromophore is sensing two unique

environments with populations that do not exchange significantly on the timescale of the
excited-state relaxation, or the chromophore exists in a dielectric gradient, where slight
changes in the location and/or orientation of the chromophore with respect to the gradient
give rise to a distribution of fluorescence lifetimes.
reported previously for the NBD chromophore.

71-72

67-70

This condition has been

While it may be tempting to ascribe

the different lifetime components to specific chemical environments, the dependence of
fluorescence lifetime on intramolecular and intermolecular properties is sufficiently
complex that it is not possible to make such assignments in the absence of additional
information.
Fluorescence anisotropy decay data, in contrast, can be interpreted within the
context of well established theoretical models.

37, 73-75

The induced orientational

anisotropy decay function, R(t), is formulated as the normalized transient difference in
emission intensity parallel, I||(t), and perpendicular, I(t), to the polarization of the
incident excitation pulse (Eq. 2.1):

R(t ) 

I (t )  I  (t )

(2.1)

I (t )  2 I  (t )

This is the functional form of R(t) that contains chemically relevant information. There
are two possible choices for the interpretation of our fluorescence anisotropy data due to
the fact that the chromophore is attached to the headgroup of the phospholipid. The

30

functional form of the anisotropy decay allows us to determine whether this tetheredchromophore behaves as a free rotor or as a hindered rotor.
A free rotor, such as a chromophore in free solution, is characterized by a zero
infinite-time anisotropy because the chromophore can assume all orientations with equal
probability. The general case for a chromophore tethered to a lipid (as is the case for
NBD-PE) is termed a ―hindered rotor‖, because it is not free to access all possible
74-75

orientations, resulting in a non-zero infinite-time anisotropy, R(∞).

The limiting

case for a hindered rotor is where the volume accessible to the chromophore is
hemispherical, and under this condition it will behave as a free rotor because of the
inherent polarization symmetry of the incident electric field. Our experimental data are
characterized by a two-component exponential anisotropy decay for each of the three
lipid systems used in this study, and for each the infinite-time anisotropy, R(∞) is zero to
within the experimental uncertainty (Figure 2.3b, Table 2.1). The absence of an infinitetime anisotropy component in the data means that the chromophore is able to reorient
within a hemispherical volume and, as such, it can be treated as a free rotor.
We can interpret our anisotropy decay data using a model for molecular
reorientation developed by Chuang and Eisenthal.

73

This model was developed for free

chromophores with full motional freedom in solution, but the same essential physics
underlying the Chuang and Eisenthal model must also apply to the dynamics of a tethered
chromophore, especially when the chromophore is free to relax in a hemispherical
volume (i.e. no confining cone). In this model, the Cartesian axes are assigned so that the
z-axis is perpendicular to the chromophore π-system plane and the exciting and emitting
31

transition dipole moments are oriented along either the x-axis (long axis) or y-axis (short
axis) of the chromophore molecular plane.

76

The two limiting cases of this model treat

the chromophore as either a Type I or a Type II rotor. The excited and emitting transition
dipoles lie coincident with the unique rotational axis for Type I (Dx ≠ Dy = Dz ) and
perpendicular to the unique rotational axis for Type II (Dx = Dy ≠ Dz) rotors. The
reorientation decay time constants are related to the Cartesian components of the
rotational diffusion constant through the anisotropy decay function, R(t), as shown in
Eqs. 2.2-2.5:

77

Transition dipole oriented along the chromophore long (x) axis
 4
r (t )    exp  6 Dz t 
 10 

Type I:

Type II:

(2.2)

1
 3
r (t )    exp    2 Dx  4 Dz  t     exp  6 Dxt 
 10 
 10 

(2.3)

Transition dipole oriented along the chromophore short (y) axis
1
 3
r (t )    exp    2 Dz  4 Dx  t     exp  6 Dz t 
 10 
 10 

Type I:

Type II:

 4
r (t )    exp    2 Dz  4 Dx  t 
 10 

(2.4)

(2.5)

In this model, the recovery of a double-exponential decay indicates that the NBD
chromophore behaves as either an x-axis polarized Type II rotor or a y-axis polarized
Type I rotor, according to Eqs. 2.3 and 2.4. Previous studies of the solution phase NBD
32

chromophore using one and two-photon excitation methods suggest that the S1 ← S0
transition of this chromophore is short-axis polarized, and a two-component anisotropy
decay would be consistent with NBD reorienting as a Type I rotor (Eq. 2.4).

57

Based on

this assignment, we evaluate the Cartesian components of the rotational diffusion
constant for NBD-PE.
The Cartesian components of the diffusion constant (Table 2.2) of tethered NBD
in all of the vesicle solutions are similar. These data all indicate that the tethered NBD
chromophore is a substantially anisotropic rotor, with an aspect ratio of Dx/Dz ~ 8. While
it may be tempting to consider the subtle differences between the Dz and Dx values and
the ratio of these constants, the uncertainty in the data is sufficiently large enough that it
is not possible to draw any system-dependent conclusions without additional information.
Next we will consider the molecular-scale translational diffusion behavior of NBD-PE in
our supported lipid bilayer systems.
Translational Diffusion in Supported Lipid Bilayers:

Rotational diffusion

measurements can provide information on the molecular-scale environment of the
reorienting chromophore, and the information we obtain for NBD-PE on the molecular
length scale suggests that the dynamics of the lipid headgroup are determined to a
significant extent by the interactions of the aqueous environment with the lipid bilayer.
This finding is consistent with our observation of a two-component fluorescence lifetime,
which can be accounted for in the context of the chromophore existing in a dielectric
gradient. The molecular-scale information does not provide a complete picture, however,

33

Table 2.2. Cartesian components of the rotational diffusion constant (MHz) of NBD-PE
chromophores in lipid bilayer vesicles.

Dx (MHz)

Dz (MHz)

Dx/Dz

DOPC

362 ± 27

45 ± 2

7.97 ± 0.82

DOPG

484 ± 148

52 ± 7

9.30 ± 4.22

DOPC/DOPG

303 ± 131

42 ± 1

7.16 ± 3.18

34

of the fluid properties of the lipid bilayers. We also need to consider motion over longer
length scales, and we accomplish this task through the use of translational diffusion
measurements.
Mobile fractions and translational diffusion coefficients were measured with
FRAPP to provide information complementary to our TCSPC fluorescence lifetime and
anisotropy decay data. Because vesicles with 98 mol% DOPG did not adsorb to fused
silica, FRAPP studies were limited to two systems: 1) 98 mol% DOPC and 2 mol%
NBD-PE, and 2) 59 mol% DOPC, 39 mol% DOPG, and 2 mol% NBD-PE. We found
that 39 mol% DOPG was the maximum concentration of the anionic lipid that would lead
to the establishment of stable fluid solid-supported lipid bilayers.
consistent with previous studies under similar buffer conditions.

This finding is

46, 78

For FRAPP experiments, a Ronchi ruling was placed in the rear image plane of
the microscope, creating a striped pattern of alternating light and dark fringes on the
solid-supported lipid bilayer. Chromophores in illuminated zones are photobleached by
exposure to a high intensity laser beam. Unbleached chromophores in the dark zones
then diffuse into the photobleached areas, leading to the recovery of a spatially
homogeneous fluorescence signal.

79

The time constant and functional form of this

recovery are given by,

f (t )  f (0) 

  8 
 4 2 Dt  1
 36 2 Dt  
m


  exp  

1  f (0) 1  
exp  

2
2  9
2 


2
a
a





  


35

(2.6)

where f(t) is the post-bleach fluorescence intensity (t > 0) normalized with respect to the
constant pre-bleach fluorescence intensity (t < 0), with t = 0 being the time of the bleach
pulse, D is the average translational diffusion coefficient of the chromophores within the
bleached area, m is the mobile fraction, and a is the stripe periodicity of the Ronchi ruling
in the sample plane.

80

Eq. 2.6 assumes a single diffusing population, corresponding to

the chromophores in the solid-supported lipid bilayer for our experiments, and an
immobile population of chromophores that reside in intact vesicles attached to the
surface.
A typical FRAPP recovery curve for a DOPC/NBD-PE bilayer is shown in Figure
2.4, and calculated translational diffusion coefficients and mobile fractions are given in
Table 2.3. The mobile fractions were determined to be statistically the same for both
systems using an ANOVA analysis; however, the translational diffusion coefficients were
statistically different. For the DOPC bilayer we measured a diffusion coefficient of 2.91
2

2

± 0.31 µm /s, and for the mixed DOPC/DOPG bilayers we measured 1.82 ± 0.41 µm /s.
It is clear from these data that translational diffusion of NBD-PE is hindered in bilayers
containing DOPG. The fact that this difference in translational diffusion times was
observed within the experimental uncertainty for the DOPG-containing bilayer, but not in
the rotational diffusion measurements suggests that the NBD-PE chromophore does not
show a substantial preference for interacting with either DOPC or DOPG lipid
headgroups. The intermolecular headgroup-headgroup interactions between lipids seem
to mediate the translational motion of the bilayer constituents. This finding is consistent
with recent molecular dynamics (MD) simulations involving PC and PG bilayers.

36

81

The

normalized fluorescence intensity

0.6
0.5
0.4
0.3
0.2
0.1
0.0
-5

0

5

10

15

20

25

30

35

40

time (s)

Figure 2.4. A typical FRAPP recovery curve for a DOPC/NBD-PE supported bilayer
membrane (above) along with residuals (below). Total photobleaching time was 300
ms.

37

Table 2.3. Mobilities and diffusion coefficients of NBD-PE in supported lipid bilayers
on fused silica. FRAPP data represent average values for three bilayers. At least five
different areas were interrogated for each bilayer.

Mobility (%)

Diffusion Coefficient (μm2/s)

DOPC

0.79 ± 0.11

2.91 ± 0.31

DOPC/DOPG

0.61 ± 0.08

1.82 ± 0.24

38

strength of the hydrogen bonding interactions ranges from ca. 10 to ca. 30 kJ/mol.

82

The

NBD chromophore contains several hydrogen bonding functionalities so, in principle it
can participate in hydrogen bonding interactions. Perhaps of more significance is the fact
that the DOPG glycerol headgroup contains two hydroxyl groups that can function as
either hydrogen bond donors or acceptors. In contrast, the DOPC headgroup does not
contain any hydrogen bonding functionalities and, as such, should not be expected to
participate in headgroup-headgroup interactions in a manner analogous to that of DOPG
headgroups.

MD

simulations

have

reported

that

pure

palmitoyloleoylphosphatidylglycerol (POPG) bilayers are ca. 6 Ǻ thicker than pure
palmitoyloleoyl-phosphatidylcholine

(POPC)

bilayers.

Therefore,

the

slower

translational diffusion coefficient observed for the DOPG system could also be attributed
its having a thicker bilayer membrane.
We next need to consider how these two bodies of data can be reconciled. The
rotational diffusion data indicate that at the molecular level, the rotational motion of
chromophores localized in the headgroup region of the bilayer is not mediated
significantly by the headgroup structure or organization of the lipids. This situation is
explained by the presence of bulk Tris® buffer solvent in close contact with the
interfacial region of the lipid bilayer structure. The fact that the translational diffusion
behavior of these chromophores in the bilayers does depend on lipid headgroup identity
arises from the fact that it is the translational motion of entire lipid molecules that is
being sensed in the measurement, and translational motion is expected to be significantly
dependent on structural heterogeneity present in the bilayer.

39

In our translational diffusion data there appear to be regions of the supported
bilayer structure that are comparatively less mobile than others. The notion of immobile
regions is not surprising for the DOPC/DOPG systems in light of the large body of
literature which proposes the existence of phase separation in mixed lipid systems. The
basis for such phase separation is understood from a thermodynamic point of view for
multi-component systems,

83-92

the origin of such features.

but for single-component systems, it is fair to question
One possibility is that interactions exist between the

supported lipid bilayer lower leaflet and the substrate interface, which result in
comparatively less mobile regions. Experiments to resolve the physical reason or reasons
for these results is a presently under investigation.

40

Conclusions
We have investigated the dynamics of a tethered chromophore (NBD-PE) in
unilamellar DOPC, DOPG, and mixed DOPC/DOPG vesicles and supported bilayer
systems. We used both phosphocholine (PC) and phosphoglycerol (PG) to evaluate the
role of headgroup hydrogen bonding interactions in mediating the fluidity of lipid bilayer
structures. Within the experimental uncertainty we obtained the same rotational diffusion
results for tethered NBD in each system for vesicles of both 100 nm and 2000 nm
diameter. Our data also showed that the molecular-scale rotational motion of the NBD
chromophore does not depend sensitively on the type of lipid in the bilayer. We also
measured the translational diffusion of NBD-PE in planar supported lipid bilayers
comprised of pure DOPC and a mixture of DOPC and DOPG. The data indicate a slower
rate of translational diffusion in supported lipid bilayers containing DOPG. We believe
this is due to hydrogen bonding interactions between the headgroup regions of
neighboring lipids, even though the effect on rotational diffusion appears to be much less
pronounced.

41

CHAPTER 3
Investigating Membrane Fluidity of Phospholipid Bilayers Using
Steady-State Fluorescence Measurements

Introduction
Membrane fluidity is an extremely complex and dynamic concept that
encompasses a number of molecular motions, including lipid flip-flop, rotational
diffusion, and translational diffusion.

23

All lipid motions contribute to membrane

fluidity and play an essential role in regulating the life of a cell. Successful membrane
protein reconstitution and functionality are largely dependent upon the behavior of lipid
bilayers at a molecular level. Therefore, in order to develop a working membrane-based
biosensor, the behavior and fluidity of lipids in bilayer vesicle systems needs to be
thoroughly understood before proceeding.
Several biological cell processes are mediated by the composition and properties
of the cell membrane, some of these include transmembrane protein function,
intercellular signaling, and enzyme activity.

42

Biological cell membranes can regulate

protein function by manipulating the fluidity of the lipid bilayer to control the diffusive
properties of the membrane. Translational diffusion of molecules through the membrane
can be mediated in several ways, including altering the phase of the lipids, adding lipids
with varied chain lengths or unsaturations, or by introducing an ―impurity‖ to the bilayer.
It is possible to vary the phase of the lipid bilayer by altering the system temperature of
the lipid environment.

All lipids have a characteristic gel-to-fluid phase transition

temperature (Tm) that is distinguished by an ordered gel phase at temperatures below Tm,
42

and a less ordered fluid phase at temperatures above Tm. Gel phase lipids have tightly
packed acyl chain regions, resulting in limited rotational and translational freedom of the
lipids. Fluid phase bilayers have disordered acyl chain regions with increased motional
freedom, resulting in more ―fluid-like‖ behavior.
Additionally, the introduction of a bilayer impurity or non-lipid constituent, such
as the weak anesthetic ethanol, is thought to disrupt the well ordered structure of the
bilayer lipids, which alters the overall fluidity of the membrane. Ethanol is thought to
displace water molecules that are associated with the gel phase lipid bilayer interfacial
region, destabilizing the interactions between the lipid headgroups, causing the bilayer to
dilate. This lipid dilation increases the average surface area of the headgroups, and once
a sufficient amount of ethanol is added to the system, it can perturb the bilayer to the
extent that it forms an intermediate phase known as the interdigitated phase.

93-95

The

interdigitated phase occurs when lipids in opposing leaflets interpenetrate with one
another, decreasing the overall bilayer thickness, and eliminating the bilayer midplane.
The formation of the interdigitated phase can affect the proper function of integral
proteins in biological cells by altering the fluidity of the lipid membrane.

96

Therefore, it

is of interest to us to study the effects of ethanol concentration on the fluidity of pure
lipid vesicles using a fluorescent molecular probe.
One way of investigating the membrane fluidity or translational mobility of lipids
is through the use of a fluorescent molecular probe imbedded in the acyl chain region of
the bilayer, which reports on the local environment of the probe. The acyl chain region is
the area of the bilayer that undergoes the most dramatic structural reorganization as a

43

result of the addition of non-lipid constituents and system temperature changes.
Therefore, the nonpolar region of the bilayer is of particular interest to us. The local
environment of a probe molecule can often directly determine its optical response.
Pyrene, a polycyclic aromatic hydrocarbon with a long excited-state lifetime and high
quantum yield, has been used extensively to investigate and gauge the local polarity of
solvents.

29-33, 97

In lipid mixtures, pyrene intercalates into the nonpolar acyl chain

region of the lipid bilayer, providing information on the local polarity of the nonpolar
region of lipid vesicles. Pyrene monomer has a characteristic fluorescence spectrum,
with five major vibronic bands appearing at 373, 379, 383, 389, and 393 nm
30, 32

(Fig. 3.1).

These vibronic bands are assigned values I through V, respectively. The

I/III band intensity ratio scales with increasing solvent polarity, and is a result of dipoleinduced dipole interactions between the solvent molecules and the vibrational modes of
pyrene.

29-33

Pyrene also possesses the ability to form an excited-state dimer structure known
as an excimer. The formation of an excimer, (MM)*, occurs when an electronically
excited molecule of pyrene (M*) collides and reacts with a ground-state monomer of
pyrene (M), and the process can be thought of as dynamic fluorescence quenching of the
pyrene monomers.

35

The reactions of interest for excimer formation are

35, 97

:

M  h1  M *

Excitation

(3.1)

M *  M  h 2

Fluorescence

(3.2)

M * M  (MM )*

Excimer Formation

(3.3)

44

(MM )*  2M  h 3

Excimer Fluorescence

(3.4)

The fluorescence spectrum of pyrene excimer is red-shifted from that of pyrene
monomer, and is characterized by the appearance of a broad, featureless emission band at
ca. 470 nm (Fig. 3.1). The formation of pyrene excimer is concentration-dependent, as
an increase in the concentration of pyrene monomer increases the probability of a
reaction between an excited-state pyrene molecule and a ground-state pyrene monomer
(Fig. 3.2). The ability of pyrene to form excimers when sequestered in the acyl chain
region of lipid bilayers also depends on the fluidity of the membrane, or the translational
mobility of pyrene molecules through the acyl chain medium. In this way, pyrene
fluorescence measurements provide data on the restrictiveness of the local environment
of the probe.

26

As more ethanol is added to solutions of pyrene-doped lipid vesicles, the

translational mobility of pyrene through the acyl chains becomes less restricted and the
propensity of pyrene to form excimers increases.

Therefore, by monitoring the

fluorescence emission of pyrene excimer in lipid vesicles as a function of ethanol
concentration, we are able to gauge the extent to which the incorporation of ethanol
mediates the fluid properties of the bilayer structure. We also want to study whether
ethanol perturbs the lipid bilayer to the extent that it can induce the formation of the
interdigitated phase.
For these experiments we use 100 nm diameter unilamellar vesicles comprised of
1,2-dimyristoyl-sn-phosphatidylcholine (14:0 DMPC, Tm = 297 K), and apply steadystate fluorescence techniques to interrogate the properties of pyrene incorporated into

45

normalized intensity (a.u.)

Pyrene

1.0

absorbance
emission

0.8
0.6
excimer band
0.4
0.2
0.0
300 325 350 375 400 425 450 475 500
wavelength (nm)

14:0 DMPC

Figure 3.1. Structures of 14:0 DMPC, pyrene, and absorbance and emission
spectra of 1e-5 M pyrene in chloroform.

46

excimer band

2.0

normalized intensity (a.u.)

3e-5 M pyrene
3e-6 M pyrene
1.5

1.0

0.5

0.0

320

340

360

380

400

420

440

460

480

500

520

wavelength (nm)

Figure 3.2. Emission spectra of pyrene demonstrating the relationship between
concentration and excimer band intensity in 3e-6 M pyrene in DMPC vesicles (dotted
line) and 3e-5 M pyrene in DMPC vesicles (solid line).

47

fluid phase (T = 298 K) vesicle bilayers. Typical biological cell membranes are primarily
composed of phospholipids with 16 to 18 carbons per acyl chain, but we use DMPC,
which has 14 carbons in each of its aliphatic tails. We chose DMPC because of its
characteristic phase transition temperature, which is just below room temperature. This
gel-to-fluid phase transition temperature (297 K) allows for easy manipulation of lipid
phase at ambient conditions. For steady-state fluorescence experiments we use pyrene
because of its aforementioned well-characterized polarity-dependent emission properties
and its propensity for excimer formation.

