, ....:\1

J. w..v:1huu\nx.£ L. 42.. E dam—éwvhrg
. > 3 . t: 3.)?)

...
4 05....
r... : . .12" 1
. ﬂan}
(uhézéﬂui

x
v.
.

x

.

.3

3.2.:
x

3.?

:21 .2 «
It?» .

1.5

smut:
1.3.... up .

a...

 

34%
'w;$.4 ,

 

,l.
rn27194

 

.
zen-(.11.: 3..
1172!!) x
A .1. 1., v

Iv

 

 

 

 

 

gulp!“ » .1... . , ., , a .
I m2 wihvwhuari hut." ... Ma ..1 ‘ v“.M «Rw.%%ﬂ.?
r, t . ‘. a7....w~w

. t. , . :

 

r

 

.1006 ? LIBRARY
Michigan State

* __ University

 

 

 

This is to certify that the
dissertation entitled

ABSORPTION AND INTERACTIONS OF BIOMACROMOLECULES
AT FLUID-LIKE INTERFACES

presented by

Sachin Shashikant Vaidya

has been accepted towards fulﬁllment
of the requirements for the

Ph.D. degree in Chemical Enﬂleering

 

 

MW W

Major Profe r’ s Sign? ure

ﬂit/05’
//

Date

MSU is an Afﬁnnative Action/Equal Opportunity Institution

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

A UIPJABEWEgg DATE DUE DATE DUE

 

350410

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 a/ciﬁcrpmm.mms

 

ADSORPTION AND INTERACTIONS OF BIOMACROMOLECULES AT FLUID-
LIKE INTERFACES

By

Sachin Shashikant Vaidya

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemical Engineering and Materials Science

2005

ABSTRACT

ADSORPTION AND INTERACTIONS OF BIOMACROMOLECULES AT FLUID-
LIKE INTERFACES

By

Sachin Shashikant Vaidya
The adsorption and interactions of biomacromolecules at biologically relevant
interfaces have been characterized in this study. While each chapter addresses a unique
issue, they all share an underlying common theme of increasing our understanding of the
behavior of biomolecules at ﬂuid-like interfaces. The primary tool for all three studies

was total internal reﬂection ﬂuorescence microscopy (T IRFM).

The ﬁrst study addressed sequential and competitive adsorption and interactions
of human plasma ﬁbronectin (HFN) and human serum albumin (HSA) at the oil-water
interface. Among key results, ﬁbronectin adsorption is rapid and essentially irreversible,
even over short time scales. This is probably due to the highly ﬂexible nature of the
protein, which allows its domains to quickly attain energetically favorable conformations.
In contrast, HSA adsorption is relatively reversible at short times, and the protein is
readily displaced by ﬁbronectin even after incubation at the interface for as long as two
hours. HSA more effectively resists displacement by ﬁbronectin at longer time scales,
although signiﬁcant ﬁbronectin adsorption occurred even under those conditions.
Displacement of ﬁbronectin by HSA was essentially negligible under all conditions. This
study is relevant to emerging research thrusts such as the development of biomimetic
interfaces, where effects of interfacial competition, adsorption time scales, and extent of

adsorption irreversibility on interfacial dynamics are important.

The second study addressed the adsorption of bovine serum albumin at a model
oil-water interface. Estimates of protein interfacial coverages were obtained using a
protocol based on a ﬂuorescence recovery after photobleaching (FRAP) technique
previously proposed by Zimmerman et a1. (1990). Protein coverages ranged from 0.02-
0.3 mg/m2 for bulk concentrations of 0.2 mg/ml to 3.5 mg/ml. These values are an order
of magnitude lower than typical estimates reported in the literature for the solid-liquid
interface. It is likely that coverages at the oil-water interface are lower because more
rapid protein relaxations at this interface result in greater interfacial area per molecule,
thus inhibiting adsorption of later-arriving proteins. Fluorescence lifetime measurements
of BSA-FITC in solution and at the interface (obtained by two-photon time correlated
single photon counting spectroscopy) yielded similar results, suggesting that the quantum

yield of BSA-FIT C at the interface and in the bulk solution are identical.