31, 34, 98-102

We found that at ethanol

concentrations of ca. 0.6 M there appeared to be an increase in the emission intensity of
pyrene excimer. This finding indicates that there is an increase in excimer formation at
this concentration of ethanol, which could be the result of a fluidizing, or reorganization,
of the lipid bilayer acyl chains by the penetration of ethanol molecules into the bilayer of
the vesicles.

48

Experimental Section
Vesicle Preparation. The phospholipid 1,2-dimyristoyl-sn-phosphatidylcholine
(14:0 DMPC) in chloroform was purchased from Avanti Polar Lipids, Inc. (Alabaster,
AL) and used as received. Pyrene (99% optical grade) and perylene (99+ %) were
purchased from Sigma-Aldrich (Milwaukee, WI) and used without further purification.
For each sample, 10 mg of lipid were mixed with 30.5 µg of chromophore, and the
chloroform solvent was evaporated. Tris

®

buffer (Sigma-Aldrich) was then added to

each sample to make the lipid concentration 1 mg/mL. Tris

®

buffer (10 mM, pH 8.0)

was prepared with purified water from a Milli-Q Plus water purification system
(Millipore, Bedford, MA), and was purged with argon prior to use. The lipid and
chromophore mixtures were processed five times through a freeze-thaw-vortex cycle to
ensure complete mixing of the constituents. Each cycle was comprised of first freezing
the solution by immersion in N2(l) (5 min.), then thawing in a 60°C water bath (5 min.),
followed by vortexing the mixture (2 min.). After completion of the freeze-thaw-vortex
process, sample solutions were extruded eleven times through a polycarbonate membrane
(Whatman, Piscataway, NJ) with a nominal pore diameter of 100 nm using a miniextruder apparatus (Avanti Polar Lipids, Inc.).

25, 103

Once the vesicle solutions were

extruded, controlled amounts of 95% ethanol (Sigma-Aldrich) were added to the bulk
mixtures, and the samples were allowed to equilibrate under ambient conditions prior to
fluorescence acquisition. A time study (results not shown) showed that the amount of
lipid-ethanol equilibration time needed to achieve reproducible fluorescence results was
ca. 24 hours.

49

Steady-State Fluorescence Measurements. Steady-state emission spectra were
acquired for three trials each of the DMPC/pyrene vesicles with ethanol held at a constant
temperature of 298 K. Samples were excited at 339 nm, and emission was collected from
350-500 nm using a Spex Fluorolog 3 emission spectrometer with 2 nm resolution for
both excitation and emission monochromators. The sample temperature was regulated to
± 0.1⁰C with a water-circulating bath (Neslab EX-100DD) connected to a temperaturecontrolled glass cell jacket which held the sample cuvette. All samples were allowed to
thermally equilibrate for 10 min. prior to data acquisition.

50

Results and Discussion
The goal of the work we report here is to gauge the effects of ethanol
concentration on the molecular-scale dynamics and fluidity of unilamellar lipid vesicles.
We have used pyrene imbedded in the acyl chain region of the vesicles for our steadystate emission measurements because of the probe’s sensitivity to polarity and its
propensity to form excimers. Pyrene excimer formation depends on the chromophore’s
ability to diffuse through the acyl chain environment. So, in other words, the more fluid
the local environment of pyrene, the higher the concentration and the more intense the
fluorescence emission band for pyrene excimer. These data provide insight into the
interdigitation process and suggest that interdigitation proceeds heterogeneously across
the bilayer surface, rather than in a coordinated manner.
Pyrene is well established as a polarity-sensitive fluorescent probe in lipid
membranes,

34, 98, 104

and because of its nonpolar nature, it partitions selectively into the

aliphatic region of lipid bilayers. The emission spectrum of pyrene consists of five
vibronic bands,

30

and the ratio of the I/III emission band intensities has been correlated

with the polarity of the chromophore immediate environment.

31

This somewhat unusual

spectral property is a consequence of the electronic structure of pyrene and is a
consequence of the differences in the nature of vibronic coupling for the I and III
transitions.

31

A I/III band intensity ratio > 1 is characteristic of pyrene being sequestered

in a polar environment, while a ratio < 1 is indicative of the chromophore being in a
nonpolar environment. The I/III pyrene emission band ratio (Py scale) remained constant
for all of our steady-state measurements with a value of 0.96, a value that suggests that
51

the chromophore remains in the nonpolar bilayer acyl chain region under all conditions
examined here. We note that the Py band ratio has a reported value of 1.18 for pyrene in
ethanol, 1.87 in water and 0.60 for hexadecane. The value we obtain is consonant with
solvents of polarity similar to p-xylene and iso-propyl ether.

105

Pyrene also has the ability to form excited-state dimers or excimers, which are
products from the reaction of electronically excited pyrene molecules with ground-state
pyrene molecules.

The characteristic signature of pyrene excimer formation is the

appearance of a broad structureless emission feature centered near 470 nm. Excimer
formation is limited by both concentration and translational diffusion through the
surrounding medium. The probability of excimer formation scales with the fluidity of the
bilayer structure, and we can use this spectroscopic feature as a gauge of local fluidity in
our work.
We are interested in understanding changes in the molecular-scale organization of
lipid bilayers as a function of ethanol concentration, and whether the subsequent
perturbation to the interfacial region of the bilayer can trigger the formation of the
interdigitated phase. It has been proposed before that lipid bilayer interdigitation results
in a subsequent decrease in membrane fluidity.

106-107

The rationale for this assertion is

that the presence of ethanol, which has been known to initiate lipid interdigitation, gives
rise to an increase in the average spacing between lipid headgroups.

108

Losses in lipid-

lipid headgroup interactions are compensated by interdigitation, which increases interleaflet acyl chain interactions. While this condition could lead to a less fluid membrane,
the interchain spacing in the interdigitated bilayer must remain substantially the same as
52

that of the non-interdigitated bilayer because of the intermolecular forces acting between
the acyl chains.

96, 109

Our steady-state fluorescence measurements can help to resolve

this issue.
Steady-state fluorescence emission data were acquired for pyrene incorporated
into 100 nm diameter DMPC vesicles as a function of solution ethanol concentration
(Fig. 3.3). There is an overall trend of increasing excimer band intensity with increasing
ethanol concentration, with a noticeable excimer intensity enhancement near 0.6 M
ethanol. We view the data presented in Figure 3.3 as showing an essentially continuous
change in the amount of excimer formed with increasing ethanol concentration, with an
anomalous increase near 0.6 M ethanol. We understand the continuous trend in the
context of increasing pyrene excimer content of the bilayer structure as the solution-phase
ethanol concentration increases. While there are no data we are aware of that relate
ethanol concentration to changes in lipid bilayer acyl chain viscosity, the dilation of the
acyl chains by ethanol could lead to increased translational mobility of constituents
within the bilayer. The anomalous increase in excimer formation in the region of 0.6 M
ethanol has been seen for other bilayer systems, albeit at different ethanol concentrations.
We believe that this anomalous feature represents the heterogeneous onset of
interdigitation, a phenomenon which should depend on the structural identity of the lipid.
If the solubility of pyrene is different in the interdigitated bilayer regions than in the noninterdigitated regions, then the formation of interdigitated areas would give rise to an
increase in pyrene concentration in the region of preferential solubility. Once the bilayer
structure became more structurally homogeneous, this concentration modulation effect
would no longer be operative, and the ethanol concentration-dependence would return
53

normalized excimer emission intensity

0.0085

0.0080

0.0075

0.0070

0.0065

0.0060

0.0055
0.0

0.5

1.0

1.5

2.0

ethanol concentration (M)

Figure 3.3. Pyrene excimer band intensity for the chromophore incorporated into 100
nm diameter DMPC vesicles, as a function of ethanol concentration. Error bars are
± 1σ for three individual determinations at each ethanol concentration. The dashed
line is a linear regression fit to the data.

54

to what is expected for a homogeneous bilayer-ethanol system.
A previous ethanol-induced interdigitation study conducted by Löbbecke et al.
found that for carboxyfluorescein-imbedded 1,2-dipalmitoyl-sn-glycero-3-phosophocholine (DPPC) bilayers, interdigitated domains began to form at ethanol concentrations
of ca. 0.45 M, while full interdigitation was not observed until 1.0 M.

110

DPPC has 16

carbons in each of its aliphatic tails and DMPC has only 14 carbons, so qualitatively
similar behavior is expected. Because the strength of van der Waals interactions is
expected to scale with acyl chain length, a higher ethanol concentration for DPPC
interdigitation may be expected. However, there are two factors that argue for the fact
that our DMPC data exhibit signs of interdigitation occurring at lower ethanol
concentrations than is seen for DPPC. If we consider the driving force for formation of
the interdigitated bilayers to be the attractive interactions between acyl chains of
opposing lipid molecules, and that all other interactions are independent of acyl chain
length, then it is reasonable to expect longer lipids to undergo interdigitation at lower
concentrations of ethanol.

96

Also contributing to the slight difference between our

findings and those of Löbbecke et al. is the fact that the use of a carboxyfluorescein dye
could be more disruptive to bilayer structure than the pyrene chromophore we use.

110

Their finding that interdigitation was not complete until a concentration of 1.0 M ethanol
was added to the system is based on calorimetric data and a correlated change in the
ethanol concentration dependence of the carboxyfluorescein emission intensity. The
pyrene emission data we report here, while loosely related to the emission processes

55

Löbbecke et al. reported, sense different phenomena and it is therefore not surprising that
these data do not provide exact correspondence.
We have also examined the average size of the vesicles as a function of solution
ethanol concentration using dynamic light scattering measurements.

Our interest in

measuring the average size of the vesicles lies in the fact that the process of
interdigitation is likely to alter vesicle diameter due to the interweaving of the lipid acyl
chains, and a consequent decrease in bilayer thickness. We assert that the intermolecular
spacing of phospholipids is dominated by the van der Waals interactions between acyl
chains. If this spacing is to remain the same in both non-interdigitated and interdigitated
bilayers, the size of the vesicle must change upon interdigitation.

We observe a

discontinuous change in vesicle size in the region of 0.6 M ethanol, consistent with
interdigitation (Fig. 3.4). We note that this anomalous behavior is consonant with that
observed for the steady-state fluorescence emission of pyrene excimer as a function of
the ethanol. There is an apparent increase in vesicle diameter with increasing alcohol
concentration up to ca. 1.0 M ethanol. We ascribe this trend to the disruption of lipid
headgroup-headgroup interactions by ethanol molecules (vide infra), which cause the
bilayer to dilate. The anomalous increase in diameter near 0.6 M likely signals the onset
of a structural phase transition, possibly to the interdigitated phase.
We note that there is significant uncertainty in some of the data, especially for the
higher ethanol concentrations (Fig. 3.4). We understand this uncertainty in the context of
there being an intrinsic vesicle size distribution, and that the functional form of this
distribution may change with increasing ethanol concentration due to vesicle-vesicle or
other unaccounted-for processes.

For ethanol concentrations exceeding 1.0 M, we
56

170

average diameter (nm)

160
150
140
130
120
110

0.0

0.5

1.0

1.5

2.0

solution ethanol concentration (M)

Figure 3.4. Dynamic light scattering data on vesicle size as a function of solution
ethanol concentration.
Uncertainties reported are ± 1 for at least three
determinations.

57

observe a plateau effect in vesicle diameter, and this may be due to the saturation of
interactions between ethanol molecules and the vesicles.

Clearly, more detailed

measurements of the vesicle size distribution will be required to resolve this issue, but the
central point of the data we present here is that there appears to be a predictable
correlation between the average diameter of the vesicles and the solution ethanol
concentration. It is important for us to note that although we were not able to sense any
resolvable change in the local environment of the chromophore upon the addition of
ethanol, we believe our vesicle systems are experiencing a structural phase transition in
the vicinity of 0.6 M ethanol.
Over the last two decades a variety of scientific techniques have been used to
study the ethanol-induced interdigitation of lipid bilayers, all of which have demonstrated
that ethanol does in fact promote lipid interdigitation.

100, 110-117

The one point,

however, that appears to be lipid-dependent is the ethanol concentration at which
interdigitation is first observed. Our steady-state spectroscopic (Fig. 3.3) and DLS (Fig.
3.4) results point to ethanol-induced lipid interdigitation in DMPC vesicles doped with
pyrene at ca. 0.6 M. However, our steady-state results alone do not clarify whether
interdigitation first begins in isolated regions of the bilayer, followed by a more
homogeneous interdigitated structure as ethanol concentration increases.

58

Conclusions
The effects of ethanol on both membrane fluidity and lipid structure have been
examined. We have used steady-state fluorescence measurements of pyrene excimer
formation to monitor changes in the translational mobility of pyrene sequestered within
the acyl chain region of lipid vesicles. We observed a linear trend of increasing excimer
formation as the bulk ethanol concentration was increased, with the exception of an
anomalous increase near 0.6 M solution ethanol concentration, which we believe to be
the point at which interdigitation begins to occur in DMPC vesicles. Dynamic light
scattering data also point to a structural anomaly in the average diameter of DMPC
vesicles in the region of 0.6 M ethanol, indicating that we are sensing a structural phase
transition which is a necessary consequence of the interdigitation process. We assert that
van der Waals interactions mediate bilayer structure, suggesting that inter-acyl chain
spacing should be the same for both interdigitated and non-interdigitated systems. In
order to better gauge the structure and molecular motion of lipids in the presence of
ethanol, we will next need to perform time-resolved fluorescence measurements to study
the dynamic behavior of both gel phase and fluid phase DMPC bilayers.
experiments are discussed in Chapter 4.

59

These

CHAPTER 4
Interrogating the Effects of Ethanol on the Organization of Lipid Vesicles
Using Time-Resolved Fluorescence Measurements

Introduction
Mammalian cell membranes are highly complex biological structures that
function as regulators of cell life, allowing specific chemical species to pass into and out
of the cell. Plasma membranes are heterogeneous structures composed primarily of
glycerophosopholipids, sphingolipids, cholesterol, and proteins arranged in a bilayer
118-120

structure.

Even in compositionally simple systems, phospholipids self-assemble

in aqueous environments to form nominally spherical bilayer liposomes, a structure
which serves to minimize interactions between hydrophobic acyl chains and the aqueous
solvent. In such structures the hydrophobic acyl chain regions of lipids comprise the
interior of a double-leaflet arrangement, while the polar headgroups act to shield the acyl
chains from the aqueous environment and, as such, are located at the surfaces of the
bilayer in direct contact with the solution in which they are suspended. In this structural
format the acyl chain terminal methyl groups of the opposing lipid monolayers are
oriented facing one another. This lipid packing arrangement is the most energetically
favorable, where the number of gauche rotamers in the acyl chain region is minimized,
and van der Waals interactions between chains are optimized. The process of optimizing
van der Waals interactions represents a balance between attractive acyl chain interactions
and the thermal energy present in the system. Lipid bilayer assemblies are known to
undergo phase transitions from ordered to disordered acyl chains, with the transition
60

temperature depending on the length of the acyl chain and the presence of any
unsaturations.
A variety of cellular processes are mediated by both lipid bilayer composition and
dynamic properties of cell membranes, including intercellular and intracellular signaling,
protein and enzyme activity, and cellular response to environmental stresses.

42

One way

in which cells can regulate these functions is by controlling or manipulating the physical
properties of the membrane, such as membrane fluidity, which is mediated by viscous
drag effects between the membrane constituents.

23

Such interactions, along with

membrane composition and structural heterogeneity, determine the motional freedom of
these constituents in the membrane.

121

Membrane fluidity as it relates to the motional

freedom of lipids refers to translational diffusion, rotational diffusion, and lipid
translocation.

23, 121

One way of experimentally altering the membrane fluidity of a

bilayer is by changing the temperature of the local environment of the membrane.
Phospholipid vesicles exhibit a characteristic phase transition at Tm, where the dynamic
properties of lipids are altered due to changes in the organization of the lipid acyl chains.
At temperatures below Tm, lipids exist in the gel phase, characterized by highly ordered
acyl chain packing, where the acyl chains are tilted slightly away from the bilayer
normal.

122

The close-packed arrangement of lipids in the gel phase results from van der

Waals interactions in the hydrophobic acyl chain region of the membrane. Inter-acyl
chain van der Waals interactions limit lipid rotational and translational freedom. At
temperatures above Tm, lipid bilayers exist in a disordered fluid phase, characterized by
61

an increase in the number of trans-gauche conformers and, as a result, an increase in the
average interfacial area subtended by the lipid headgroups.

24

Lipids in the fluid phase

experience greater translational and rotational freedom than lipids in the gel phase,
resulting in increased overall membrane fluidity.

The fluid phase has been shown

experimentally to support a number of protein functions.

123-124

The gel-to-fluid phase transition described above is dominated by acyl chain
interactions, with the lipid headgroups playing a secondary role. Membrane fluidity and
organization can also be controlled by the introduction of short-chain alcohols to the
interfacial region of the bilayer. Alcohols such as ethanol are amphipathic and can
displace some of the water molecules associated with the gel phase lipid bilayer
interfacial region, penetrating to some extent into the nonpolar region of the bilayer.

96

The displacement of water molecules in the vicinity of the lipid headgroups weakens the
interactions between the headgroups, allowing the bilayer structure to dilate, thus
increasing the average surface area per headgroup. The addition of a sufficient amount of
alcohol can perturb the bilayer interfacial region to the extent that a phase transition to an
interdigitated structure occurs.

93-95, 125

Ethanol has been used widely to induce the

formation of the interdigitated phase, with much of the early work being done by Rowe et
al.

114, 126-127

109, 128

and Simon et al.

The interdigitated phase occurs when the acyl

chains of lipids from one leaflet interpenetrate with the chains from the opposing leaflet,
eliminating the bilayer mid-plane, and leading to a consequent decrease in bilayer
thickness. The decrease in bilayer thickness can have a dramatic effect on the function of
proteins imbedded within biological cell membranes as well.
62

96

This explanation does

not resolve whether the interdigitation process is spatially homogeneous or
heterogeneous, or whether there are any molecular-scale changes in the acyl chain region
as a consequence of interdigitation. It is the purpose of this work to address the issue of
ethanol concentration-dependent changes in the molecular-scale organization of these
bilayers.
Several analytical techniques have been used in the past to elucidate the packing
arrangement and properties of phospholipids in interdigitated bilayer arrangements,
including differential scanning calorimetry,
134

and X-ray diffraction.

110, 113

129-130

fluorescence,

111, 131-132

NMR,

133-

A significant limitation of conventional diffraction

techniques for interrogating the structure and organization of unilamellar vesicles is their
relatively low sensitivity, and calorimetric data provide limited molecular-scale insight.
Fluorescence and NMR methods provide the requisite molecular-scale information and
fluorescence is easily capable of detecting low chromophore concentrations. For timeresolved fluorescence measurements, a chromophore can be selected to interrogate a
specific region of the bilayer (i.e. the nonpolar acyl chain region or the polar headgroup
region). We are interested in interrogating the interior acyl chain region of phospholipid
vesicles because this region undergoes substantial structural and dynamic changes as a
result of temperature- or alcohol-induced phase transitions.
In this work we use 100 nm diameter unilamellar vesicles comprised of 1,2dimyristoyl-sn-phosphatidylcholine (14:0 DMPC, Tm = 297 K),

135

and time-resolved

fluorescence techniques to interrogate the properties of a chromophore incorporated into
the vesicle bilayer. We measured the anisotropy decay dynamics of the fluorescent probe

63

perylene in vesicle-containing solutions at 289 K (gel phase DMPC) and 303 K (fluid
phase DMPC) with controlled amounts of ethanol to gauge the extent to which the lipid
phase and ethanol concentration mediate the properties of the bilayer structure, as sensed
by perylene rotational motion. We have used perylene for time-resolved fluorescence
measurements because of its well characterized anisotropy decay dynamics

136-138

and

because its fluorescence lifetime and anisotropy decay time constants are similar,
allowing for efficient collection of the requisite data. We measured the reorientation
dynamics of perylene using time-correlated single photon counting (TCSPC), and found
that this chromophore appears to be sensitive to the ethanol-induced phase transition from
the gel phase to the interdigitated phase in DMPC at 289 K. We also observed a similar,
more pronounced change in the reorientation behavior of perylene as a function of
ethanol concentration at 303 K, a temperature at which DMPC is in the fluid phase. The
interactions between lipid bilayers and ethanol produce measurable changes in lipid
bilayer organization in both the gel and fluid phases.

64

Experimental Section
Vesicle Preparation. The phospholipid 1,2-dimyristoyl-sn-phosphatidylcholine
(14:0 DMPC) in chloroform was purchased from Avanti Polar Lipids, Inc. (Alabaster,
AL) and used as received.

Perylene (99+ %) was purchased from Sigma-Aldrich

(Milwaukee, WI) and used without further purification. For each sample, 10 mg of lipid
was mixed with 30.5 µg of chromophore, and the chloroform solvent was evaporated.
Tris

®

buffer (Sigma-Aldrich) was then added to each sample to make the lipid
®

concentration 1 mg/mL. Tris buffer (10 mM, pH 8.0) was prepared with purified water
from a Milli-Q Plus water purification system (Millipore, Bedford, MA), and was purged
with Ar prior to use. The lipid and chromophore mixtures were processed five times
through a freeze-thaw-vortex cycle to ensure complete mixing of the constituents. Each
cycle was comprised of first freezing the solution by immersion in N2(l) (5 min.), then
thawing in a 60°C water bath (5 min.), followed by vortexing the mixture (2 min.). After
completion of the freeze-thaw-vortex process, sample solutions were extruded eleven
times through a polycarbonate membrane (Whatman) with a nominal pore diameter of
100 nm using a mini-extruder apparatus (Avanti Polar Lipids, Inc.).

25, 103

Once the

vesicle solutions were extruded, controlled amounts of 95% ethanol (Sigma-Aldrich)
were added to the bulk mixtures, and the samples were allowed to equilibrate under
ambient conditions prior to fluorescence acquisition. A time study (results not shown)
indicated that the amount of lipid-ethanol equilibration time needed to achieve
reproducible fluorescence results was ca. 24 hours.

65

Time-Resolved Fluorescence Measurements.

All fluorescence lifetime and

anisotropy decay data were acquired using a time-correlated single photon counting
(TCPSC) instrument. The light source is a CW passively mode-locked, diode-pumped
Nd:YVO4 laser (Spectra Physics Vanguard) that produces 2.5 W average power at 355
nm and 2.5 W of average power at 532 nm at 80 MHz repetition rate, with 13 ps pulses at
both wavelengths. The output of the Nd:YVO4 laser pumps a cavity-dumped dye laser
(Coherent 702-2), which operates in the range of 430 nm to 850 nm, producing 5 ps
pulses. The repetition rate of the dye laser is adjustable between 80 MHz and 80 kHz by
means of the cavity dumping electronics (Gooch & Housego).

The dye laser was

operated with Stilbene 420 dye (Exciton) to produce the excitation wavelength of ca. 430
nm. The fundamental excitation pulse from the dye laser is divided, with one portion of
the pulse directed to a reference photodiode (Becker & Hickl PHD-400-N), and the other
portion directed to the sample. Emission is collected using a 40x reflecting microscope
objective. The collected emission is separated into polarization components parallel (0⁰)
and perpendicular (90⁰) to the vertically polarized excitation pulse using a polarizing
cube beam splitter. The parallel and perpendicular polarized signal components are
detected simultaneously using microchannel plate photomultiplier tubes (MCP-PMT,
Hamamatsu R3809U-50), each equipped with a subtractive double monochromator
(Spectral Products CM-112) for wavelength selection. The detection electronics (Becker
& Hickl SPC-132) resolve the parallel and perpendicular transients separately, yielding
ca. 30 ps response functions for each detection channel.

The detection electronics

include a time-to-amplitude converter (TAC) and a constant fraction discriminator (CFD)
that temporally resolves the fluorescence signal for each polarization component. Data
66

are collected using multichannel analyzers (MCAs), which are integral components of the
SPC-132 electronics. Data acquisition, detector bias, and collection wavelength are all
controlled using an in-house written LabVIEW

®

(National Instruments) program on a

PC. The sample temperature was regulated to ± 0.1⁰C with a water-circulating bath
(Neslab RTE-110) connected to a temperature-controlled brass cell jacket which held the
sample cuvette. All samples were allowed to thermally equilibrate for 10 minutes prior
to data acquisition.

67

Results and Discussion
The goal of the work we report here is to gain a better understanding of the effects
of ethanol concentration and system temperature on the fluidity and molecular-scale
organization of the lipid acyl chain region of 100 nm diameter DMPC vesicles. As noted
above, we have used perylene for time-resolved measurements because its anisotropy
decay dynamics are well understood and are sensitive to changes in its local environment.
Perylene (Fig. 4.1) partitions into the hydrophobic region of DMPC vesicles.

The

reorientation dynamics of perylene thus provide information useful for understanding the
behavior of the bilayer acyl chain region as it experiences structural changes associated
with temperature and exposure to solution phase ethanol. The results of our timeresolved fluorescence measurements show that the organization of DMPC in both the gel
and fluid phases is influenced by the presence of ethanol.
When incorporating a chromophore into the lipid bilayer structure, there is always
a chance that the fluorescent probe can perturb the bilayer to the extent that it could alter
the phase transitions sensed by the probe. Previous work has shown, however, that
perylene does not change the observed gel-to-fluid phase transition temperature (297
135

K)

measurably in DMPC bilayers.

25

We have used an even smaller probe-to-lipid

ratio in this work than we have in the past, and for this reason we do not expect to perturb
the gel-to-fluid phase transition temperature from its reported value.

135

Another

important characteristic of perylene to be considered is where, within the bilayer, the
chromophore localizes. Perylene is thought to locate amongst the acyl chains of the

68

Perylene

normalized intensity (a.u.)

14:0 DMPC

1.0

excitation

emission

0.8
0.6
0.4
0.2
0.0
350

400

450

500

550

600

wavelength (nm)

14:0 DMPC

Figure 4.1. Structures of 14:0 DMPC and perylene, and normalized absorption and
emission spectra of perylene in DMPC vesicles.

69

bilayer leaflets. Some recent work has indicated that perylene moves within the bilayer
to be in closest proximity to the acyl chain terminal methyl groups for vesicles of ca. 1
μm diameter and larger.

76

Those changes in perylene local environment were sensed by

anisotropy decay measurements. We consider next the time-resolved emission behavior
of perylene in both the gel and fluid phases of 100 nm diameter lipid vesicles as a
function of solution phase ethanol concentration to gain insight into the alcohol-mediated
changes in the organization of DMPC bilayer structures.
It is useful to consider at this point the nature of the interactions between the lipid
bilayers and the solution phase ethanol. Because the lipid bilayers are present in the form
of unilamellar vesicles, and ethanol is introduced to the vesicle-containing solutions, at
least the initial interactions between the vesicles and ethanol must of necessity be at the
outer vesicle leaflet. The asymmetric perturbation to the structure of the outer leaflet
must therefore be the first step in the ethanol-induced changes we observe.

If the

chromophore is sequestered within the acyl chain region in a location that is asymmetric
with respect to the inter-leaflet gallery, as our previous data on small vesicles suggest,
76

25,

then we would expect more pronounced dynamical effects for the chromophore

residing in the outer leaflet than in the inner leaflet.
For fluorescence anisotropy decay measurements, polarized fluorescence
transients are collected for polarizations parallel and perpendicular to the vertically
polarized excitation pulse. These quantities, I (t ) and I  (t ) , respectively, are combined
to produce the induced orientational anisotropy function, R(t), (Eq. 4.1),

70

R(t ) 

I (t )  I  (t )

(4.1)

I (t )  2 I  (t )

The chemical information available from these measurements is contained in the
functional form of the R(t) decay. There are several theoretical treatments that relate the
anisotropy decay to the Cartesian components of the rotational diffusion constant and the
transition dipole moment orientation(s) of the reorienting chromophore.

139-141

We treat

the reorientation dynamics of perylene in DMPC bilayers in the context of a free rotor
rather than a hindered rotor

74-75

because perylene is not tethered to the lipid acyl chains,

and the planar shape and aspect ratio of perylene is such that when incorporated into the
bilayer structure of DMPC, there is no intrinsic limit on the rotational motion the
chromophore can execute. We note that the nonpolar chromophore 1,6-diphenyl-1,3,5hexatriene (DPH) has been used previously in the examination of lipid bilayer structures,
and in that work it was treated as a hindered rotor.

142

We assert that perylene rotational

motion cannot be treated in the same context as that of DPH in lipid bilayers because of
the substantially different structures and aspect ratios of these two molecules.
For the majority of anisotropy decay data, a single exponential decay is observed,
and the chemical information contained in such data is limited. Perylene is well suited as
a probe of local organization because it can exhibit either a one- or two-component
anisotropy decay, depending on its immediate environment.

We observe a two-

component anisotropy decay in this work, a condition which yields information on each
of the Cartesian components of the chromophore rotational diffusion constant, D. It is

71

the relative values of Dx, Dy, and Dz that are useful in probing the ethanol- and
temperature-induced lipid bilayer organizational changes of interest.
Relating the functional form of the anisotropy decay to the components of the
rotational diffusion constant requires the assignment of axes and the assumption of a
rotor shape, i.e. the volume swept out by the reorienting molecule. The excited and
emitting transition dipole moments of perylene accessed in our experiments (S1 ↔ S0) lie
along the chromophore long in-plane axis, which we designate as the x-axis. The short
in-plane axis is the y-axis, and the z-axis is perpendicular to the chromophore π-system
plane. Given these assignments, a single-exponential anisotropy decay indicates the
probe reorients as a prolate rotor (Dx > Dy = Dz),

139

R(t )  0.4exp(6Dzt )

(4.2)

and a two-component exponential decay indicates the probe reorients as an oblate rotor
(Dz > Dx = Dy), and is described by,

139

R(t )  0.1exp((2Dx  4Dz )t )  0.3exp(6 Dxt )

(4.3)

For the data we report here, the decay of R(t) was fit best by a two-component
exponential decay (Fig. 4.2), consistent with Eq. 4.3. We have reported previously on the
reorientation dynamics of perylene in DMPC vesicles.

25, 76

For vesicles ranging in size

from 100 nm to 5 μm in diameter, a two-component exponential decay was found. The
data we report here, both above and below the known Tm of DMPC, show perylene to

72

a

1800
1600

counts

1400
1200
1000
800
600
400
200
0
0

1000 2000 3000 4000 5000 6000 7000 8000 9000

time (ps)
0.15

b

R(t)

0.10

0.05

0.00

-0.05
0

1000 2000 3000 4000 5000 6000 7000 8000 9000

time (ps)
Figure 4.2. (a) Experimental data for perylene reorientation in a 100 nm diameter
DMPC vesicle, with 0.25 M ethanol concentration at 289 K. (b) Induced
orientational anisotropy decay function for the data shown in panel (a), along with the
best-fit function to the data. For this fit the function is:
f(t) = (0.11 ± 0.004)exp(-t/(339±24 ps)) + (0.04 ± 0.003)exp(-t/(2483±185 ps)).

73

reorient as an oblate rotor. From the experimental anisotropy decay data we extract the
Cartesian components of the rotational diffusion constant, D. As noted above, Eq. 4.3
assumes Dz > Dx = Dy. We present in Table 4.1 the experimental values of Dz, Dx (=Dy)
and the ratio Dz/Dx, as a function of system temperature and solution phase ethanol
concentration. These data contain several interesting features.
At the outset of this discussion it is important to consider the extent to which
subtle trends in reorientation data can be relied upon. The rotational diffusion constants
we report here are the average and standard deviation (1) for at least four individual
determinations.

We understand that the data we report reveal subtle ethanol

concentration dependencies, a fact that results from the nature of the physico-chemical
phenomena we are examining.
In previous work, the ratio of Dz/Dx was shown to be a useful gauge of changes in
the immediate environment of the chromophore,

136

but such a permutation of our data

provides limited insight into the systems we report on here. Of more immediate benefit,
however, is a direct examination of the ethanol concentration-dependence of the Dz and
Dx values at both 289 K and 303 K. We present these data in Fig. 4.3, revealing
measurable differences in the effect of solution phase ethanol on the local environment of
perylene for the two temperatures.

We ascribe the differences in the ethanol

concentration dependencies of Dz and Dx at the two different temperatures to be the result
of the lipid bilayer existing in the gel phase at 289 K and the fluid phase at 303 K, at least
for modest ethanol concentrations.

25

The Dz values reported at 289 K over all
74

Table 4.1. Cartesian components of the rotational diffusion constant and aspect ratio
Dz/Dx as a function of solution phase ethanol concentration at 289 K and 303 K.
Errors are ±1 for at least four individual determinations at each ethanol
concentration.
T = 289 K, Gel Phase

T = 303 K, Fluid Phase

[Ethanol],
M

Dz (MHz)

Dx (MHz)

Dz/Dx

Dz (MHz)

Dx (MHz)

Dz/Dx

0

600 ± 50

55 ± 2

10.9 ± 1.2

888 ± 80

70 ± 5

12.7 ± 1.1

0.25

703 ± 117

59 ± 3

11.8 ± 1.9

851 ± 65

65 ± 5

13.1 ± 1.7

0.50

689 ± 51

54 ± 3

12.8 ± 1.0

879 ± 92

73 ± 4

12.1 ± 1.5

0.58

665 ± 35

49 ± 2

13.6 ± 0.8

923 ± 73

58 ± 5

16.0 ± 1.8

0.67

737 ± 89

61 ± 7

12.1 ± 2.0

762 ± 50

52 ± 3

14.8 ± 0.5

0.75

651 ± 72

56 ± 3

11.5 ± 1.0

816 ± 89

57 ± 3

14.3 ± 1.8

1.00

676 ± 36

52 ± 2

12.9 ± 0.4

882 ± 76

68 ± 8

13.0 ± 1.2

75

T = 289 K

800

Dz

a

Dx, Dz (MHz)

700
600
500
Dx

60
50
0.0

0.2

0.4

0.6

0.8

1.0

ethanol concentration (M)
T = 303 K

1000

Dz

b

Dx, Dz (MHz)

900
800
700
70
60
Dx

50
40

0.0

0.2

0.4

0.6

0.8

1.0

ethanol concentration (M)
Figure 4.3. (a) Dependence of perylene Dx and Dz on ethanol concentration for 100
nm diameter DMPC vesicles at 289 K. (b) Dependence of perylene Dx and Dz on
ethanol concentration for 100 nm diameter DMPC vesicles at 303 K. Errors are
±1 for at least four individual determinations at each ethanol concentration.

76

ethanol concentrations (Fig. 4.3a) were found to be statistically the same using an
ANOVA analysis at the 99% (α = 0.01) confidence level, while the Dx values were found
to be statistically different for ethanol concentrations in the proximity of 0.6 M. It should
be noted that all ANOVA analyses were tested at both a 95% (α = 0.05) and 99% (α =
0.01) confidence level. In all cases the same conclusion was reached in both instances.
The value of Dx at 0.58 M is anomalously low and the value at 0.67 M appears to be
slightly high relative to the corresponding values at other ethanol concentrations, albeit
within the uncertainty of the Dx values at 0.50 and 0.75 M. These results suggest that
near 0.6 M ethanol, there is some perturbation to the lipid bilayer structure that makes
rotation of perylene about its x-axis somewhat more restricted than for other ethanol
concentrations, and that for 0.67 M ethanol, the value of Dx has at least recovered, if not
increased slightly. It is thought that for small diameter vesicles, perylene resides in the
bilayer between acyl chains, and if this is the case, the chromophore may be sensing the
onset of an interdigitation process. Such a process would be most pronounced in the
region between leaflets, and perylene is likely in close proximity to this portion of the
bilayer. The fact that there is no observable change in Dz over this ethanol concentration
range suggests that the confining, quasi-lamellar structure created within the bilayer acyl
chain region is not disrupted significantly by the structural events responsible for the
changes in Dx.
The Dx (=Dy) and Dz data for perylene in DMPC vesicles at 303 K provide a
qualitatively different picture of ethanol-dependent changes in bilayer organization (Fig.
77

4.3b). For these data we also find, through the use of an ANOVA analysis, that values of
Dx in the vicinity of 0.6 M ethanol are statistically different, while the values of Dz are
statistically not distinguishable. Interestingly, as ethanol concentration increases at 303
K, the value of Dx decreases. If the ethanol-dependent change in behavior we observe is
associated with an interdigitation process, it appears that the onset of interdigitation
causes the acyl chain region of the bilayer to become more restrictive to perylene out-ofplane rotation. The same general trend is seen for Dz, despite the fact that ANOVA
analysis shows that Dz is statistically independent of ethanol concentration.

However,

the ethanol concentration at which the apparently anomalous behavior in Dz occurs is not
commensurate with that observed for Dx (Fig. 4.3b). This finding may be related to
modulation of the acyl chain density by the interdigitation process.

The ethanol-

dependent reorientation dynamics of perylene at 289 K and 303 K are measurably
different from one another, and we correlate this difference with DMPC being in the gel
phase at 289 K and the fluid phase at 303 K, at least for modest ethanol concentrations.

25

In comparing the data taken at 289 K to those taken at 303 K, we note that the
rotational diffusion constants at 303 K are uniformly faster than those at 289 K. This is
not surprising, for two reasons. The first is that the local frictional (viscous) interactions
experienced by perylene are expected on theoretical grounds to be temperature
dependent,

39, 143-144

and the second is that, for pure DMPC, there is a phase transition

between the two temperatures at which we recorded data. The difference in phases is
seen in the different Dz/Dx ratios for perylene at the two temperatures. ANOVA analysis
78

showed that the value of Dz/Dx for perylene at 289 K is independent of ethanol
concentration to within the experimental uncertainty, at a value of 12.2 ± 0.5. The same
analysis applied to the Dz/Dx ratio for perylene at 303 K is ethanol-dependent. The acyl
chain environment for perylene is slightly more anisotropic at 303 K (Table 4.1) than it is
at 289 K. Such behavior may reflect either a difference in the relative location of the
chromophore in the acyl chain region at the two temperatures, or a change in the
environment at a specific location within the bilayer acyl chain region. Regardless, the
local environment sensed by perylene appears to be different at the two temperatures.
It is possible to gauge the frictional intermolecular interactions experienced by
perylene in the bilayer structures. We can calculate the rotational diffusion constant, D,
using the Cartesian components Dx (=Dy) and Dz given in Table 4.1.
D

Dx  D y  Dz

(4.4)

3

The rotational diffusion constant is related to the viscosity of the local environment
sensed by the probe according to the modified Debye-Stokes-Einstein equation,

39, 143-

144

1
Vf

6 D k BTS

(4.5)

where η is the viscosity of the surrounding medium of the reorienting chromophore, V is
the hydrodynamic volume of the chromophore, f is a frictional term to compensate for
solvent-solute interactions, kB is the Boltzmann constant, T is the temperature, and S is a
shape factor determined from Perrin’s equations to account for the nonspherical shape of
79

the chromophore.

39

We used the rotational diffusion constants calculated from Eq. 4.4,

the shape factor, S = 0.70,

145

3 146

and the hydrodynamic volume of perylene, V = 225 Å ,

to determine the viscosity, , of the perylene local environment as a function of ethanol
concentration. We find that there is no resolvable ethanol concentration dependence of
the acyl chain micro-viscosity sensed by perylene using ANOVA analysis, indicating that
the ethanol is not penetrating fully into the acyl chain region, but is instead acting
primarily on the lipid headgroup region. For data taken at 289 K, we calculate  = 7.9 ±
0.3 cP and at 303 K,  = 6.6 ± 0.2 cP. These values are consistent with those we have
measured previously, indicative of an acyl chain region that does not contain constituents
other than phospholipids.

25

We attribute the difference in calculated viscosity to the

temperature difference between the two measurements and the fact that DMPC is initially
in the gel phase at 289 K and the more disordered fluid phase at 303 K. It is significant
that these values are consistent with those reported for DMPC without the addition of
25

ethanol,

suggesting that ethanol penetration of the DMPC bilayer does not proceed to

the extent of perturbing the acyl chain region of the structure, at least prior to the point of
causing interdigitation.
It is known that the ethanol-induced interdigitated phase results when amphipathic
molecules disrupt the lipid headgroup interactions to the extent that the bilayers dilate
and eventually interpenetrate.

94

When lipid interdigitation takes place, it is characterized

by a closely-packed gel phase structure, and a consequent decrease in membrane
fluidity.