The ﬁnal study was on the adsorption of liposomes composed of 1,2-Dioleoyl-sn-
Glycero—3 -Phosphocholine (DOPC) and l ,2-Dioleoyl-sn-Glycero—3-Phosphate
Monosodium Salt) (DOPA) to poly(dimethyldiallylammonium chloride) (PDAC) and
poly(allylamine hydrochloride) (PAH). The liposomes adsorbed preferentially on PDAC,
but in much smaller amounts on sulfonated poly(styrene) (SPS) due to electrostatic
repulsion between the negatively charged liposomes and the SPS coated surface.
Poly(ethylene glycol) (m-dPEG acid) coated surfaces also resisted liposome adsorption.
These results were exploited to create arrays of lipid bilayers by exposing PDAC, PAH
and m-dPEG patterned substrates to DOPA/DOPC vesicles of various compositions.
Such arrays may be useful for high-throughput screening of compounds that interact with

cell membranes.

Dedicated to my mother, father and sister,
and to Salil for all that he missed

iv

i”

ACKNOWLEDGEMENTS

I would like to thank my advisor Dr. Robert Y. Ofoli for his invaluable guidance
and for his friendship during these long years. I am grateful to him for recruiting me and
giving me the opportunity to advance my studies at Michigan State. His open-door policy
has always encouraged me to speak freely with him and he has taught me to be an
independent thinker. I would like to thank my committee members for agreeing to mentor
me and for their advice and support during the course of my dissertation. I would like to
thank Dr. Melissa Baumann for her inputs and for her careful critique of this dissertation,
Dr. Gary Blanchard for his assistance with the two-photon spectroscopy measurements
and for his patience when I was incessantly bugging him about experiments and Dr.
Ilsoon Lee and Dr. Mark Worden for their mentorship, fruitful collaborations during

these past years, and for providing me with generous use of their laboratory facilities.

During my years at MSU, I developed friendships that I will never forget and it
remains the single most valuable possession that I carry forward. Sunder and Bhaskar are
my earliest friends here, right from the Shaw Hall days and although everyone has gone
their separate ways, we share the best of memories. I would like to thank the A-24 gang,
Arun, Vinay, Pradeep, Mandar, Srini and Akshay for being the best roommates. Arun has
been my closest friend over the last 12 years and his friendship has meant a lot to me. I
would also like to thank Tejas, Soumya, Abhi, Kiran, Amit, Prajakt, Giri, and Prema for
being great friends.

I would like to thank Lavanya Parthasarathy for been a great friend and colleague

Without her, it would have been impossible to navigate the treacherous TIRFM waters.

We started our PhDs together and have learned far more than this dissertation will ever
tell. Ari Gajraj was a great mentor in the early years and I would like to thank him for his
guidance, and for the music. I would also like to thank Neeraj Kohli for endless scientiﬁc
discussions and for all our collaborations and wonderful research ideas and goals that
Nobel prize winners always seemed to beat us to (by at least a couple of years!). I
would also like to thank Nancy Albright and JoAnn Peterson for all their help during

these past six years.

I have been most inﬂuenced in my life, by my parents and sister who have
sacriﬁced a lot for me. They have always encouraged me to seek higher goals and have
never doubted me. This dissertation would not have been possible if not for their love and
support. This dissertation is dedicated to the memory of my brother, who I wish, could be
here. Finally I would like to thank my wife Mukta for her support during the ﬁnal years
of this dissertation. She has always been there for me, especially at the worst times when
the end seemed just as far away as the beginning. Her constant love and encouragement

has been the beacon that has guided me to the end.