107

At a concentration of ca. 0.6 M ethanol in our systems, perylene experiences

80

a decrease in its motional freedom for Dx at 289 K and 303 K. These data are consistent
with the DMPC lipid bilayers becoming interdigitated at ca. 0.6 M ethanol. One might
expect there to be a requisite increase in Dx and/or Dz with the onset of lipid headgroup
dilation preceding interdigitation, but this is not seen. We believe that this is due to the
fact that perylene resides in the lipid acyl chain region, relatively well isolated from the
phospholipid headgroups, and that up to the point of interdigitation the van der Waals
interactions between acyl chains dominate the local environment experienced by
perylene. For ethanol concentrations above 0.6 M, it is likely that the lipid headgroup
interactions become further disrupted, eventually giving rise to the loss of bilayer
integrity. Up to the point of bilayer destruction, however, it is likely that the perturbation
to the headgroup organization is resisted by the van der Waals attractive interactions
between lipid acyl chains. For ethanol concentrations of ca. 1 M, we observe dynamics
that differ little from those measured for aqueous vesicle solutions, underscoring the
importance of dispersion forces in mediating lipid local organization.
Our data on 100 nm diameter DMPC lipid vesicles at 303 K, a temperature above
the known gel-to-fluid phase transition temperature (Fig. 4.3b), show an ethanol
concentration-dependence that is qualitatively similar, albeit different in detail, from that
seen for the 289 K data (Fig. 4.3a). Much work has been conducted in the last two
decades on ethanol-induced interdigitation of gel phase lipid bilayers, all of which have
demonstrated through a variety of techniques that ethanol does promote lipid
interdigitation.

95, 110-114

While there are no data we are aware of that have reported on

the existence of an ethanol-induced interdigitation in fluid phase lipid bilayers, our data

81

reveal fluid phase behavior that is commensurate with gel phase behavior. Our data
reveal ethanol concentration-dependent changes in Dx in the vicinity of 0.6 M ethanol.
While the details of changes in Dx at 303 K differ from those seen at 289 K, they do show
a change in lipid acyl chain environment, consistent with an interdigitation phase
transition seen for gel phase DMPC. Further experiments will be required to resolve the
details of this anomalous organizational change in fluid phase DMPC, and those
experiments are underway in our laboratory.

82

Conclusions
We have found that the presence of solution phase ethanol has an effect on the
organization of DMPC lipid bilayer structures, both below and above the gel-to-fluid
phase transition temperature.

At 289 K, we observe that the rotational diffusion

dynamics of perylene incorporated in the bilayer acyl chain region experience a
perturbation about their long in-plane rotational axis for ethanol concentrations in the
vicinity of 0.6 M. Interestingly, there is no measurable effect on the rotation of perylene
about the axis normal to its molecular plane. For the same measurements performed at
303 K, we similarly observe a perturbation to out-of-plane (Dx, Dy) perylene rotational
motion when the solution phase ethanol concentration is in the vicinity of 0.6 M,
although the onset of the perturbation occurs at slightly different ethanol concentrations,
which is likely the result of the extent of structural perturbation experienced by the lipid
acyl chain region.

Again, there is no measurable perturbation to the in-plane (Dz)

rotational motion of perylene. Our findings are consistent with the known phases of the
lipid vesicles. Phospholipids are known to undergo interdigitation phase transitions in
their gel phases at ethanol concentrations in the vicinity of 0.6 M.

110

Our findings point

to the influence of ethanol in mediating a structural perturbation for DMPC in the fluid
phase as well, although the structural details of this perturbation remain to be resolved
fully. Further work is underway to understand this effect more thoroughly.

83

CHAPTER 5
Effects of Localized Energy Dissipation on the Motional Dynamics of
Unilamellar Vesicles

Introduction
There has been a great deal of research interest in the organization and dynamics
of lipid bilayer structures because of their intrinsic importance to life processes

42

and

their complexity. Mammalian plasma membranes contain between 100 and 500 different
chemical constituents, with the identity and amount of these constituents depending on
the type of cell. This structural complexity appears to be a central factor in mediating the
organization and function of transmembrane proteins, and it is the broad issue of how
lipid bilayer constituents interact with transmembrane proteins and other bilayer
constituents that remains to be understood in detail. Specific areas of importance relating
to the organization of lipid bilayer structures are the relationship between bilayer
composition and fluidity, and the nature of the coupling between the bilayer structures
and transmembrane proteins.

23

There are a variety of ways to address issues of

intermolecular organization, and one of them is to evaluate the efficiency of energy
transfer between constituents.
Lipid bilayers are known to behave as fluids, and this factor allows for the facile
incorporation of transmembrane proteins. The viscosity of a bilayer membrane has been
the subject of some debate since the early 1970s.

84

147-150

Initial estimates of the average

bilayer viscosity predicted a value of 100 cP.

150

Subsequent estimates, however, have

likened the bilayer lipid viscosity to that of olive oil or ―crocodile fat on a warm
summer’s day‖.

150

The fluidity of lipid bilayer structures has been examined somewhat

more quantitatively by a number of groups, and perhaps one of the most direct means of
evaluating this property is through orientational relaxation measurements of selected
species incorporated in the bilayer structure.

This type of measurement, which is

sensitive to molecular rotational motion, has been used to improve our understanding of
lipid bilayer fluidity. It has been found from this work that the addition of selected
bilayer constituents can mediate the microscopic viscosity of the acyl chain environment
and lower the temperature at which phase transitions occur in the bilayers.
Multicomponent systems can exhibit viscosities on the order of several cP, consistent
with high mobility of bilayer constituents within the membrane, but the details of this
fluidity depend sensitively on the composition of the bilayer.

The term ―fluidity‖

encompasses a number of physical and chemical properties within the bilayer. These
include microscopic viscosity, translational diffusion of the bilayer constituents, and
inter-leaflet translocation of lipids. Addressing the full complexity of the fluid properties
of lipid bilayers ultimately requires the examination of these structures from a number of
perspectives.
In simple bilayer systems, composed of a single phospholipid species, membrane
viscosity can be altered by changing the system temperature. Such lipid bilayers are
characterized by a phase transition temperature (Tm), where the extent of order within the
acyl chain region of the bilayer goes from an ordered (gel) phase to a less ordered (fluid)
85

phase. The transition temperature Tm depends sensitively on the functionality of the
headgroup, the presence of any acyl chain unsaturations, and the length of the acyl
chains.

24

The phase of the lipid bilayer can affect the activity of transmembrane proteins

contained within the bilayer structure.

123, 151

The gel-to-fluid phase transition within the bilayer can be sensed on the molecular
level by measuring the orientational relaxation time, OR, of a fluorescent probe molecule
imbedded in the nonpolar acyl chain region of the bilayer.

25

The acyl chain interior is

the region of the bilayer most affected structurally and dynamically by phase changes,
and as such is the moiety of interest.

Several studies have employed molecular

reorientation measurements to study the dynamic behavior of lipid vesicles as a function
of temperature,

10

1, 25

impurities.

76

vesicle diameter,

headgroup functionality,

152-154

and the presence of

These reports demonstrate collectively that the use of molecular

reorientation measurements is well suited to the interrogation of molecular scale
properties within lipid bilayer structures.
We are concerned in this work with the issue of local, rather than systemic,
temperature change within lipid bilayer structures. Specifically, we are interested in
understanding how the localized deposition of thermal energy into a bilayer structure
perturbs the system, and to gauge the physical extent of this perturbation on a timescale
relevant to molecular motion.

In order to address these issues, we have used the

chromophore perylene as a local ―radiator‖, and compare its molecular reorientation

86

dynamics when little if any excess thermal energy is deposited into the chromophore
immediate environment versus when ca. 2 eV of thermal energy are dissipated
nonradiatively. We use the phospholipid 1,2-dimyristoyl-sn-phosphatidylcholine (14:0
135

DMPC, Tm = 297 K)

to create unilamellar vesicles of ca. 100 nm diameter and

compare the rotational diffusion behavior of perylene1, 25 when excited to two different
electronic states. We find that there is indeed an excitation state-dependence to the
reorientation dynamics of perylene and we relate our findings to the dissipation of excess
excitation energy.
We excite the S1 ← S0 and the S2 ← S0 transitions of perylene and monitor
emission from the S0 ← S1 transition. Because of the small Stokes shift exhibited by
perylene and its high fluorescence quantum yield, there is little if any local heating that
results from excitation of the S1 ← S0 transition. The perylene S2 ← S0 transition is 1.92
eV higher in energy than the S1 ← S0 transition, and subsequent to S2 ← S0 excitation,
rapid nonradiative relaxation to the S1 state occurs. The dissipation of 1.92 eV of excess
excitation energy into the surrounding lipid acyl chain bath will result in local heating.
By monitoring the state-dependent change in reorientation dynamics of the chromophore
and through the application of a simple model, we can estimate the temperature change
experienced by the chromophore local environment.

155

The transient temperature

change of the lipid bilayer is on the order of 4 K, with a possible dependence on the
phase of the lipids surrounding the chromophore. Our data also suggest that the spatial

87

extent of the local heating phenomenon depends sensitively on the extent of organization
of the lipid acyl chains.

88

Experimental Section
Vesicle Preparation. The phospholipid 1,2-dimyristoyl-sn-phosphatidylcholine
(14:0 DMPC) in chloroform was purchased from Avanti Polar Lipids, Inc. (Alabaster,
AL) and used as received.

Perylene (99+ %) was purchased from Sigma-Aldrich

(Milwaukee, WI) and used without further purification. For each sample, 10 mg of lipid
was mixed with 30.5 µg of chromophore, and the chloroform solvent was evaporated.
Tris

®

buffer (Sigma-Aldrich) was then added to each sample to make the lipid
®

concentration 1 mg/mL. Tris buffer (10 mM, pH 8.0) was prepared with purified water
from a Milli-Q Plus water purification system (Millipore, Bedford, MA), and was purged
with Ar prior to use. The lipid and chromophore mixtures were processed through five
freeze-thaw-vortex cycles to ensure complete mixing of the constituents. Each cycle
consisted of freezing the solution by immersion in N2(l) (5 min.), followed by thawing in
a 60°C water bath (5 min.), then vortexing the mixture (2 min.). After completion of the
freeze-thaw-vortex process, sample solutions were extruded eleven times through a
polycarbonate membrane (Whatman) with a nominal pore diameter of 100 nm using a
mini-extruder apparatus (Avanti Polar Lipids, Inc.) to produce unilamellar vesicles.25, 103
Time-Resolved Fluorescence Measurements.

All fluorescence lifetime and

anisotropy decay data were acquired using a time-correlated single photon counting
(TCPSC) instrument. The light source is a CW passively mode-locked, diode-pumped
Nd:YVO4 laser (Spectra Physics Vanguard) that produces 2.5 W average power at 355
nm and 2.5 W of average power at 532 nm at 80 MHz repetition rate, with nominally the

89

same 13 ps pulses at both wavelengths. The output of the Nd:YVO4 laser pumps a
cavity-dumped dye laser (Coherent 702-2), which operates in the range of 430 nm to 850
nm, producing 5 ps pulses. The repetition rate of the dye laser is adjustable between 80
MHz and 80 kHz by means of the cavity dumping electronics (Gooch & Housego). For
excitation of the S1 ← S0 transition, the dye laser was operated with Stilbene 420 dye
(Exciton) to produce the excitation wavelength of ca. 435 nm.

For the S2 ← S0

transition, Coumarin 500 dye (Exciton) was used to produce an excitation pulse of 520
nm.

For both excitation wavelengths, the laser output is linearly polarized with a

polarization extinction ratio of ca. 100. The 520 nm excitation pulses were frequency
doubled using a Type I BBO crystal to produce a pulse of ca. 260 nm. The 260 nm pulse
polarization was rotated to vertical using a UV-grade quartz polarization rotating Fresnel
rhomb pair (CVI). The fundamental excitation pulse from the dye laser is divided, with
one portion of the pulse directed to a reference photodiode (Becker & Hickl PHD-400N), and the other portion directed to the sample. Emission is collected using a 40x
reflecting microscope objective (Ealing).

The collected emission is separated into

polarization components parallel (0⁰) and perpendicular (90⁰) to the vertically polarized
excitation pulse using a polarizing cube beam splitter (Newport, extinction ratio ≥ 500:1).
The parallel and perpendicular polarized signal components are detected simultaneously
using microchannel plate photomultiplier tubes (MCP-PMT, Hamamatsu R3809U-50),
each equipped with a subtractive double monochromator (Spectral Products CM-112) for
wavelength selection. The detection electronics (Becker & Hickl SPC-132) resolve the
parallel and perpendicular transients separately, yielding ca. 30 ps response functions for

90

each detection channel. The detection electronics include a time-to-amplitude converter
(TAC) and a constant fraction discriminator (CFD) that temporally resolves the
fluorescence signal for each polarization component.

Data are collected using

multichannel analyzers (MCAs), which are integral components of the SPC-132
electronics. Data acquisition, detector bias, and collection wavelength are all controlled
using an in-house written LabVIEW

®

(National Instruments) program on a PC. The

sample temperature was regulated to ± 0.1⁰C with a water-circulating bath (Neslab RTE110) connected to a temperature-controlled brass cell jacket which held the sample
cuvette. All samples were allowed to equilibrate thermally for 10 minutes prior to data
acquisition.
Steady-State Fluorescence Measurements. Excitation and emission spectra were
acquired for our vesicle samples in order to determine the perylene band positions. We
used a Hitachi F-4500 fluorescence spectrophotometer for excitation measurements, and
a Spex Fluorolog 3 emission spectrometer for emission measurements. Both instruments
were set to a spectral band pass of 1 nm for both excitation and emission
monochromators.

91

Results and Discussion
Although it is clear that there are thermally driven phase transitions within lipid
bilayer structures, it is not clear whether or not such transitions can proceed in isolation
or in areas of molecular-dimension for a bilayer exposed to a point source thermal
transient. This issue may appear to be relatively esoteric, but it is of central importance
to the study of bilayers. Many spectroscopic and electrochemical studies of bilayer
structure, organization and dynamics are performed using probe molecules to report on
their local environment. The results of such studies make the underlying assumption that
the presence of the probe and any excess energy that it may dissipate do not alter the
properties of the lipid bilayer. We assert that this is, at least in certain cases, a poor
assumption, and seek to explore the consequences of a spectroscopic probe being present
in a bilayer system.

It is clear that probe molecules are not the only non-lipid

constituents in plasma membrane structures, and the data we present here bear, in
principle, on the influence of any bilayer non-lipid constituent.
We use perylene as a probe of local organization in lipid bilayers. We use this
chromophore in low concentrations to minimize large scale disruption of bilayer
organization.

Previous work using higher perylene concentrations showed that the

presence of the chromophore did not lower the gel-to-fluid phase transition
temperature.

25

The work we report here, where two different perylene transitions are

accessed, makes use of nonradiative energy transfer from the chromophore to the lipid
bath modes to infer information on local organization within the lipid bilayer. The
spectroscopic properties of perylene are well understood and characterized, and the

92

measurement we use, fluorescence anisotropy decay, can provide information on the
short range dissipation of energy that is not readily obtainable by other means.
The goal of the work presented here is to gain insight into the effects of transient
local heating in simple lipid vesicle systems, and to gauge the spatial extent of the
thermal perturbation to the bilayer structure on the timescale of molecular motion.
Understanding the extent to which an energetic perturbation within a lipid bilayer can
influence its fluidity is important because of the dependence of cellular regulation
processes on plasma membrane properties.

49, 156-157

The chromophore we have chosen

for this work partitions selectively into the nonpolar acyl chain region of the bilayer, and
it is thought that this region undergoes the most marked organizational changes during a
phase transition. Previous studies have shown that the location of perylene within the
DMPC acyl chain region depends on the diameter of the extruded vesicle.

76

For

relatively small vesicles (i.e. 100 - 800 nm diameter), the phase transition of DMPC (Tm
= 297 K) influences the motional freedom of the chromophore, while in larger diameter
vesicles (i.e. ≥ 1 μm) chromophore motion is not sensitive to the phase transition. This
vesicle size-dependent change in the sensitivity of the chromophore to the phase
transition is associated with a discontinuous change in its reorientation dynamics,
indicating that the location of perylene within the bilayer acyl chain region changes
according to vesicle size and curvature. One explanation consistent with these findings is
that perylene is located interstitially, between acyl chains, in small DMPC vesicles, and
that as the diameter of the vesicle increases, minimizing any structural or organizational
difference between inner and outer leaflets, the chromophore relocates into the bilayer
93

mid-plane region. It is important to consider the location of perylene in the lipid bilayer
as we interpret the results of our fluorescence anisotropy decay measurements for
perylene in unilamellar vesicles as a function of system temperature. Because of the
diameter of the vesicles we use here, we assert that perylene is located within the acyl
chains and not in the inter-layer gallery.
We measured the fluorescence anisotropy decay of perylene (Fig. 5.1) in DMPC
unilamellar vesicles at temperatures above and below Tm for this phospholipid. For such
measurements, polarized fluorescence transients are acquired for polarizations parallel,
I||(t), and perpendicular, I(t), to the vertically-polarized excitation pulse. These two
quantities are combined to produce the induced orientational anisotropy function, R(t ) ,
(Eq. 5.1).

R(t ) 

I (t )  I  (t )

(5.1)

I (t )  2 I  (t )

The functional form of the R(t ) decay contains the chemical information of interest.
Chromophores can produce anisotropy decays with as many as five exponential
components, but in most cases only a one- or two-component anisotropy decay is
observed experimentally.

Perylene can exhibit a one- or two-component decay,

depending upon the nature of its local environment. For our experiments, we observe a
two-component anisotropy decay for excitation of the S1 ← S0 transition ( S1 = 435 nm)
and a one-component decay for excitation of the S2 ← S0 transition ( S 2 = 260 nm).

94

For both excitation conditions, we monitor emission from the S0 ← S1 transition. The
excitation and emission spectra of perylene are shown in Figure 5.1. To understand the
information content of these data, we first consider their treatment in the context of
ellipsoids of rotation.
There have been several theoretical treatments relating time-domain fluorescence
anisotropy decay to the Cartesian components (Dx, Dy, and Dz) of the rotational diffusion
constant, D, and the orientation of the emitting transition dipole moment of the
38, 139-140

chromophore.

D

The rotational diffusion constant is defined as:

Dx  Dy  Dz

(5.2)

3

Chuang and Eisenthal derived a series of expressions for the functional form of R(t )
based on the assignment of Cartesian axes and the assumption of a rotor shape.

139

The

rotor shapes are described as being either oblate or prolate ellipsoids, and they represent
the volume swept out by the reorienting chromophore. For perylene, both the excited and
emitting transition dipole moments for the S1 ↔ S0 transition lie along the long in-plane
axis, designated as the x-axis. The y-axis is defined as the short in-plane axis and the zaxis is perpendicular to the chromophore π-system plane. For excitation of the perylene
139

S1 ← S0 transition, reorientation as a prolate rotor (Dx > Dy = Dz)

gives rise to a

single-component exponential anisotropy decay,

R(t )  0.4exp(6Dz t )

(5.3)

95

normalized intensity (a.u.)

Perylene

S1 S0 excitation

1.2
1.0
0.8

S2  S0 excitation
S0 S1 emission

0.6
0.4
0.2

0.0
200 250 300 350 400 450 500 550 600

wavelength (nm)

14:0 DMPC
Figure 5.1. Structures of 14:0 DMPC and perylene, and normalized excitation and
emission spectra of perylene in DMPC vesicles.

96

and reorientation as an oblate rotor (Dz > Dx = Dy) produces an anisotropy decay with
two exponential decay components,

R(t )  0.1exp((2Dx  4Dz )t )  0.3exp(6 Dxt )

(5.4)

Our data for excitation of the S1 ← S0 transition were fit best by a doubleexponential decay (Fig. 5.2b), indicating that perylene behaves as an oblate rotor (Eq.
5.4) in DMPC, above and below Tm. These findings are fully consistent with previous
1, 25, 76

reports.

Because perylene reorients as an oblate rotor, we are able to extract the

values of the Cartesian components of the diffusion constant. These values, along with
the ratio Dz/Dx, are presented as a function of system temperature in Table 5.1. We will
consider their significance in the context of our data for excitation of the S2 ← S0
transition (vide infra).
The S2 ← S0 transition of perylene is polarized nominally perpendicular to the
S1 ← S0 transition and for excitation of S2 ← S0 and monitoring S0 ← S1, R(t) for an
oblate rotor is given by

R(t )  0.1exp(6Dxt )  0.3exp((2Dx  4Dz )t )

(5.5)

Based on the results for excitation of the S1 ← S0 transition, indicating that perylene
reorients as an oblate rotor, we assume that the rotor shape and aspect ratio (Dz/Dx) of
perylene are independent of the state to which the molecule is excited. Eq. 5.5 predicts
97

1600

a

parallel

1400

counts

1200
1000
800

perpendicular

600
400
200
0
0

1000 2000 3000 4000 5000 6000 7000 8000 9000

time (ps)
0.20

b

R(t)

0.15
0.10
0.05
0.00
-0.05
0

1000 2000 3000 4000 5000 6000 7000 8000 9000

time (ps)
Figure 5.2. (a) Experimental data for perylene reorientation in 100 nm diameter DMPC
vesicles at 292 K, excited at 435 nm (S1 ← S0). (b) Induced orientational anisotropy
decay function for the data shown in panel (a), along with the best-fit function to the data.

98

Table 5.1. The rotational diffusion constant, aspect ratio Dz/Dx, and calculated local
viscosity as a function of temperature.

λ

S1

= 435 nm, S1 ← S0

λ

η (cP)

Temp. (K)

Dz/Dx

292

11.8 ± 0.8

254 ± 13 8.3 ± 0.4

11.8

294

11.8 ± 0.1

269 ± 1

7.8 ± 0.1

11.8

296

11.9 ± 0.4

294 ± 7

7.2 ± 0.2

11.9

296.5

12.6 ± 0.7

322 ± 14 6.6 ± 0.3

12.6

298

12.5 ± 0.4

386 ± 10 5.5 ± 0.1

12.5

300

12.5 ± 0.3

396 ± 8

5.4 ± 0.1

12.5

302

13.2 ± 0.4

427 ± 10 5.1 ± 0.1

13.2

a

D (MHz)

S2

Dz/Dx
a

a

a

a

a

a

a

= 260 nm, S2 ← S0

D (MHz)

η (cP)

311 ± 5

6.7 ± 0.1

346 ± 6

6.1 ± 0.1

406 ± 2

5.2 ± 0.1

438 ± 2

4.8 ± 0.1

527 ± 8

4.1 ± 0.1

530 ± 16

4.1 ± 0.1

561 ± 22

3.9 ± 0.1

The values of the aspect ratio (Dz/Dx) for the S2 ← S0 transition were assumed to

remain unchanged from those determined experimentally for the S1 ← S0
transition. Errors are ±1 for at least four individual determinations at each
temperature.

99

that the recovered anisotropy decay function will be characterized by a negative zerotime anisotropy because of the perpendicular polarizations of the excited and emitting
transitions. Despite the prediction of a two-component anisotropy decay, we resolve only
one of the decay components (Fig. 5.3). We understand the reason for this finding in the
context of more than one fluorophore being present in the solution. An examination of
the steady-state and time-resolved anisotropy measurements of solutions containing only
10 mM Tris

®

buffer solution reveals the presence of a weak emission component (data
®

not shown). Exciting the Tris buffer solution at 260 nm produces emission in the range
of 375 nm to 475 nm, with the excitation and emission transition moments for the
®

contaminant polarized parallel to one another. In solutions containing TRIS -buffered
DMPC but no perylene, we recover an anisotropy decay for the contaminant that is given
by



t
R(t )  0.06exp 

 3000  225 ps 

(5.6)

This emissive species, when present in solutions containing small amounts of perylene,
interferes with the slow (OR ~ 2800 ps) (negative) anisotropy decay of the chromophore
and as a result we can only resolve the fast component of the perylene anisotropy decay.
With the assumption that the rotor shape and aspect ratio are the same for the two
excitation conditions, we can recover Dz and Dx using Eq. 5.7.

 OR1 

1
2 Dx  4 Dz

(5.7)

100

a

perpendicular

3000

counts

2500

parallel

2000
1500
1000
500
0
0

500 1000 1500 2000 2500 3000 3500 4000 4500 5000

time (ps)
0.04

b

0.02

R(t)

0.00
-0.02
-0.04
-0.06
-0.08
-0.10
0

1000

2000

3000

4000

5000

time (ps)

Figure 5.3. (a) Experimental data for perylene reorientation in 100 nm diameter
DMPC vesicles at 292 K, and excited at 260 nm (S2 ← S0). (b) Induced
orientational anisotropy decay function for the data shown in panel (a), along with
the best-fit function to the data.
101

We present these results in Table 5.1. The data show that excitation of the S2 ← S0
transition gives rise to faster rotational diffusion constants than excitation of the
chromophore S1 ← S0 transition. We have observed this phenomenon before for a
rhodamine chromophore

155

and understand this behavior in the context of local heating

due to the dissipation of excess excitation energy by nonradiative transfer to the bath.
Based on the data presented in Table 5.1, it is clear that D for excitation of the
S2 ← S0 transition of perylene is larger than D for excitation of the S1 ← S0 transition.
We can relate the rotational diffusion constant, D, to the viscosity of the medium in
which the chromophore resides through the modified Debye-Stokes-Einstein (DSE)
equation. This equation is intended to describe the rotational motion of an ellipsoidal
rotor in a continuum solvent, a situation that does not describe the environment formed
within a lipid bilayer accurately. Despite this issue, we apply the DSE model to the same
chromophore under conditions of excitation at two different wavelengths. Thus any
issues relating to the application of this model to our data are consistent throughout. The
39, 143

modified Debye-Stokes-Einstein equation is given by
1
Vf

6 D k BTS

(5.8)

In this equation η is the solvent bulk viscosity, V is the hydrodynamic volume of the
reorienting chromophore, f is a frictional term used to compensate for interactions
between the solute and the solvent, kB is the Boltzmann constant, T is the temperature of

102

the system, and S is a shape factor that is calculated using Perrin’s equations to account
for the non-spherical shape of the chromophore.
hydrodynamic volume of perylene,

158

39

3

We take V = 225 Å for the

S = 0.70 for the shape factor,

145

and f = 1 for the

frictional coefficient. In this equation the only temperature-dependent terms are the
solvent bulk viscosity and the temperature of the system.

The essential issues to

understanding these data are knowledge of the amount of excess excitation energy
deposited in the system and an understanding of the temperature dependence of the
viscosity of the lipid bilayer medium. The amount of excess energy deposited into the
system can be calculated in a straightforward manner. This energy is the difference in
energy between the S2 ← S0 transition (ex = 260 nm, Eex = 4.77 eV) and the S1 ← S0
transition (0-0 = 435 nm, E0-0 = 2.85 eV), which is 1.92 eV. This 1.92 eV of energy is
dissipated nonradiatively into the bath modes of the surrounding lipid acyl chains.
Figure 5.4 shows that the dissipated energy from the excitation of the S 2 ← S0 transition
results in faster reorientation times for this transition, as opposed to the S1 ← S0
transition, because the excess excitation energy is presumably disordering the local
environment of the chromophore, and decreasing the local viscosity.
We relate the state-dependent rotational diffusion behavior to a change in both the
viscosity of the medium and the temperature of the system using the modified DSE
equation.

    Vf  Vf
 OR     

 T   k B S  k BTS



(5.9)

103

400
380
360

reorientation time (ps)

340

S1 excitation

320
300
280
260
240
220

S2 excitation

200
180
160
292

294

296

298

300

302

temperature (K)

Figure 5.4. Dependence of reorientation time on temperature for 100 nm diameter
DMPC vesicles, for excitation of the S1 ← S0 transition (○) and the S2 ← S0
transition (●). Errors are ±1 for at least four individual determinations at each
temperature.

104

Using Eq. 8 we can calculate the value of Δη at each temperature for which fluorescence
anisotropy data were recorded, and we show these results in Table 5.2. In order to
determine the change in local temperature (ΔT) that gives rise to the state-dependent
reorientation times, we must have a model for the temperature dependence of the lipid
acyl chain viscosity.
The temperature dependence of viscosity depends on the physical properties of
the medium, and has been determined experimentally for a variety of liquids.

159

The

same is not true for lipids, however, and we know of no reliable evaluation of the
temperature-dependent viscosity characteristics of DMPC.

150

Previous rotational

diffusion studies of perylene in DMPC have shown that the viscosities of the lipid acyl
chain region, at temperatures in the vicinity of the lipid Tm, are similar to those of normal
alcohols.

25

Based on the data we report here, we estimate the viscosity of the DMPC acyl
159

chain region to be similar to that of n-octanol (η = 7.4 cP)

for temperatures below Tm,

and similar to the viscosity of n-heptanol (η = 5.8 cP) for temperatures above Tm (Table
5.1). If we use the model of n-alcohols for the lipid acyl chains, and assume that their
local viscosity obeys the same temperature dependence (on length scales short enough
that we do not have to consider a mesoscopic phase transition), we can use Eq. 5.10 to
estimate the temperature change associated with the experimentally observed OR. A
parameterized fit can be used to relate η to T

159

105

Table 5.2. Experimental reorientation times and calculated changes in local viscosity
and temperature associated with internal conversion of perylene.

Temp.

τOR1 (S1)

τOR1 (S2)

ΔτOR1 (S2 - S1)

Δη

ΔT

(K)

(ps)

(ps)

(ps)

(cP)

(K)

292

370 ± 22

300 ± 5

-70 ± 23

-0.88 ± 0.28

3.6 ± 1.2

294

349 ± 3

270 ± 4

-79 ± 5

-1.00 ± 0.06

4.1 ± 0.3

296

317 ± 10

231 ± 2

-86 ± 10

-1.09 ± 0.13

4.5 ± 0.5

296.5

289 ± 15

212 ± 1

-77 ± 15

-0.98 ± 0.19

4.1 ± 0.8

298

241 ± 7

176 ± 3

-65 ± 8

-0.83 ± 0.10

4.0 ± 0.5

300

235 ± 5

175 ± 5

-60 ± 7

-0.77 ± 0.09

3.7 ± 0.4

302

216 ± 6

165 ± 6

-51 ± 8

-0.66 ± 0.11

3.2 ± 0.5

Errors are ±1 for at least four individual determinations at each temperature.

106

T

A



 B  C

(5.10)

A, B, and C in this expression are empirical constants. The temperature dependence of
the viscosity is given by Eq. 5.11.

155

d
 2

dT T
A  2B

(5.11)

0

where T0 is the ambient temperature of the vesicle-containing solution. Literature values
for A and B were used along with the viscosities for n-octanol and n-heptanol, and the
experimentally derived values of Δη ~ dη (Eq. 5.9) in order to calculate dT ~ ΔT (Table
5.2). The calculated value of ΔT is an estimate of the amount of transient heating that
155

occurs to the lipid acyl chains, which in turn gives rise to the observed ΔOR.

We

calculate an average value of ΔT of 4 K, but for temperatures below the DMPC phase
transition (< 297 K) the average value of ΔT is slightly higher (i.e. 4.1 K) than it is for
system temperatures above the Tm of DMPC (i.e. 3.7 K),

although this apparent

difference lies within the uncertainty of the measurements.
We note in passing that it is possible, in principle, to treat the transient
temperature change associated with point source heating with the well-known timedependent partial differential heat equation.

160

We have chosen not to present an

analysis of our data in the context of the heat equation because the results calculated in
this manner are highly dependent on assumed input parameters, and the qualitative result
is always that the heat deposited into the system dissipates very rapidly. There is little or
107

no insight to be gained into the intermolecular processes responsible for the effects we
observe using this model.
We have discussed our data in the context of a transient change in the temperature
of the chromophore immediate environment. Because of the time scales involved in both
the energy dissipation process and in our measurement of the associated system changes,
we sense an average temperature change rather than an instantaneous change. The time
scale of energy dissipation in this system is determined by the time required to transfer
energy from the chromophore to the vibrational modes of the surrounding lipid acyl
chains. The nonradiative relaxation from the perylene S2 state to the S1 state will
proceed within a picosecond.

161

The heating process for the sample as a whole will

occur over the duration of the excitation pulse, but for any given chromophore, the
relaxation process will be on the order of a ps or less. This conservative estimate of the
relaxation time includes IVR (intramolecular vibrational relaxation) within the S2
manifold and S2  S1 relaxation. Once the excitation energy has been deposited into the
lipid bath modes, the rate at which intermolecular vibrational energy transfer proceeds is
determined by the nature of the energy transfer process and the efficiency of
intramolecular vibrational relaxation within the lipid acyl chains. The former process,
assuming resonant energy transfer dominates, will proceed on the ca. 10 ps time
136-137, 162-165

scale,

and the latter process, which depends on the density of vibrational

states within the acyl chains, may take a similar period of time, at most. Thus it is fair to
estimate that the energy dissipation into the bath is complete within less than 20 ps, and

108

quite possibly much less than 20 ps. This time scale is fast relative to the time scale
associated with perylene reorientation. Our experimental data show that the reorientation
time of perylene varies between 165 ps and 370 ps, depending on the state excited and
the temperature of the system (Table 5.2). If we average the transient temperature change
over several hundred ps, it is clear that the instantaneous temperature change associated
with excitation is much higher than the ΔT values we report. While it is tempting to
calculate the transient temperature change based on the time scales involved, doing so is
not straightforward because of changes in the order of the lipid acyl chains. Specifically,
the transient temperature change can drive phase transition(s) in the lipid acyl chain
region, leading to a more disordered system. The persistence time for the thermally
induced disorder may be significantly longer than the actual thermal transient used to
create the disorder and, as a consequence, the chromophore is sensing a disordered
environment that does not re-organize fully on the time scale of the measured
reorientation. Thus, owing to the time scales and intrinsic molecular-scale complexity of
the systems we have examined, it is not possible to quantitate either the maximum ΔT
achieved or the duration of the thermal transient associated with the S2  S1 relaxation
of perylene.

109

Conclusions
We have examined the time-resolved fluorescence anisotropy decay behavior of
perylene in DMPC vesicles as a function of system temperature, lipid phase and the state
to which perylene is excited. The excitation state-dependent reorientation dynamics of
perylene reveal the effect of the deposition of excess excitation energy on the lipid
bilayer acyl chain environment in closest proximity to the chromophore. Our findings are
consistent with a transient temperature change of ca. 4 K for excitation of perylene to its
S2 state, with there being a subtle difference in the temperature change for the lipid gel
and fluid phases, albeit not outside of the experimental uncertainty. It appears that
thermal energy may be transferred to lipid (bath) molecules somewhat more efficiently
when the bath is relatively well organized, and that for phases of comparatively low
order, such as the lipid fluid phase, the absence of organization can potentially limit the
efficiency of vibrational intermolecular energy transfer on the timescale of the
reorientation experiment. Our data provide insight into the effect of transient heating in
lipid bilayer systems and suggest that the least perturbative way to perform spectroscopic
experiments in any medium is to excite the chromophore to the same excited state from
which the experimental signal is derived.

110

CHAPTER 6
Consequences of Transient Heating on the Motional Dynamics of
Cholesterol-Containing Phospholipid Vesicles

Introduction
Biological plasma membranes are dynamic fluid structures that are composed of
an assortment of chemical constituents, which work together to promote the regulation
and function of the cell.

42

The interactions of these chemical constituents, which include

phospholipids, glycerophospholipids, sphingolipids, sterols, and proteins, are crucial to
the organization of the membrane and the proper function of integral proteins. Among
the primary constituents in plasma membranes are phospholipids and cholesterol. It is
well-established that the amount of cholesterol present in the membrane supports cell
structure, dynamics, and function.

166-168

It is for this reason that a number of studies on

synthetic bilayer structures have focused on comparatively simple systems that contain
phospholipid, sphingomyelin, and cholesterol.

153, 169-171

Cholesterol is an amphipathic

molecule (Fig. 6.1), and the β-hydroxyl group on its A ring produces an orientational
preference in bilayer structures, with the fused-ring system aligning parallel to the lipid
acyl chains and the β-hydroxyl group directed toward the bilayer headgroup region.

166

The incorporation of cholesterol into phospholipid bilayers is believed to disrupt the
interactions between neighboring lipids, which leads to the creation of an intermediate

111

―liquid-ordered‖ phase and consequently alters the lipid’s main phase transition
172

temperature (Tm).

Bilayers that contain cholesterol exhibit regions of phase separation,

173

and are

known to exist in a liquid-ordered phase at low temperatures and a liquid-disordered
phase at higher temperatures.

174

The temperature at which this phase transition occurs

depends sensitively on the functionality of the lipids comprising the bilayer and the
amount of cholesterol present. This phase transition is analogous to the ―melting‖ gel-tofluid phase transition seen in bilayers comprised solely of phospholipid. The addition of
cholesterol to a phospholipid bilayer typically leads to a decrease in the observed phase
transition temperature, consistent with the introduction of a loss of molecular-scale
organization within the system.
Phase separation in model lipid bilayer systems has been associated with the
notion of ―lipid rafts‖, the existence of which in plasma membranes remains a subject of
175-179

debate.

It is clear that in model systems the presence of cholesterol in

phospholipid bilayers mediates lateral organization and local motional freedom of
individual phospholipids within the membrane. In most bilayer systems cholesterol is
believed to act in concert with other constituents. Cholesterol and sphingomyelin, in
particular, are thought to play a cooperative role in mediating bilayer organization, as
well as protein function in plasma membranes.

179

We are interested in learning about the consequences of local temperature change
within bilayer structures containing cholesterol.

We have recently investigated the

structural effects of local temperature change in bilayers comprised entirely of a
112

180

phospholipid, DMPC.

In the work we present here, we examine the effects of

transient heating on vesicles comprised of phospholipid and cholesterol. We measure the
rotational diffusion dynamics of perylene imbedded in the non-polar region of the
vesicles, as a function of electronic excitation energy. Perylene functions as a local
―radiator‖ of thermal energy in these systems. For S1 ← S0 excitation, little excess
energy is released nonradiatively, owing to the high fluorescence quantum yield and
small Stokes shift of this chromophore. For S2 ← S0 excitation, rapid nonradiative
relaxation (ca. 1 ps)

28

from the S2 manifold to the S1 manifold deposits ca. 1.92 eV of

thermal energy into the immediate environment of the chromophore, resulting in a
transient change in local temperature and, possibly, organization within the bilayer.
Consistent with our earlier work,

180

the reorientation dynamics of perylene in

DMPC/cholesterol vesicles are excitation state-dependent, and from this state dependence
we are able to estimate the transient temperature change in the chromophore local
environment, averaged over its reorientation time. The average temperature change we
observe is on the order of ca. 7 K below what we believe to be the liquid-ordered to
liquid-disordered phase transition, and ca. 13 K above this phase transition.

The

reorientation data suggest a significant change in the chromophore local environment at
the onset of a phase transition, with a lower degree of organization and, consequently,
molecular order above the phase transition.

113

Experimental Section
Vesicle Preparation. The phospholipid 1,2-dimyristoyl-sn-phosphatidylcholine
(14:0 DMPC) dissolved in chloroform was purchased from Avanti Polar Lipids, Inc.
(Alabaster, AL) and was used without further purification. Cholesterol powder (ovine
wool, > 98 %) was also purchased from Avanti Polar Lipids, Inc. and used as received.
Perylene (99+ %) was purchased from Sigma-Aldrich (Milwaukee, WI) and used without
further purification. For each sample, DMPC and cholesterol were first mixed in a 2:1
ratio of 66.7 mol % and 33.2 mol %, respectively, with 0.1 mol % perylene added and the
chloroform solvent was evaporated under vacuum. 10 mM Tris® buffer (Fluka) was then
added to each sample, making the final sample concentration 1 mg/mL. Tris ® buffer was
prepared with water from a Milli-Q Plus water purification system (Millipore, Bedford,
MA),

buffered

to

pH

8,

and

purged

with

Ar

(g)

prior

to

use.

Lipid/cholesterol/chromophore mixtures were mixed with five freeze-thaw-vortex cycles.
Each cycle consisted of freezing the solution by immersion in N2 (l) for 5 min., followed
by thawing of the sample in a 60⁰C water bath for 5 min., then vortexing the mixture for
2 min. The resulting solutions were extruded eleven times through a polycarbonate
membrane (Whatman) with a nominal pore diameter of 100 nm using a mini-extruder
apparatus (Avanti Polar Lipids, Inc.) to produce unilamellar vesicles.
Time-Resolved Fluorescence Measurements.