vi

TABLE OF CONTENTS
LIST OF FIGURES ......................................................................................................... x
NOMENCLATURE ..................................................................................................... xvi
1 INTRODUCTION ................................................................................................... 1
1. I Background ...................................................................................................... l
1.2 Sequential and competitive adsorption between human serum albumin and
human plasma ﬁbronectin at the oil-water interface ..................................................... 3
1.3 Estimating protein interfacial coverage at the liquid-liquid interface ................ 4
1.4 Fabrication and characterization of 3D arrays of lipid-bilayers on
polyelectrolyte multilayers .......................................................................................... 6
2 ADSORPTION AND INTERACTION OF FIBRONECTIN AND ALBUMIN AT
THE LIQUID-LIQUID INTERFACE ............................................................................. 9
2.1 ABSTRACT .................................................................................................... 9
2.2 INTRODUCTION ......................................................................................... 10
2.3 Theory ........................................................................................................... 13
2.4 Experimental Protocol .................................................................................... 14
2.4.1 Materials: ................................................................................................... 14
2.4.2 TIRFM Setup ............................................................................................. 16
2.5 Results and Discussion ................................................................................... 19
2.5.1 Sequential adsorption of HSA and HFN ..................................................... 20
2.5.2 Effect of adsorption timescales ................................................................... 26
2.5.3 Competitive adsorption of HFN and HSA .................................................. 29
2.6 Conclusions ................................................................................................... 31
2.7 Recommendations for future work ................................................................. 32

3 ESTIMATING PROTEIN INTERFACIAL COVERAGE AT THE LIQUID-

LIQUID INTERFACE: A TIRF - FRAP STUDY .......................................................... 44
3.1 ABSTRACT .................................................................................................. 44
3.2 Introduction ................................................................................................... 45
3.3 Theory ........................................................................................................... 48

3.3.1 Total internal reﬂection .............................................................................. 48
3.3.2 The calibration protocol ............................................................................. 49
3.3.3 Two-photon ﬂuorescence lifetime spectroscopy and the quantum yield
problem ................................................................................................................. 54
3.3.3.1 Rationale 54
3.3.3.2 Theory 57
3.3.3.3 Measurement of Fluorescent Lifetimes 59
3.4 Experimental Methods ................................................................................... 61
3.4.1 Materials and Methods ............................................................................... 61
3.4.2 Protein labeling .......................................................................................... 61

vii

3.4.3 Preparation of surfaces ............................................................................... 62
3.4.4 TIRFM instrumentation and experimental set-up: ....................................... 62
3.4.5 Two photon lifetime spectroscopy: Experimental set-up. ............................ 64
3.4.5.1 Data analyses for TIRFM and calibration protocol 65
3.4.5.2 2-pTCSPC lifetime measurements 65
3.5 Results and Discussion ................................................................................... 66
3.5.1 Adsorption isotherm for BSA-FIT C at the oil-water interface ..................... 66
3.5.2 Molecular relaxations and its inﬂuence on ultimate coverages at the oil-water
interface ................................................................................................................. 68 .
3.5.3 Relaxations at the OTS-water interface ...................................................... 73
3.5.4 Sequential adsorption experiments ...................................................... 75
3.5.5 Fluorescence Lifetime measurements ......................................................... 76
3.5.6 Other sources of error ................................................................................. 77
3.5.6.1 Possibility of incomplete bleaching of interfacial ﬂuorophores 77
3.5.6.2 Elevation of the bulk-signal contribution by light scattering: 78
3.6 Conclusions ................................................................................................... 81
3.7 Recommendations for future work ................................................................. 81
4 3-D ARRAYS OF LIPID BILAYERS ON POLYELECTROLYTES
MULTILAYERS ........................................................................................................... 99
4.1 ABSTRACT ......................................................................... 99
4.2 Introduction ................................................................................................. 100
4.3 Experimental section .................................................................................... 103
4.3.1 Materials .................................................................................................. 103
4.3.2 Preparation of stamps: .............................................................................. 104
4.3.3 Preparation of liposomes: ......................................................................... 104
4.3.4 Preparation of arrays: ............................................................................... 105
4.3.5 Total internal reﬂection microscopy: ........................................................ 106
4.3.6 Fluorescence recovery after photobleaching (TIRF-FRAP): ..................... 106
4.3.7 Fluorescence recovery after pattern photobleaching (FRAPP): ................. 107
4.4 Theory and Data Analyses for Fluorescence recovery after pattern
photobleaching ........................................................................................................ 109
4.4.1 Theory ..................................................................................................... 1 10
4.4.1.1 Samples containing a single diffusive population: 112
4.4.1.2 Samples containing multiple diffusive populations (n 2 1) 112
4.4.2 Data Analyses and curve ﬁtting ................................................................ 113
4.5 Results and Discussions ............................................................................... 114
4.5.1 TIRFM Adsorption experiments: .............................................................. 114
4.5.2 Arrays of lipid bilayers: ............................................................................ 118
4.5.3 TIRF spot photobleaching: ....................................................................... 119
4.5.4 Fluorescence recovery after pattern photobleaching: ................................ 120
4.6 Conclusions ................................................................................................. 123
4.7 Recommendations for future work ............................................................... 124
5 APPENDICES ..................................................................................................... 143