103

Fluorescence lifetime and

anisotropy decay data were acquired using a time-correlated single photon counting
1

(TCPSC) instrument. This system has been described in detail elsewhere, and we
briefly summarize the pertinent features below. The light source is a CW passively

114

mode-locked, diode-pumped Nd:YVO4 laser (Spectra Physics Vanguard) that produces
13 ps pulses at 1064 nm at a repetition rate of 80 MHz. The output is frequency doubled
and tripled, providing 2.5 W average power with the same pulse characteristics at 355 nm
and 532 nm. The second and third harmonic outputs of the Nd:YVO4 laser excite cavitydumped dye lasers (Coherent 702-2), which can operate in the range of 430 nm to 850
nm, producing 5 ps pulses with a repetition rate of 4 MHz. For excitation of the S 1 ← S0
transition at 430 nm, the dye laser was operated with Stilbene 420 dye (Exciton) to
produce the excitation wavelength. For the S2 ← S0 transition at 260 nm, Coumarin 500
dye (Exciton) was used to produce 520 nm pulses that were frequency doubled using a
Type I BBO crystal to produce 260 nm pulses. For both excitation wavelengths, the laser
output is linearly polarized with a polarization extinction ratio of ca. 100:1.

The

polarization of the 260 nm pulse was rotated to vertical using a UV-grade quartz
polarization rotating Fresnel rhomb pair (CVI), which ensured an extinction ratio >
500:1. The fundamental excitation pulse from the dye laser is divided, with one portion
of the pulse directed into a reference photodiode (Becker & Hickl PHD-400-N), and the
other portion directed into the sample. Emission is collected at right angles to the
excitation axis and directed to the detectors using a 40x reflecting microscope objective
(Ealing). The collected emission is separated into polarization components parallel (0⁰)
and perpendicular (90⁰) to the vertically polarized excitation pulse using a polarizing
cube beam splitter. The polarized signal components are simultaneously detected with
microchannel plate photomultiplier tubes (MCP-PMT, Hamamatsu R3809U-50), each
equipped with a subtractive double monochromator (Spectral Products CM-112). The
115

detection electronics (Becker & Hickl SPC-132) record the parallel and perpendicular
transients separately, yielding ca. 30 ps response functions for each detection channel.
The instrument is controlled using an in-house written LabVIEW® (National Instruments)
program on a PC. The sample temperature was regulated to ± 0.1° C with a watercirculating bath (Neslab RTE-110) connected to a temperature-controlled brass cell jacket
which held the sample cuvette. All samples were allowed to equilibrate thermally for 10
minutes prior to data acquisition.
Steady-State Fluorescence Measurements. Excitation and emission spectra were
acquired for our vesicle samples to determine the perylene band positions. We used a
Hitachi F-4500 fluorescence spectrophotometer to collect both the excitation and
emission spectra of perylene in our samples. The instrument was set to a spectral band
pass of 1 nm for both excitation and emission monochromators.

116

Results and Discussion
The purpose of this work is to study the effects of thermal energy dissipation in
lipid bilayer structures. Cholesterol in the membrane is critical to the regulation of
cellular processes.

49, 167, 181-183

Also important to the understanding of cellular

function and overall bilayer behavior is the mechanism by which thermal energy is
dissipated within the lipid matrix. We do know that thermal energy influences the
organization of the bilayer constituents,

25, 76, 184

but it remains unclear whether the

presence of a point source of thermal energy is capable of creating a local or longer range
change in that organization. In order to study this phenomenon in greater detail, we use a
fluorescent probe molecule to obtain information about its local environment. In this
case, the local environment of interest is the acyl chain region of a lipid bilayer, as
changes in this region can help to elucidate the structure, packing, and overall dynamics
of the bilayer. This is the area of the bilayer most greatly affected by the addition of
1

―impurities‖, such as ethanol or cholesterol, and by changes in the thermal energy of the
system.

180

The use of any molecule as a ―probe‖ can, in principle, lead to the acquisition of
information that is not necessarily reflective of the system of interest. This is because the
probe molecule is itself a perturbation to the system. In an attempt to minimize the
consequences of this limitation, relatively low concentrations of chromophore are
typically used. The concentration of perylene we use is sufficiently low that, for DMPC
vesicles less than 800 nm in diameter, we observe the same phase transition temperature
(Tm = 297 K) that has been measured using variable temperature NMR spectroscopy.
117

180

For this reason, we assert that the phenomena we report here are reflective of the
chromophore local environment formed by the lipids and cholesterol, and that the
chromophore functions as a largely non-disruptive probe.
As noted above, we have reported previously on the dissipation of thermal energy
in pure phospholipid bilayers using the same spectroscopic technique. In that work, we
observed a ca. 4 K change in local temperature associated with the nonradiative
relaxation of perylene from the S2 to S1 manifold.

We calculated the average

temperature change of the chromophore immediate environment using a model that
relates the excitation state-dependent change in the anisotropy decay time constant to a
change in local viscosity, and then related the change in viscosity to a change in
temperature.

180

The data we report here indicate that the addition of cholesterol to this

bilayer system reduces the efficiency of intermolecular vibrational energy transfer, thus
keeping the thermal energy in close spatial proximity to the probe and giving rise to a
larger measured temperature increase.
Perylene was chosen for these experiments because of our past experience with
this probe,

1, 25

and because it is a nonpolar polycyclic aromatic hydrocarbon (Fig. 6.1)

that partitions selectively into the hydrophobic region of the lipid bilayer The actual
location of the chromophore within the acyl chain region of the bilayer has been shown to
depend sensitively on the diameter of the extruded vesicles.

76

As the diameter of the

vesicle exceeds ca. 800 nm, perylene becomes insensitive to the gel-to-fluid phase
transition in DMPC unilamellar vesicles, suggesting its relocation to the bilayer
midplane.

76

In small vesicles (< 800 nm diameter), the gel-to-fluid phase transition is
118

Perylene

Cholesterol

normalized intensity (a.u.)

4.0

14:0 DMPC

3.5
3.0

S2 S1 excitation

2.5

S0 S1 emission

2.0
1.5

S1 S0 excitation

1.0
0.5

0.0
200 250 300 350 400 450 500 550 600
wavelength (nm)
Figure 6.1. Structures of 14:0 DMPC, perylene, cholesterol, and normalized
excitation and emission spectra of perylene in DMPC vesicles.

119

sensed, suggesting that perylene is most likely situated between the lipid acyl chain
moieties.

76, 180

Understanding the approximate location of the probe within the bilayer

acyl chains allows for a more meaningful interpretation of the fluorescence dynamics
measurements.
Two sets of experiments were performed to study the consequences of transient
heating in DMPC/cholesterol vesicles. The first was to measure the rotational diffusion
dynamics of perylene excited to its S1 state ( S1 = 430 nm) at a series of system
temperatures, and the second was to perform the same measurements, except exciting the
chromophore to the S2 state ( S 2 = 260 nm). For all measurements, emission from the
S1 state is monitored, and excitation state-dependent differences in the chromophore
reorientation dynamics are related to transient heating in the bilayer.
For fluorescence anisotropy measurements, the polarized fluorescence transients
are combined to produce the induced orientational anisotropy function, R(t),

R(t ) 

I (t )  I  (t )

(6.1)

I (t )  2 I  (t )

where I (t ) is the emission transient polarized parallel to the vertically polarized
excitation pulse and I  (t ) is the emission transient polarized perpendicular to the
excitation pulse. For these experiments, the pertinent chemical information is contained
in the number of exponential decays exhibited in R(t) and the time constant of each
decay. There are several theoretical treatments that relate the decay functionality of R(t)
38, 139-

to the Cartesian components of the chromophore rotational diffusion constant, D.

120

140

Perylene is known to produce either a one- or two-exponential component

fluorescence anisotropy decay, depending on its local environment. In this work, we
have observed a two-component anisotropy decay for excitation of the S1 ← S0
( S1 = 430 nm) transition at all temperatures studied (Fig. 6.2b). Excitation of the
S2 ← S0 transition ( S 2 = 260 nm) produces a two-component decay for temperatures of
278 K and below (Fig. 6.3b), and a one-component decay for temperatures above 278 K
(Fig. 6.4b). We consider the implications of these findings and the chemical information
content of these data next.
Chuang and Eisenthal developed theoretical expressions relating fluorescence
anisotropy decay functionality to the Cartesian components of the diffusion constant,
D,139
D

Dx  D y  Dz

(6.2)

3

It is the magnitude of D and relative values of Dx, Dy, and Dz that reflect the interactions
with and constraints imposed upon the chromophore by its local environment. Because
of the number of independent variables in the model and the number of experimental
observables, it is necessary to assume that molecules reorient according to the general
condition Di ≠ Dj = Dk. It is not possible to unambiguously resolve three Cartesian
components of D. It is most common to interpret reorientation data in the context of the
chromophore reorienting as either a prolate or oblate rotor,

139

but these representations

can lead to ambiguity in the interpretation of anisotropy decay data under certain

121

3500

a

parallel

3000

counts

2500
2000

perpendicular

1500
1000
500
0
0

1000 2000 3000 4000 5000 6000 7000 8000 9000

time (ps)
0.15

b

R(t)

0.10

0.05
0.00

-0.05
0

1000 2000 3000 4000 5000 6000 7000 8000 9000

time (ps)

Figure 6.2. (a) Experimental data for perylene reorientation in 100 nm diameter
DMPC vesicles at 270.5 K, excited at 430 nm (S1 ← S0). (b) Induced
orientational anisotropy decay function for the data shown in panel (a), along with
the best-fit function to the data.

122

5000

perpendicular

a

counts

4000

parallel

3000
2000
1000
0
0

1000 2000 3000 4000 5000 6000 7000 8000 9000

time (ps)
0.04

b

0.03
0.02

R(t)

0.01
0.00
-0.01
-0.02
-0.03
-0.04
0

1000 2000 3000 4000 5000 6000 7000 8000 9000

time (ps)

Figure 6.3. (a) Experimental data for perylene reorientation in 100 nm diameter
DMPC vesicles at 273 K, excited at 260 nm (S2 ← S0). (b) Induced
orientational anisotropy decay function for the data shown in panel (a), along
with the best-fit function to the data.

123

3500
perpendicular

3000

a

counts

2500
parallel

2000
1500
1000
500
0
0

1000 2000 3000 4000 5000 6000 7000 8000 9000

time (ps)
0.04

b

0.03
0.02
0.01

R(t)

0.00
-0.01
-0.02
-0.03
-0.04
-0.05
0

1000 2000 3000 4000 5000 6000 7000 8000 9000

time (ps)
Figure 6.4. (a) Experimental data for perylene reorientation in 100 nm
diameter DMPC vesicles at 283 K, excited at 260 nm (S2 ← S0). (b)
Induced orientational anisotropy decay function for the data shown in panel
(a), along with the best-fit function to the data.

124

circumstances, such as those we encounter here. It is the relative orientations of the
absorbing and emitting transition dipole moments with respect to the major axis of
rotation that defines the types of possible rotors. We designate these rotors as Type I,
Type II, and Type III,

76

and define them below.

In addition to the perylene S2 ← S0 transition being 1.92 eV higher in energy than
the S1 ← S0 transition, these transitions are polarized perpendicular to one another.
Thus, using Chuang and Eisenthal’s treatment of R(t), there will be different anisotropy
decay equations to describe each excitation condition. For perylene, we define the long
in-plane axis as the x-axis, the short in-plane axis as the y-axis, and the out-of-plane axis
perpendicular to the π-system plane as the z-axis. A Type I rotor has a unique axis of
rotation coincident with the chromophore emission transition moment (the axis of
observation). For perylene, both the absorption and emission transition dipole moments
for the S1 ↔ S0 transition lie along the x-axis.
nomenclature for a Type I rotor,

139

Using Chuang and Eisenthal’s

where q is the unit vector describing the orientation

of the absorbing transition dipole moment and  is the unit vector describing the emitting
2

2

transition dipole moment, qx = 1, x = 1 and Dx ≠ Dy = Dz. The anisotropy decay
function is:

R(t )  0.4exp(6Dzt )

(6.3)
2

2

For the perylene S2 ← S0 transition, which is y-axis polarized (qy = 1, x = 1), the
functional form of the anisotropy decay is:

R(t )  0.2exp(6Dzt )

(6.4)
125

Note that Eq. 6.4 predicts a negative zero-time anisotropy because the absorption
transition is polarized perpendicular to the emission transition.
A Type II rotor is defined as having its unique axis of rotation oriented
perpendicular to its transition dipole moment(s).

For perylene this corresponds to

Dz ≠ Dx = Dy. For excitation of the S1 ← S0 transition,

R(t )  0.1exp((2Dx  4Dz )t )  0.3exp(6 Dxt )

(6.5)

And for excitation of the S2 ← S0 transition,

R(t )  0.1exp(6Dxt )  0.3exp((2Dx  4Dz )t )

(6.6)

A Type III rotor is defined as having its unique axis of rotation perpendicular to
the emission transition dipole moment, but not necessarily perpendicular to the absorbing
transition dipole moment. For perylene, this condition corresponds to Dy ≠ Dx = Dz. For
excitation of the S1 ← S0 transition, where the absorbing and emitting transition
moments are oriented perpendicular to the unique rotational axis,

R(t )  0.3exp((4Dy  2Dz )t )  0.1exp(6 Dzt )

(6.7)

And for the S2 ← S0 transition, where the absorption transition moment is parallel to the
unique rotational axis,

R(t )  0.2exp(6Dzt )

(6.8)

For perylene, a Type I rotor is characterized by a single exponential anisotropy decay for
both excitations, a Type II rotor is characterized by a two-component exponential
anisotropy decay for both excitations, and a Type III rotor is characterized by a two-

126

component anisotropy decay for S1 ← S0 excitation and a one component anisotropy
decay for S2 ← S0 excitation.
In a previous report, we recovered two-component exponential anisotropy decays
when exciting the perylene S1 ← S0 transition in DMPC vesicles at a series of
temperatures.

1, 180

We also have recovered two-component anisotropy decays for

perylene in DMPC/cholesterol vesicles (S1 ← S0 excitation) (Fig. 6.2b). In our earlier
report, we predicted that S2 ← S0 excitation of perylene would give rise to a twocomponent exponential decay, consistent with this chromophore behaving as a Type II
rotor in DMPC vesicles, both above and below the gel-to-fluid phase transition.
However, due to the presence of a fluorophore contaminant in our samples, we were only
able to resolve one exponential decay component. It is also important to note that the
functionality of R(t) did not change with respect to temperature for these experiments.
This contamination issue has since been resolved, and we expect perylene to behave as a
Type II rotor upon S2 ← S0 excitation. In the work we report here, we observe a
temperature-dependent anisotropy decay for S2 ← S0 excitation.

Specifically, for

temperatures of 278 K and lower, we observe a two-component anisotropy decay,
consistent with a Type II rotor. For temperatures above 278 K, however, we observe a
one component anisotropy decay for S2 ← S0 excitation. This change is indicative of a
transition from a Type II rotor to a Type III rotor.

127

Because of the change in rotor shape observed for S2 ← S0 excitation at
temperatures above 278 K, and the transient heating effect associated with this excitation
(vide infra), it is not possible to discern unambiguously whether S1 ← S0 excitation data
manifest the analogous change of rotor shape at the same system temperature. Eqs. 6.5
and 6.7 are similar to one another, with the only difference being the exponential
prefactors, indicating a transposition of the unique axis of rotation from the z- to the yaxis. The quantitation of anisotropy prefactors is not as reliable as the measurement of
the decay time constants, rendering the resolution of the change of rotor type somewhat
less apparent for S1 ← S0 excitation. Regardless of this uncertainty, it is clear that
perylene in our bilayers undergoes a change from a Type II to a Type III rotor in the
vicinity of 278 K.
For temperatures below 278 K, the Type II rotor behavior we observe is
consistent with other reports of perylene reorientation in lipid bilayers.

1, 76, 185

Above

278 K we recover anisotropy decay functionalities that are different for the two
excitations, consistent with perylene reorienting as a Type III rotor. The change of
unique rotational axis from Dz to Dy reflects a substantial alteration in the environment of
the reorienting chromophore.
Assuming that at temperatures below 278 K perylene behaves as a Type II rotor,
all the components of the diffusion constant can be extracted from the measured
reorientation times (Table 6.1) using Eqs. 6.9 and 6.10, and the model condition that
Dz ≠ Dx = Dy.

128

Table 6.1. Measured reorientation times as a function of system temperature.
λ

S1

= 430 nm, S1 ← S0

λ

S2

= 260 nm, S2 ← S0

Temp. (K)

τOR1 (ps)

τOR2 (ps)

τOR1 (ps)

τOR2 (ps)

268

543 ± 27

2550 ± 14

405 ± 18

3102 ± 56

270.5

548 ± 16

2401 ± 66

445 ± 24

2756 ± 78

273

546 ± 12

2255 ± 40

435 ± 27

2531 ± 80

278

489 ± 11

2099 ± 35

353 ± 21

2224 ± 64

280.5

387 ± 56

2063 ± 83

1277 ± 57

-

283

405 ± 8

2014 ± 42

1167 ± 78

-

288

340 ± 17

1871 ± 51

1036 ± 51

-

Errors are ±1σ for at least four individual determinations at each temperature.

129

 OR1 

1
2 Dx  4 Dz

(6.9)

 OR 2 

1
6 Dx

(6.10)

The aspect ratio, Dmajor/Dminor, (Dz/Dx for a Type II rotor) can be calculated for each
temperature (Table 6.2). The aspect ratio is related to the nature of the surrounding
chromophore environment (i.e. the acyl chains). For example, a chromophore in an
unhindered environment, such as a continuum solvent, would exhibit reorientation about
the x-, y-, and z-axes, with the relationship between Dx, Dy, and Dz being determined by
the structure of the chromophore. For a spherical rotor, Dmajor/Dminor ≈ 1. The larger
the aspect ratio, the more influence is being exerted on the chromophore by its local
environment. Our experimental data for temperatures below 278 K indicate Dz/Dx ~ 6
for S1 excitation and ~ 9 for S2 excitation. This change in aspect ratio for the two
excitation conditions indicates that the chromophore is becoming more anisotropic as a
result of the transient heating effect and the subsequent change in local temperature. We
believe that at system temperatures in the vicinity of 278 K, our system is nearing a phase
transition, and that the excess excitation energy dissipated by perylene is sufficient
enough to initiate a phase transition in these systems. This phase transition, and its
consequent structural reorganization, could then trigger a change in the location of the
chromophore within the bilayer, presumably from between the acyl chains to the bilayer
midplane. A past study demonstrated that it is possible for perylene to relocate from the
acyl chains to the midplane.

76

Perylene molecules located in this region of the bilayer

130

would sense a local environment that is structurally much different than one located
between acyl chain moieties, which could explain why the aspect ratios for the
chromophore after S2 excitation are more anisotropic than those for S1 excitation. The
role of cholesterol in mediating the dynamics of these systems is yet to be determined.

For temperatures above 278 K, perylene reorients as a Type III rotor (Eqs. 6.7 and
6.8), and we can use Eqs. 6.11 and 6.12, with the measured reorientation time constants
for S1 excitation (Table 6.1) to solve for the Cartesian components of D.

 OR1 

1
4 D y  2 Dz

(6.11)

 OR1 

1
6 Dz

(6.12)

Again, the aspect ratio, Dmajor/Dminor, (Dy/Dz for a Type III rotor) can be calculated
using this set of equations (Table 6.2).

131

Table 6.2. The rotational diffusion constant, aspect ratio Dmajor/Dminor, and calculated
local viscosity as a function of temperature.
λ

S1

= 430 nm, S1 ← S0

6.6 ± 0.4

D
(MHz)
186 ± 8

η
(cP)
10.3 ± 0.4

270.5

6.1 ± 0.3

187 ± 4

273

5.7 ± 0.1

278

λ

S2

= 260 nm, S2 ← S0

11.0 ± 0.6

D
(MHz)
233 ± 9

η
(cP)
8.2 ± 0.3

10.4 ± 0.2

8.8 ± 0.6

218 ± 10

8.9 ± 0.4

190 ± 4

10.3 ± 0.2

8.1 ± 0.6

226 ± 14

8.7 ± 0.5

5.9 ± 0.1

210 ± 4

9.5 ± 0.2

9.0 ± 0.5

274 ± 15

7.3 ± 0.4

280.5

7.7 ± 1.4

259 ± 31

7.8 ± 0.9

7.7

a

421 ± 19

4.8 ± 0.2

283

7.0 ± 0.2

247 ± 4

8.2 ± 0.1

7.0

a

439 ± 34

4.7 ± 0.4

288

7.8 ± 0.6

290 ± 12

7.1 ± 0.3

7.8

a

525 ± 27

3.9 ± 0.2

Temp.
(K)
268

Dmaj./Dmin.

Dmaj./Dmin.

a

The values of the aspect ratio (Dmajor/Dminor) for the S2 ← S0 transition were assumed
to remain constant from those at the corresponding temperatures determined for the
S1 ← S0 transition. Errors are ±1σ for at least four individual determinations at each
temperature.

132

For excitation of the S2 ← S0 transition at temperatures > 278 K, perylene
behaves as a Type III rotor (Eq. 6.8), allowing the recovery of one component of the
diffusion constant (Dz) according to Eq. 6.13.

 OR 

1
6 Dz

(6.13)

To calculate D (Eq. 6.2) from our S2 excitation data above 278 K, we assume that
Dmajor/Dminor is the same as for S1 excitation (Table 6.2). While it may be tempting to
use the same change in Dmajor/Dminor as for the Type II rotor (temperatures ≤ 278 K),
we believe this assumption would not be justified based on the bilayer phase transition
and, more importantly, the change in rotor type. These data indicate that excitation of the
S2 ← S0 transition results in larger rotational diffusion constants than S1 excitation for
perylene. We understand this phenomenon in terms of nonradiative energy transfer
between perylene and its immediate environment.

155, 180

The change in the rotor shape

from a Type II to a Type III rotor at ~ 278 K is likely the result of a change in the
chromophore local environment. As perylene relaxes nonradiatively from the S 2 state to
the S1 state, excess excitation energy released into the surroundings can alter the
organization and viscosity of the lipid medium. We consider next the extraction of
approximate average transient temperature changes that result.
The rotational diffusion constant, D, is related to the viscosity of the local
environment of the chromophore through the modified Debye-Stokes-Einstein (DSE)
equation.

39, 143

This model describes the rotational motion of a chromophore in a
133

continuum solvent, which is not the case for perylene imbedded in lipid bilayers. Despite
this limitation, we apply the DSE model to our data. Any errors that arise from the
application of the model to our data will apply to all of the data uniformly, so that
excitation-dependent changes in D will accurately reflect changes in . The modified
Debye-Stokes-Einstein equation is

39, 143

1
Vf

6 D k BTS

(6.14)

In this equation η is the viscosity of the bath (cP), V is the hydrodynamic volume of the
3

chromophore (Å ), f is a frictional term used to account for interactions between the
solvent and solute molecules, kB is the Boltzmann constant, T is the system temperature,
and S is a solute shape factor calculated using Perrin’s equations

39

non-spherical (ellipsoidal) shape of the reorienting chromophore.
perylene,

158

S = 0.70 for perylene,

145

to account for the
3

V = 225 Å for

and f is taken to be 1 for the frictional coefficient.

The DSE equation contains only two terms that depend on temperature: solvent bulk
viscosity (η) and system temperature (T). Knowing the temperature dependence of the
viscosity of the lipid bilayer and the amount of excess excitation energy that is dissipated
by the chromophore is vital for accurate interpretation of our data.
The first step in the determination of the temperature dependence of the lipid
bilayer viscosity is to use the DSE equation (Eq. 6.14) to calculate the bulk viscosity of
the DMPC/cholesterol lipid bilayers at each experimental system temperature and each
excitation (S1 and S2). We can then determine the change in bulk viscosity (Δη) by

134

difference (Table 6.2). To determine the amount of excitation energy released into the
surrounding bath, we simply calculate the difference in energy between the two
excitation transitions according to

E

hc

(6.15)



So, for the S2 ← S0 transition, λex = 260 nm and Eex = 4.77 eV, and for the S1 ← S0
transition, λ0-0 = 435 nm and E0-0 = 2.85 eV.

Therefore, 1.92 eV of energy is

nonradiatively dissipated into the surrounding bath of lipid acyl chains. To calculate the
change in local temperature (ΔT) that results from the nonradiative dissipation of ca. 1.92
eV of energy into the lipid environment surrounding the chromophore, we will use a
model for the temperature dependence of the lipid acyl chain viscosity.
The chemical identity of the medium plays a central role in the temperature
dependence of the viscosity, and this relationship has been determined experimentally for
a variety of liquids.

159

The temperature dependence of DMPC/cholesterol lipid bilayer

viscosity is, however, not known.

150

It was found previously that the lipid acyl chain

environment sensed by perylene in DMPC vesicles had a viscosity similar to those of
25

normal alcohols at temperatures near Tm.

Although we measure dynamics at

temperatures lower than the expected Tm of DMPC, the viscosity data reported in Table
6.2 show what appear to be three distinct viscosity ranges for both excitations, which we
refer to as ―phases‖. The first phase is observed from 268 – 273 K, a second one at 278
K, and a third from 280.5 – 288 K. It is important to note that the viscosity values
calculated for the S1 ← S0 transition are significantly larger than those determined for the
135

S2 ← S0 transition at all measured temperatures. These data are consistent with a phase
transition in our bilayer at ca. 278 K, which is sensed by the rotational diffusion
dynamics of perylene.
To estimate the overall temperature change (ΔT) that corresponds to the measured
159

change in local viscosity (Δη), we use Eq. 6.16, a parameterized fit relating η to T.

T

A



 B  C

(6.16)

A, B, and C are empirical constants. We do not have an independent measure of the
viscosity of DMPC acyl chains at ambient temperature. We estimate the behavior of the
lipid acyl chains to be similar to that of normal alcohols.

Using the viscosities

determined for the first excited state transition (S1 ← S0) as our reference, we estimate
the viscosity of the DMPC acyl chain region to be similar to that of n-decanol (η = 10.9
cP) over the range of 268 – 273 K, n-nonanol (η = 9.1 cP) at 278 K, and n-octanol (η =
7.3 cP) above 278 K.

159

By substituting the viscosity of the n-alcohol (at 298 K) for the

lipid acyl chain region viscosity, and assuming that the local viscosity of the acyl chains
has a similar temperature dependence to that of the alcohols, we can use the expression
below to calculate ΔT,
d
 2

dT T 0 A   2 B

(6.17)

where T0 is the ambient temperature of the DMPC/cholesterol vesicle solution. In Eq.
6.17, A and B are literature values used along with the viscosities of the normal alcohols
(n-decanol, n-nonanol, and n-octanol) and the experimentally determined values of Δη ~
dη to solve for dT ~ ΔT (Table 6.3). ΔT is the estimated average amount of transient
136

Table 6.3. Experimental reorientation times and calculated changes in local viscosity and
temperature associated with internal conversion of perylene.
Temp. (K)

η (S1) (cP)

η (S2) (cP)

Δη (S2 – S1) (cP)

ΔT (K)

268.0

10.3 ± 0.4

8.2 ± 0.3

-2.1 ± 0.5

7±2

270.5

10.4 ± 0.2

8.9 ± 0.4

-1.5 ± 0.5

5±2

273.0

10.3 ± 0.2

8.7 ± 0.5

-1.6 ± 0.5

6±2

278.0

9.5 ± 0.2

7.3 ± 0.4

-2.2 ± 0.5

8±2

280.5

7.8 ± 0.9

4.8 ± 0.2

-3.0 ± 0.9

12 ± 4

283.0

8.2 ± 0.1

4.7 ± 0.4

-3.5 ± 0.4

14 ± 2

288.0

7.1 ± 0.3

3.9 ± 0.2

-3.2 ± 0.4

13 ± 2

Errors are ±1σ for at least four individual determinations at each temperature.

137

heating that occurs to the lipid acyl chains.

155

At temperatures below 278 K the average

value of ΔT is ca. 7 K, and at temperatures above 278 K the average ΔT is ca. 13 K.
These results indicate that we are either detecting a phase transition in the temperature
range studied, or that the dissipated transient heat is altering the local environment of the
chromophore in such a way that perylene is sensing a change in its location within the
bilayer. This explanation is consistent with our finding that perylene changes from a
Type II rotor at temperatures below 278 K to a Type III rotor at temperatures above 278
K.

138

Conclusion
We have examined the time-resolved fluorescence anisotropy decay behavior of
perylene in vesicles comprised of DMPC and cholesterol as a function of both system
temperature and chromophore excitation state (Table 6.1).

The rotational diffusion

constants we have acquired demonstrate that the dynamics of the chromophore are
excitation state-dependent (Table 6.2). The state dependence arises from the deposition
of excess excitation energy into the surrounding lipid bath.

Rotational diffusion

measurements allow for estimation of the viscosity of the lipid bilayer in the vicinity of
the chromophore.

For the bilayers we consider here, which contain both the

phosphocholine lipid DMPC and cholesterol, we observe a phase transition in the vicinity
of 278 K, based on a distinct change in the reorientation dynamics of perylene at this
temperature. We infer from these data a change in the unique axis of rotation for
perylene as a function of bilayer temperature.

This unexpected finding suggests a

substantial change in the chromophore local environment.

Such a change was not

observed for lipid bilayers comprised solely of DMPC, indicating that cholesterol has a
role in this behavior. Using our reorientation data, we estimate the transient temperature
change of the lipid local environment due to the dissipation of excitation energy to be ca.
7 K for temperatures below 278 K and ca. 13 K for temperatures above 278 K. These
data suggest that thermal energy is transferred to the surrounding solvent bath (lipid acyl
chains) more efficiently at lower temperatures (Table 6.3), consistent with the lipid
bilayer existing in a more ordered state below the phase transition.

139

CHAPTER 7
CONCLUSIONS
The ultimate long-term goal of this project is to incorporate functional
transmembrane proteins into an artificial lipid bilayer environment for the development
of a biomimetic sensor. Planar supported lipid bilayers act as excellent models of the cell
membrane, and are attractive platforms for the construction of membrane-based
biosensors. However, incorporating transmembrane proteins into supported lipid bilayers
while maintaining the proteins’ bioactivity, presents a challenge.

Transmembrane

proteins span the full thickness of the bilayer and require contact with the polar aqueous
environment at both terminal ends in order to function properly. In addition, a lipid
bilayer fused directly to a solid substrate causes the lipid membrane to lose its fluidity, a
property that is crucial for the activity of biological components. This loss of membrane
fluidity, and the lack of space for the incorporation of a functional transmembrane
protein, led us to consider an alternative lipid bilayer system in the form of unilamellar
lipid vesicles. The main advantage of using lipid vesicles over planar bilayers is that both
the interior and exterior interfacial regions of the bilayer are in direct contact with the
aqueous environment, so in theory, incorporating proteins is possible.
Before attempting to introduce proteins into lipid vesicles, however, it was of the
utmost importance for us to investigate the structural organization and dynamic behavior
of simple lipid vesicle systems. In order to interrogate our lipid vesicle systems, we used
fluorescent molecular probes or chromophores to investigate the interfacial region of the
bilayer. Chromophores act as ―reporter‖ molecules, and provide valuable information

140

regarding the local environment of the probe, such as rotational and translational
dynamics, membrane fluidity, and local viscosity.

We have designed a series of

experiments to examine the effects of global- and molecular-scale perturbations on the
organization and fluidity of the lipid vesicle by measuring the motional dynamics of
incorporated chromophore molecules. We used low concentrations of chromophores, so
as not to further disrupt the bilayer structure due to the presence of the probes.
Chapter 2 discusses our experiments on the hydrogen bonding interactions
between lipids with one of two unique headgroup functionalities using a headgrouptethered chromophore, and how these interactions influence the fluidity of the
systems.

154

These lipid systems consisted of either unilamellar vesicles or supported

bilayers comprised of phosphocholine (PC), phosphoglycerol (PG), or a mixture of the
two lipids. PC lipids are zwitterionic and are believed to experience hydrogen bonding
interactions between the phosphate group oxygen atoms of neighboring lipid
53-56

headgroups,

while PG lipids are anionic and are thought to form hydrogen bonds

between neighboring headgroups and other membrane constituents.

57-60

Our time-

resolved fluorescence anisotropy decay measurements show that the rotational diffusion
of lipids comprising vesicles is independent of headgroup functionality. However, our
fluorescence recovery after pattern photobleaching (FRAPP) measurements on solidsupported lipid bilayers show slower rates of translational diffusion in PG lipid bilayers
compared to mixed PC/PG lipid systems, indicating stronger headgroup-headgroup
interactions in systems composed primarily of PG. We reconcile these two bodies of data
by first considering that translational diffusion measurements sense the motion of the
141

entire lipid through the bilayer. This movement is mediated by the structure of the
bilayer and the strength of the interactions involving the headgroups as well as the acyl
chains.

It is thus not surprising that a headgroup-tethered chromophore would be

sensitive to a portion of the overall molecular motion. Conversely, the rotational motion
of tethered chromophores is more directly influenced by the bulk solvent molecules in
closest proximity to the chromophores, than by the hydrogen bonding interactions
between the headgroups.
In other experiments, we chose to investigate the dynamic behavior of the lipid
acyl chain regions of lipid vesicles because these regions are thought to undergo the most
structural and organizational changes as a consequence of chemical and thermal
perturbations. In Chapter 3, we used steady-state fluorescence techniques to measure the
fluorescence emission of pyrene excimers sequestered in the acyl chain regions of 1,2dimyristoyl-sn-phosphatidylcholine (DMPC) lipid vesicles, as a function of ethanol
concentration. Pyrene excimer formation is diffusion-limited, and therefore the more
fluid the local environment of the probe, the more intense the characteristic pyrene
excimer fluorescence emission band (~ 472 nm) becomes. These experiments confirm
the dependence of pyrene excimer formation on the bulk ethanol concentration, and
consequent fluidity of the bilayer. The steady-state data also show a deviation from the
otherwise linear relationship between ethanol concentration and excimer band intensity in
the vicinity of 0.6 M ethanol, which points to a structural reorganization of the lipid
bilayer.
Building on the steady-state fluorescence experiments discussed in Chapter 3, we
next used time-resolved fluorescence anisotropy decay measurements to study the effects
142

of solution phase ethanol concentration on the rotational motion of the chromophore
perylene imbedded in gel and fluid phases of DMPC lipid vesicles. These experiments
are presented in Chapter 4. The anisotropy decay data show that ethanol concentration
affects the organization of both gel and fluid phase lipid bilayers. In both cases, the
rotational diffusion dynamics of perylene molecules incorporated in the bilayer acyl
chain region experience a perturbation about their long in-plane rotational axis (Dx) for
ethanol concentrations in the vicinity of 0.6 M. Taken collectively, the steady-state and
dynamic fluorescence data dealing with the effects of ethanol on DMPC lipid vesicles
seem to indicate a structural phase transition occurring after ca. 0.6 M ethanol has been
added to the system, and there is evidence that this transition marks the onset of the
interdigitated phase in gel and fluid phase bilayers.
We also chose to examine the time-resolved fluorescence anisotropy decay
behavior of perylene in DMPC vesicles as a function of system temperature, lipid phase
and the state to which perylene is excited. The specific purpose of these experiments was
to study the effect of the deposition of excess excitation energy on the lipid bilayer acyl
chain environment in closest proximity to the chromophore. By exciting imbedded
perylene to its second excited electronic singlet state (S2) and comparing the recovered
fluorescence anisotropy decay(s) (emission from the S0 ← S1 transition) over a range of
bulk system temperatures to those obtained for excitation of the S1 state, the average
transient temperature change of the chromophore local environment was estimated to be
ca. 4 K, with there being a subtle difference in the temperature change for the lipid gel
and fluid phases, albeit not outside of the experimental uncertainty. Chapter 5 also
143

discusses how lipid phase and organization appear to influence the efficiency of thermal
energy dissipation. The gel phase is highly ordered and the lipid acyl chains are tightly
packed, which promotes vibrational intermolecular energy transfer from the ―radiating‖
chromophore molecule to the surrounding (bath) lipids. Chapter 6 expands on this work
by investigating the same excitation energy dissipation from the S2 state of perylene
imbedded in DMPC lipid systems with cholesterol. Our results appear to show a phase
transition in the vicinity of 278 K based on a distinct change in the reorientation
dynamics of perylene at this temperature – a change that is not observed in pure DMPC
lipid systems. We calculate the average transient temperature change to be on the order
of 7 K for system temperatures below 278 K, and ca. 13 K for temperatures greater than
278 K. Like the results of Chapter 5, these results indicate a more efficient transfer of
energy in more ordered systems (i.e. the gel phase vs. the fluid phase). It appears that
cholesterol plays an important role in regulating the fluidity and phase of model bilayer
systems analogous to its behavior in biological cell membranes. Having greater control
over the structural and fluid properties of the bilayer will allow us to successfully
incorporate biologically-active transmembrane proteins into our bilayer systems.
The results of our experiments have laid the groundwork for a greater
understanding of the dynamic behavior of lipid bilayer systems, and have provided us
with a means of manipulating the structural properties of simple lipid bilayer systems.
Our work with lipid vesicles has demonstrated the importance of understanding the
effects and extent of both global- and molecular-scale perturbations on the fluid behavior
of lipid bilayers. We believe that by using this knowledge we will be able to incorporate
functional transmembrane proteins into these lipid systems, and we expect that our
144

findings will ultimately lead to the successful development of a transmembraneincorporated membrane-based biosensor.

145

LITERATURE CITED

146

Literature Cited
1.
Pillman, H. A.; Blanchard, G. J., The Journal of Physical Chemistry B 2010, 114,
3840.

2.

Newman, J. D.; Turner, A. P. F., Biosensors and Bioelectronics 2005, 20, 2435.

3.

Wu, T.-Z., Biosensors and Bioelectronics 1999, 14, 9.

4.

Nikolelis, D. P.; Hianik, T.; Krull, U. J., Electroanalysis 1999, 11, 7.

5.

de Mendoza, D.; Cronan, J. E., Trends in biochemical sciences 1983, 8, 49.

6.

Cooper, R. A., Journal of Supramolecular Structure 1978, 8, 413.

7.
Xu, G.; Wu, Y.; Li, R.; Wang, P.; Yan, W.; Zheng, X., Chinese Science Bulletin
2002, 47, 1849.

8.
Pancrazio, J. J.; Whelan, J. P.; Borkholder, D. A.; Ma, W.; Stenger, D. A., Annals
of Biomedical Engineering 1999, 27, 697.

9.

Chaki, N. K.; Vijayamohanan, K., Biosensors and Bioelectronics 2002, 17, 1.

10.

Richter, R. P.; Bérat, R.; Brisson, A. R., Langmuir 2006, 22, 3497.

11.

Groves, J. T.; Dustin, M. L., Journal of Immunological Methods 2003, 278, 19.

12.

Sackmann, E., Science 1996, 271, 43.

13.

Domi ska, M.; Blanchard, G. J., Langmuir 2009, 26, 1043.

14.

Wagner, M. L.; Tamm, L. K., Biophysical Journal 2000, 79, 1400.

147

15.

Tanaka, M.; Sackmann, E., Nature 2005, 437, 656.

16.

Castellana, E. T.; Cremer, P. S., Surface Science Reports 2006, 61, 429.

17.
51.

Lösche, M.; Helm, C.; Mattes, H. D.; Möhwald, H., Thin Solid Films 1985, 133,

18.

Tamm, L. K.; McConnell, H. M., Biophysical Journal 1985, 47.

19.

Brian, A. A.; McConnell, H. M., Proc. Nat. Acad. Sci. USA 1984, 81, 6159.

20.
McConnell, H. M.; Watts, T. H.; Weis, R. M.; Brian, A. A., Biochimica Et
Biophysica Acta 1986, 864.

21.
Klee, B.; Duveneck, G. L.; Oroszlan, P.; Ehrat, M.; Widmer, H. M., Sensors and
Actuators B: Chemical 1995, 29, 307.

22.

Doherty, G. J.; McMahon, H. T., Annual Review of Biophysics 2008, 37, 65.

23.
Gawrisch, K., The Dynamics of Membrane Lipids. In The Structure of Biological
Membranes, 2nd ed.; Yeagle, P. L., Ed. CRC Press: New York, 2005.