viii

5.1 Appendix A: Labview ﬂowsheets ................................................................. 143
5.1.1 Labview ﬂowsheets for TIRFM experiments conducted in analog mode ..143

5.1.2 Labview ﬂow sheets for photon counting ................................................. 145
5.2 Appendix B: Matlab programs for ﬁltering and averaging analog measurements
148

5.2.1 Filtering program ..................................................................................... 148
5.2.2 Averaging Script ...................................................................................... 149

5.3 Appendix C: Protocol for estimation of spot diameters for TIRF-
photobleaching ........................................................................................................ 150
5.4 APPENDIX D: Quartz crystal microbalance studies ..................................... 154
LIST OF REFERENCES ............................................................................................. 155

ix

r: .

LIST OF FIGURES

Figure 2.1: a) Experimental layout for TIRFM set-up: NDF: Neutral density ﬁlter, OC:
optical chopper, F1,F2: optical ﬂats, Sh: shutter, M1,M2: Mirrors, C: coatings b)
Layout of ﬂow cell for TIRF experiments. ............................................................. 34

Figure 2. 2: Sequential adsorption of unlabeled HFN (0.1 mg/ml) at the oil- water
interface, following adsorption of labeled HSA (0.130 mg/ml) to near interfacial
equilibrium. In this experiment, a mixture of labeled HSA and unlabeled HFN was

introduced into the ﬂow cell approximately 75 minutes after the adsorption of -1“
labeled HSA had been initiated. The plot gives strong evidence of displacement of
adsorbed HSA at the interface. ............................................................................... 35

Figure 2.3: Sequential adsorption of HFN/HSA at the oil~water interface. In this
experiment, a mixture of labeled HFN and unlabeled HSA was introduced into the
ﬂow cell approximately 60 minutes after adsorption of labeled HFN had been
initiated. The proﬁles show that HSA had no apparent effect on HFN adsorption. In
the top graph (a), HSA and HFN are at the same bulk concentration (0.1 mg/ml); in
the bottom graph (b), the concentration of HSA is an order of magnitude higher (1
mg/ml). (c) Desorption of HFN from the interface by interfacial rinsing. Based on
this data, adsorption of HFN at the oil-water interface is essentially irreversible.
Little reduction in HFN ﬂuorescence emission intensity is observed upon
introduction of a protein-free rinse (solid circles) or circulation of unlabeled HSA
(solid triangles) after 60 minutes of HFN at the interface. ...................................... 36

Figure 2.4: Sequential adsorption of HSA and I-IFN at the oil-water interface. Figure 2.4a
shows that pre-adsorption of HFN completely frustrates subsequent HSA adsorption.
In this experiment, unlabeled HFN was adsorbed for 60 minutes. Subsequently,
labeled HSA was introduced into the ﬂow cell. In the top graph (a) HFN is present
in the bulk solution at a concentration of 0.1 mg/ml; in the bottom graph (b), the
concentration of HFN is at a much lower bulk concentration of 0.02 mg/ml. Curve 2
in this plot (solid circles) is for HSA adsorbing by itself at the same bulk
concentration, and is given for comparison. ........................................................... 37