24.
Pabst, G.; Amenitsch, H.; Kharakoz, D. P.; Laggner, P.; Rappolt, M., Physical
Review E 2004, 70, 021908.

25.
Koan, M. M.; Blanchard, G. J., Journal of Physical Chemistry B 2006, 110,
16584.

26.
Slavik, J., Fluorescent Probes in Cellular and Molecular Biology. CRC Press:
Ann Arbor, 1994.

27.

Valeur, B., Molecular Fluorescence. Wiley-VCH: New York, 2002.

148

28.

Kasha, M., Discussions of the Faraday Society 1950, 14.

29.
Acree, W. E.; Tucker, S. A.; Cretella, L. E.; Zvaigzne, A. I.; Street, K. W.; Fetzer,
J. C.; Nakasuji, K.; Murata, I., Applied Spectroscopy 1990, 44, 951.

30.

Chen, S.-H.; McGuffin, V. L., Applied Spectroscopy 1994, 48, 596.

31.

Karpovich, D. S.; Blanchard, G. J., J. Phys. Chem. 1995, 99, 3951.

32.
248.

Ray, G. B.; Chakraborty, I.; Moulik, S. P., J. Colloid Interface Sci. 2006, 294,

33.
Waris, R.; Street, K. W.; Acree, W. E.; Fetzer, J. C., Applied Spectroscopy 1989,
43, 845.

34.

Galla, H. J.; Sackmann, E., Biochimica Et Biophysica Acta 1974, 339, 103.

35.

Muenter, J. S.; Deutsch, J. L., Journal of Chemical Education 1996, 73, 580.

36.
Bergmann, W. L.; Dressler, V.; Haest, C. W. M.; Deuticke, B., Biochimica et
Biophysica Acta (BBA) - Biomembranes 1984, 772, 328.

37.

Debye, P., Polar Molecules Chemical Catalog Co.: New York 1929; p 84.

38.

Hu, C. M.; Zwanzig, R., Journal of Chemical Physics 1974, 60, 4354.

39.

Perrin, F., Journal de Physique et le Radium 1936, 7, 1.

40.

Chuang, T. J.; Eisenthal, K. B., Journal of Chemical Physics 1972, 57, 5094.

41.
Garcia, J. J.; Martinez-Ballarin, E.; Millan-Plano, S.; Allue, J. L.; Albendea, C.;
Fuentes, L.; Escanero, J. F., Journal of Trace Elements in Medicine and Biology 2005,
19, 19.
149

42.

Beney, L.; Gervais, P., Applied Microbiology and Biotechnology 2001, 57, 34.

43.
Lapinski, M. M.; Castro-Forero, A.; Greiner, A. J.; Ofoli, R. Y.; Blanchard, G. J.,
Langmuir 2007, 23, 11677.

44.
581.

Chan, Y. H. M.; Boxer, S. G., Current Opinion in Chemical Biology 2007, 11,

45.
M'Baye, G.; Mely, Y.; Duportail, G.; Klymchenko, A. S., Biophysical Journal
2008, 95, 1217.

46.
Cambrea, L. R.; Haque, F.; Schieler, J. L.; Rochet, J. C.; Hovis, J. S., Biophysical
Journal 2007, 93, 1630.

47.
Raguse, B.; Braach-Maksvytis, V.; Cornell, B. A.; King, L. G.; Osman, P. D. J.;
Pace, R. J.; Wieczorek, L., Langmuir 1998, 14, 648.

48.
Jadhav, S. R.; Sui, D. X.; Garavito, R. M.; Worden, R. M., Journal of Colloid and
Interface Science 2008, 322, 465.

49.
Miersch, S.; Espey, M. G.; Chaube, R.; Akarca, A.; Tweten, R.; Ananvoranich,
S.; Mutus, B., Journal of Biological Chemistry 2008, 283, 18513.

50.
Seu, K. J.; Cambrea, L. R.; Everly, R. M.; Hovis, J. S., Biophysical Journal 2006,
91, 3727.

51.
Sakurai, I.; Mizusawa, N.; Ohashi, S.; Kobayashi, M.; Wada, H., Plant
Physiology 2007, 144, 1336.

52.

Wada, H.; Murata, N., Photosynthesis Research 2007, 92, 205.

53.

Pandit, S. A.; Bostick, D.; Berkowitz, M. L., Biophysical Journal 2003, 84, 3743.

150

54.
PasenkiewiczGierula, M.; Takaoka, Y.; Miyagawa, H.; Kitamura, K.; Kusumi, A.,
Journal of Physical Chemistry A 1997, 101, 3677.

55.
Bockmann, R. A.; Hac, A.; Heimburg, T.; Grubmuller, H., Biophysical Journal
2003, 85, 1647.

56.
Pasenkiewicz-Gierula, M.; Takaoka, Y.; Miyagawa, H.; Kitamura, K.; Kusumi,
A., Biophysical Journal 1999, 76, 1228.

57.
558.

Greenough, K. P.; Blanchard, G. J., Journal of Physical Chemistry A 2007, 111,

58.
Uragami, M.; Miyake, Y.; Tokutake, N.; Zhang, L. H.; Regen, S. L., Langmuir
2000, 16, 8010.

59.
125.

Dicko, A.; Bourque, H.; Pezolet, M., Chemistry and Physics of Lipids 1998, 96,

60.

Zhang, Y. P.; Lewis, R.; McElhaney, R. N., Biophysical Journal 1997, 72, 779.

61.

Hull, M. C.; Cambrea, L. R.; Hovis, J. S., Analytical Chemistry 2005, 77, 6096.

62.

Richter, R. P.; Brisson, A. R., Biophysical Journal 2005, 88, 3422.

63.
DelaCruz, J. L.; Blanchard, G. J., Journal of Physical Chemistry B 2003, 107,
7102.

64.
Dewitt, L.; Blanchard, G. J.; Legoff, E.; Benz, M. E.; Liao, J. H.; Kanatzidis, M.
G., Journal of the American Chemical Society 1993, 115, 12158.

65.

Gajraj, A.; Ofoli, R. Y., Langmuir 2000, 16, 4279.

66.
Schroeder, R.; London, E.; Brown, D., Proceedings of the National Academy of
Sciences of the United States of America 1994, 91, 12130.
151

67.

Drexhage, K. H., J. Lumin. 1970, 1,2, 693.

68.

Girard, C.; Martin, O. J. F.; Dereux, A., Physical Review Letters 1995, 75, 3098.

69.

Lukosz, W.; Kunz, R. E., Opt. Commun. 1977, 20, 195.

70.

Lukosz, W.; Kunz, R. E., J. Opt. Soc. Am. 1978, 67, 1607.

71.

Chattopadhyay, A.; London, E., Biochimica Et Biophysica Acta 1988, 938, 24.

72.
Raghuraman, H.; Shrivastava, S.; Chattopadhyay, A., Biochimica Et Biophysica
Acta-Biomembranes 2007, 1768, 1258.

73.

Chuang, T. J.; Eisentha.Kb, Journal of Chemical Physics 1972, 57, 5094.

74.

Lipari, G.; Szabo, A., Biophysical Journal 1980, 30, 489.

75.

Szabo, A., Journal of Chemical Physics 1984, 81, 150.

76.
130.

Lapinski, M. M.; Blanchard, G. J., Chemistry and Physics of Lipids 2008, 153,

77.

Blanchard, G. J., Journal of Chemical Physics 1987, 87, 6802.

78.

Shoemaker, S. D.; Vanderlick, T. K., Biophysical Journal 2002, 83, 2007.

79.

Smith, L. M.; Weis, R. M.; McConnell, H. M., Biophysical Journal 1981, 36, 73.

80.

Starr, T. E.; Thompson, N. L., Biophysical Chemistry 2002, 97, 29.

81.

Dickey, A.; Faller, R., Biophysical Journal 2008, 95, 2636.

152

82.
Garret, R. H., Grisham, C.M., Biochemistry. 3rd ed.; Thompson Brooks/Cole:
Belmont, CA, 2007; p 13.

83.

Anderson, T. G.; McConnell, H. M., Biophysical Journal 2001, 81, 2774.

84.

Anderson, T. G.; McConnell, H. M., Biophysical Journal 2002, 83, 2039.

85.
Benvegnu, D. J.; McConnell, H. M., Journal of Physical Chemistry 1993, 97,
6686.

86.

Keller, S. L.; McConnell, H. M., Physical Review Letters 1999, 82, 1602.

87.
Keller, S. L.; Pitcher, W. H. I.; Huestis, W. H.; McConnell, H. M., Physical
Review Letters 1998, 81, 5019.

88.
Keller, S. L.; Radhakrishnan, A.; McConnell, H. M., Journal of Physical
Chemistry B 2000, 104, 7522.

89.

Lee, K. Y. C.; Klingler, J. F.; McConnell, H. M., Nature 1994, 263, 655.

90.
Radhakrishnan, A.; McConnell, H. M., Journal of the American Chemical Society
1999, 121, 486.

91.
Radhakrishnan, A.; McConnell, H. M., Journal of Physical Chemistry B 2002,
106, 4755.

92.
Subramaniam, S.; McConnell, H. M., Journal of Physical Chemistry 1987, 91,
1715.

93.

Seeman, P., Pharmacol. Rev. 1972, 24, 583.

94.

Slater, J. L.; Huang, C.-H., Progress in Lipid Research 1988, 27, 325.

153

95.

Vierl, U.; Lobbecke, L.; Nagel, N.; Cevc, G., Biophysical Journal 1994, 67, 1067.

96.
Slater, J. L.; Huang, C.-H., Lipid Bilayer Interdigitation. In The Structure of
Biological Membranes, Yeagle, P. L., Ed. CRC Press: New York, 2005; p 121.

97.
Van Dyke, D. A.; Pryor, B. A.; Smith, P. G.; Topp, M. R., Journal of Chemical
Education 1998, 75, 615.

98.

Vanderkooi, J. M.; Callis, J. B., Biochemistry 1974, 13, 4000.

99.

Chong, P. L. G.; Thompson, T. E., Biophys. J. 1985, 47, 613.

100.

Komatsu, H.; Rowe, E. S., Biochemistry 1991, 30, 2463.

101. Daems, D.; Vandenzegel, M.; Boens, N.; Deschryver, F. C., European Biophysics
Journal with Biophysics Letters 1985, 12, 97.

102. Lianos, P.; Duportail, G., European Biophysics Journal with Biophysics Letters
1992, 21, 29.

103. Mayer, L. D.; Hope, M. J.; Cullis, P. R., Biochimica Et Biophysica Acta 1986,
858, 161.

104. Vanderkooi, J. M.; Fischkoff, S.; Andrich, M.; Podo, F.; Owen, C. S., J. Chem.
Phys. 1975, 63, 3661.

105.

Dong, D. C.; Winnik, M. A., Canadian Journal of Chemistry 1984, 62, 2560.

106. Boon, J. M.; McClain, R. L.; Breen, J. J.; Smith, B. D., J SUPRAMOL CHEM
2001, 1, 17.

107.

Bartucci, R.; Pali, T.; Marsh, D., Biochemistry 1993, 32, 274.

154

108.

Tierney, K. J.; Block, D. E.; Longo, M. L., Biophysical Journal 2005, 89, 2481.

109.

Simon, S. A.; McIntosh, T. J., Biochimica Et Biophysica Acta 1984, 773, 169.

110.

Lobbecke, L.; Cevc, G., Biochimica Et Biophysica Acta 1995, 1237, 59.

111.

Bondar, O. P.; Rowe, E. S., Biochimica Et Biophysica Acta 1998, 1370, 207.

112.

Hata, T.; Matsuki, H.; Kaneshina, S., Biophysical Chemistry 2000, 87, 25.

113. McIntosh, T. J.; Lin, H.; Li, S. H.; Huang, C. H., Biochimica Et Biophysica Acta
2001, 1510, 219.

114.

Rowe, E. S., Biochemistry 1983, 22, 3299.

115.

Vierl, U.; Lobbecke, L.; Nagel, N.; Cevc, G., Biophys. J. 1994, 67, 1067.

116.

Zeng, J.; Chong, P. L., Biophys. J. 1995, 68, 567.

117.

Zeng, J. W.; Smith, K. E.; Chong, P. L. G., Biophys. J. 1993, 65, 1404.

118.

Lee, A., Current Biology 2001, 11, R811.

119. Ikeda, M.; Kihara, A.; Igarashi, Y., Biological & Pharmaceutical Bulletin 2006,
29, 1542.

120.

Singer, S. J.; Nicolson, G. L., Science 1972, 175, 720.

121. Harris, R. A.; Burnett, R.; McQuilkin, S.; McClard, A.; Simon, F. R., Annals of
the New York Academy of Sciences 1987, 492, 125.

155

122. Lewis, R. N. A. H.; McElhaney, R. N., The Mesomorphic Phase Behavior of
Lipid Bilayers. In The Structure of Biological Membranes, Yeagle, P. L., Ed. CRC Press:
New York, 2005; Vol. 2nd, p 53.

123.

McElhaney, R. N., Biochimica Et Biophysica Acta 1984, 779, 1.

124.

McElhaney, R. N., Crit. Rev. Microbiol. 1989, 17, 1.

125. Ryhanen, S. J.; Alakoskela, J.-M. I.; Kinnunen, P. K. J., Langmuir 2005, 21,
5707.

126.

Rowe, E. S., Biochimica Et Biophysica Acta 1985, 813, 321.

127.

Rowe, E. S., Biochemistry 1987, 26, 46.

128. Simon, S. A.; McIntosh, T. J.; Hines, M. L., The influence of anesthetic. In
Molecular and Cellular Mechanisms of Anaesthesia, Plenum Press: New York, 1986; p
297.

129.

Chen, C. H.; Hoye, K.; Roth, L. G., Arch. Biochem. Biophys. 1996, 333, 401.

130.

Tran, R.; Ho, S.; Dea, P., Biophys. Chem. 2004, 110, 39.

131. Yamazaki, M.; Miyazu, M.; Asano, T., Biochimica Et Biophysica Acta 1992,
1106, 94.

132.

Li, Q. T.; Kam, W. K., J. Biochem. Biophys. Methods 1997, 35, 11.

133.

Zbigniew, F.; Stefan, J., Molecular Physics Reports 2001, 33, 172.

134.
783.

Singh, H.; Emberley, J.; Morrow, M., European Biophysics Journal 2008, 37,

156

135.

Marsh, D., CRC Handbook of Lipid Bilayers. CRC Press: Boca Raton, FL, 1990.

136.

Goldie, S. N.; Blanchard, G. J., J. Phys. Chem. A 1999, 103, 999.

137.

Goldie, S. N.; Blanchard, G. J., J. Phys. Chem. A 2001, 105, 6785.

138.

Jiang, Y.; Blanchard, G. J., J. Phys. Chem. 1994, 98, 6436.

139.

Chuang, T. J.; Eisenthal, K. B., Journal of Chemical Physics 1972, 57, 5094.

140.

Favro, D. L., Physical Review 1960, 119, 53.

141.

Hu, C.-M.; Zwanzig, R., J. Chem. Phys. 1974, 60, 4354.

142.

Hildenbrand, K.; Nicolau, C., Biochimica Et Biophysica Acta 1979, 553, 365.

143.

Debye, P., Polar Molecules. In Chemical Catalog Co.: New York, 1929.

144.

Zwanzig, R.; Harrison, A. K., Journal of Chemical Physics 1985, 83, 5861.

145.

Jiang, Y.; Blanchard, G. J., Journal of Physical Chemistry 1994, 98, 6436.

146.

Edward, J. T., Journal of Chemical Education 1970, 47, 261.

147.

Frye, L. D.; Edidin, M., Journal of Cell Science 1970, 7, 319.

148.

Cone, R. A., Nature-New Biology 1972, 236, 39.

149.

Poo, M.-m.; Cone, R. A., Nature 1974, 247, 438.

150.

Edidin, M., Nat Rev Mol Cell Biol 2003, 4, 414.
157

151. Tocanne, J.-F.; Cézanne, L.; Lopez, A.; Piknova, B.; Schram, V.; Tournier, J.-F.;
Welby, M., Chemistry and Physics of Lipids 1994, 73, 139.

152. Greenough, K. P.; Blanchard, G. J., Journal of Physical Chemistry B 2006, 110,
6351.

153. Greenough, K. P.; Blanchard, G. J., Spectrochimica Acta Part A: Molecular and
Biomolecular Spectroscopy 2009, 71, 2050.

154. Greiner, A. J.; Pillman, H. A.; Worden, R. M.; Blanchard, G. J.; Ofoli, R. Y.,
Journal of Physical Chemistry B 2009, 113, 13263.

155. Dela Cruz, J. L.; Blanchard, G. J., Journal of Physical Chemistry A 2001, 105,
9328.

156.

Schroder, M., Cellular and Molecular Life Sciences 2008, 65, 862.

157. Horvath, I.; Multhoff, G.; Sonnleitner, A.; Vigh, L., Biochim. Biophys. ActaBiomembr. 2008, 1778, 1653.

158.

Edward, J. T., Journal of Chemical Education 1970, 47, 261.

159. Riddick, J. A., Techniques of Chemistry. In Organic Solvents, Wiley: New York,
1986; Vol. II.

160. Crank, J.; Nicolson, P.; Hartree, D. R., Proceedings of the Cambridge
Philosophical Society 1947, 43, 50.

161.

Kasha, M., Faraday Discussions of the Chemical Society 1950, 9, 14.

162.

Hambir, S. A.; Jiang, Y.; Blanchard, G. J., J. Chem. Phys. 1993, 98, 6075.

163.

Jiang, Y.; Blanchard, G. J., J. Phys. Chem. 1994, 98, 9411.

158

164.

Jiang, Y.; Blanchard, G. J., J. Phys. Chem. 1994, 98, 9417.

165.

Jiang, Y.; Blanchard, G. J., J. Phys. Chem. 1995, 99, 7904.

166. Yeagle, P. L., The Roles of Cholesterol in the Biology of Cells. In The Structure
of Biological Membranes, 2nd ed.; Yeagle, P. L., Ed. CRC Press: New York, 2005.

167. Burger, K.; Gimpl, G.; Fahrenholz, F., Cellular and Molecular Life Sciences
2000, 57, 1577.

168.

McIntosh, T. J., Biochimica et Biophysica Acta 1978, 513, 43.

169. Guo, W.; Volker, K.; Huber, T.; Afdhal, N. H.; Beyer, K.; Hamilton, J. A.,
Biophysical Journal 2002, 83, 1465.

170. Lindblom, G.; Orädd, G., Biochimica et Biophysica Acta (BBA) - Biomembranes
2009, 1788, 234.

171. Nicolini, C.; Kraineva, J.; Khurana, M.; Periasamy, N.; Funari, S. S.; Winter, R.,
Biochimica et Biophysica Acta (BBA) - Biomembranes 2006, 1758, 248.

172.

Oldfield, E.; Chapman, D., Febs Letters 1972, 21, 303.

173.

Nyström, J. H.; Lönnfors, M.; Nyholm, T. K. M., 2010, 99, 526.

174. Lewis, R. N. A. H.; McElhaney, R. N., The Mesomorphic Phase Behavior of
Lipid Bilayers. In The Structure of Biological Membranes, 2nd ed.; Yeagle, P. L., Ed.
CRC Press: New York, 2005.

175.

Simons, K.; Ikonen, E., Nature 1997, 387, 569.

176.

Simons, K.; Vanmeer, G., Biochemistry 1988, 27, 6197.

159

177. Brown, D. A.; London, E., Biochemical and Biophysical Research
Communications 1997, 240, 1.

178.

Brown, D. A.; London, E., Journal of Membrane Biology 1998, 164, 103.

179.

Brown, D. A.; Rose, J. K., Cell 1992, 68, 533.

180. Pillman, H. A.; Blanchard, G. J., Journal of Physical Chemistry B 2010, 114, in
press.

181.

Maxfield, F. R.; Tabas, I., Nature 2005, 438, 612.

182.

Simons, K.; Ikonen, E., Science 2000, 290, 1721.

183.

Veatch, S. L.; Keller, S. L., Physical Review Letters 2002, 89, 268101.

184.

Nagle, J. F., Annual Review of Physical Chemistry 1980, 31, 157.

185. Chong, P. L.-G.; van der Meer, B. W.; Thompson, T. E., Biochimica et
Biophysica Acta (BBA) - Biomembranes 1985, 813, 253.

160