Figure 2.5: Effect of pre-adsorption of HSA on HFN adsorption at the oil-water interface:
Curve (1) (solid circles) shows adsorption of labeled HFN to a bare oil-water
interface. Curve (2) (open squares) shows adsorption of labeled HFN to an oil-water
interface at which unlabeled HSA had been pre—adsorbed for 2 hrs. Curve (3) (open
triangles) shows labeled HFN adsorption at an oil-water interface at which unlabeled
HSA had been pre-adsorbed for a period of 6 hrs. Bulk concentrations of HSA and
HFN are 0.1 mg/ml and 0.05 mg/ml, respectively. ................................................. 38

Figure 2.6: Cartoon of a model proposed for adsorption of HFN at an oil-water interface
at which HSA has been pre-adsorbed. Curves A, B and C represent models for
proposed interfacial organization, under conditions corresponding to curves 1, 2 and
3 in Figure 5. ......................................................................................................... 39

 

Figure 2.7: Displacement of labeled HSA by unlabeled HFN at the oil-water interface. In
each experiment, unlabeled HFN (0.1 mg/ml) is introduced into the ﬂow cell after
HSA-FITC (0.130 mg/ml) adsorption for 1 hour (solid circles) and 3 hours (open
triangles), respectively (t=0 minutes). In each set, the data has been normalized by
the pseudo—saturation intensity obtained prior to the buffer wash. Clearly, HSA is
more effective at mitigating displacement by HFN after three hours at the interface
than after one hour. ................................................................................................ 40

Figure 2.8: Effect of pre-adsorbed HSA concentration on subsequent HFN adsorption.
Clearly, pre-adsorption of a higher bulk concentration of albumin (0.5 mg/ml) does
not signiﬁcantly inhibit subsequent HFN (0.05 mg/ml) adsorption at the interface. 41

Figure 2.9: Competitive adsorption of HSA and HFN at the oil-water interface: This plot
shows preferential adsorption of I-IFN from a mixture of HFN (0.05 mg/ml) and
HSA (0.125 mg/ml) to a bare oil-water interface. The bottom curve (open triangles)
is adsorption of a mixture of unlabeled HFN and labeled HSA to a bare oil-water
water interface, and is shown for comparison. ........................................................ 42

Figure 2.10: Interfacial tension measurements of a mixture of HFN (0.1 mg/ml) and HSA
(0.125 mg/ml) adsorbing to an oil-water interface (middle curve) are shown in
Figure 10a. For comparison, interfacial tension measurements of pure HSA (top
curve) and pure HFN (bottom curve) are also presented. In Figure 10b, the
concentration of HSA is increased to 1mg/ml. Over long times, the interfacial
tension of the mixture approaches that of pure ﬁbronectin. .................................... 43

Figure 3.1: A typical photobleaching experiment for calibrating interfacial coverages.
F(t) represents the total ﬂuorescence prior to bleaching. Fbulk (t) is the total
ﬂuorescence measured after bleaching is terminated. ........ ' ..................................... 83

Figure 3.2: Schematic showing the photobleaching steps involved in the calibration
protocol. The black circles represent bleached ﬂuorophores while the white circles
represent unbleached ﬂuorophores. Due to the high diffusion coefﬁcient of the bulk
associated ﬂuorophores, the bleached ﬂuorophores in the bulk are rapidly replaced
by unbleached ﬂuorophores thereby resulting in an “unbleachable bulk fraction”,
while the surface bound ﬂuorophores remain bleached. ......................................... 84

Figure 3.3: Experimental layout for 2-photon time correlated single photon counting. G-
T: Glan-Taylor prism; MCP-PMT: microchannel-plate photomultiplier tube; CFD:
constant fraction discriminator; TAC: Time to amplitude converter ....................... 85

Figure 3.4: Adsorption isotherm of BSA-FITC at the oil-water interface using TIRFM-
FRAP protocol ....................................................................................................... 86

Figure 3.5: Proposed model of interfacial relaxations for albumin at an oil-water
interface. Molecules that arrive early, occupy a small area per molecule. However
with time, they spread and occupy larger areas, thereby restricting available area for
molecules that arrive much later, due to slow transport rates. ................................. 87

xi

 

 

 

Figure 3.6: Adsorption of BSA-FITC (0.2 mg/ml) at the oil-water interface as a function
of ﬂow rate. Three ﬂow rates were used in this experiment: (1) 0.024 ml/min
0:013 s“); 0.048 ml/min 0:027 s") and (3) 0.096 ml/min 0:054 s") At a ﬂow
rate of 0.024 ml/min, considerably less protein adsorbs to the oil-water interface
presumably due to molecular relaxations of adsorbed protein that reduce the
available interfacial area. ............. . .......................................................................... 88

Figure 3.7: Inﬂuence of molecular relaxations on the adsorption of BSA-FITC (0.2
mg/ml) to the oil-water interface:. The data shown above has been re-plotted from
Figure 3.6 for y=0.27 s'l and y=0.13 s". The departure of the adsorption proﬁle from
the straight line, at different interfacial levels for different transport conditions
suggests the presence of relaxations. ...................................................................... 89

Figure 3.8: Adsorption of BSA-FITC (0.2mg/ml) to octadecyltrichlorosilane self
assembled monolayers at ﬂow rates of 0.47 mein (7:2.7 3", top curve) and 0.047
ml/min (7:0.27 s'l, bottom curve). ......................................................................... 90

Figure 3.9: Sequential adsorption experiments at the oil-water interface: BSA-FITC at a
concentration of 0.2, 0.5 and 1 mg/ml was sequentially introduced into the ﬂow-cell
and in each case, interfacial coverages were computed using the calibration protocol.
.............................................................................................................................. 91

Figure 3.10: Typical ﬂuorescence decay proﬁles obtained for BSA-FITC adsorbing to the
oil-water interface at a concentration of 1 mg/ml using 2-photon TCSPC ............... 92

Figure 3.11: Single exponential ﬁt to ﬂuorescence decay proﬁle for BSA-FITC (1 mg/ml)
adsorbed at the oil-water interface .......................................................................... 93

Figure 3.12: Single exponential ﬁt to ﬂuorescence decay proﬁles for BSA-FITC

(lmg/ml) in phosphate buffer ................................................................................. 94
Figure 3.13: Single exponential ﬁt to decay proﬁles of unbound FITC (0.15mg/ml) in
phosphate buffer. ................................................................................................... 95
Figure 4.1: Structures for a) common polyelectrolytes and b) m d-PEG. ...................... 127

Figure 4.2: Illustration showing creation of arrays of lipid bilayers on PEMs with PDAC
or PAH as the topmost layer. ............................................................................... 128

Figure 4.3: Illustration showing formation of lipid bilayers on PEMs with m.dPEG acid
as the uppermost layer. ........................................................................................ 129

Figure 4.4: Experimental set-up for Fluorescence recovery after pattern photobleaching
using EPI-illumination. ........................................................................................ 130

Figure 4.5: a) Fringe pattern in illuminated region obtained using 100 lines per inch
ruling. b) Observation area restricted by placing an aperture in the camera/PMT
image plane. ........................................................................................................ 131

xii

 

 

Figure 4.6: Adsorption curves of (A) liposomes (10%DOPA, 90% DOPC) on PDAC. (B)
Liposomes (10%DOPA, 90% DOPC) on SPS. (C) Liposomes (20%DOPA, 80%
DOPC) on PDAC. (D) Liposomes (20%DOPA, 80% DOPC) on SPS. ................. 132

Figure 4.7: Shear induced desorption of liposomes from PDAC/SPS PEMs with SP8 as
the topmost layer. Binding of liposomes to these surfaces is relatively loose and a
large fraction is removed upon introduction of buffer solution at a ﬂow rate of 0.34
ml/min. ................................................................................................................ 133

Figure 4.8: Adsorption curves of liposomes (10%DOPA, 90% DOPC) on glass slide
coated with PEMs with PDAC (upper curve) and m-dPEG acid (lower curve) being
the topmost layer. As can be seen, the lower ﬂuorescence intensitites obtained are
indicative of the ability of PEG to resist liposome adsorption ............................... 134

Figure 4.9: a) Adsorption of liposomes (10%DOPA, 90% DOPC) on a glass slide coated
with PEMs with m-dPEG acid being the topmost layer. b) Buffer-wash experiments
to study liposome desorption from PEMs. The top and bottom curves depict
desorption of liposomes from PEMs with PDAC and m—dPEG as the top layer,
respectively. At t=0, adsorption of liposomes (which have adsorbed for at least 45
minutes) is halted by introducing liposome-free buffer. In each curve, the
ﬂuorescence intensity has been normalized by the corresponding ﬂuorescence value
obtained prior to initiation of the buffer wash in each case. .................................. 135

Figure 4.10: Fluoresence images showing (a) line patterns on a PDAC patterned substrate
(b) circular patterns on a PDAC patterned substrate (c) line patterns on a PAH
patterned substrate. .............................................................................................. 136

Figure 4.11: Fluorescence images showing (a) line patterns on a m-dPEG acid patterned
substrate (b) circular patterns on a m-dPEG patterned substrate. .......................... 137

Figure 4.12: a) TIRF-FRAP data for DOPA/DOPC liposomes adsorbed on PEMs with
PDAC as the top layer. b) TIRF-FRAP data under conditions of intermittent
monitoring to prevent monitoring beam induced photobleaching over longer time
scales. .................................................................................................................. 138

Figure 4.13: Fluoresence recovery after pattern photobleaching (EPI-FRAPP) proﬁles on
PDAC. The solid line in the plate A represent ﬁts to the recovery data set with a
model159 that describes the sample as containing containing a single mobile and
immobile fraction, while in plate B the solid line is the ﬁt to a model describing a
sample with two mobile populations (with different mobilities) and an immobile
fraction. Average values obtained with these models are summarized in Table 4.1
and Table 4.2. Also shown below each ﬁt is a plot of the residuals vs. time as an
indication of the goodness of ﬁt. .......................................................................... 139

Figure 4.14: Fluoresence recovery after pattern photobleaching (EPI-FRAPP) proﬁles on
PAH. The solid line in the plate A represent ﬁts to the recovery data set with a
model159 that describes the sample as containing containing a single mobile and
immobile fraction, while in plate B the solid line is the ﬁt to a model describing a

xiii

 

 

 

sample with two mobile populations (with different mobilities) and an immobile
fraction. Average values obtained with these models are summarized in Table 4.3
and Table 4.4. Also shown below each ﬁt is a plot of the residuals vs. time as an
indication of the goodness of ﬁt. .......................................................................... 140

Figure 5.1: Labview ﬂow sheets for TIRF data acquisition in analog mode. The ﬂow
sheet is continued on the next page. ..................................................................... 143

Figure 5.2: Flow sheets for computer interfacing with SR400 photon counter. These ﬂow
sheets have been continued on the succeeding pages. These ﬂowsheets have been
derived from Labview templates obtained from National Instruments, Austin, TX
............................................................................................................................ 145

Figure 5.3: Stage speed calibrations displaying speed (in arbitrary units) against actual
speed for both X- and Y motion controllers .......................................................... 152

Figure 5.4: Estimating the dimensions of the elliptical spot. This picture shows a typical
scan in the x-coordinate. A spot is bleached and the specimen is translated in the x-
direction across the bleached spot. The data between the dotted lines represents the
duration during the scan when the objective collects light from the bleached area of
the sample ............................................................................................................ 153

Figure 5.5: Adsorption of liposomes as studied by QCM on (a) PDAC (b) PAH. Note the
presence of two distinct phases for adsorption on PAH (suggesting vesicle
adsorption and rupture), in contrast with the presence of one phase for PDAC
(suggesting vesicle adsorption without rupture). .................................................. 154

xiv

LIST OF TABLES

Table 3.1: Comparison of initial adsorption rates obtained using the photobleaching
model with that predicted by the Leveque equation. ............................................... 96

Table 3.2: Results of ﬁtting of the ﬂuorescence decay to the single exponential model
given by Equation (3.16) ........................................................................................ 97

Table 3.3: Results of ﬁtting of ﬂuorescence decay to multi-exponential model given by
Equation (3.17). ..................................................................................................... 98

Table 4.1: Summary of model parameters obtained from ﬁt of BLM recovery on PDAC
to Equation (4.6) (single mobile-species model) ................................................... 141

Table 4.2: Fit parameters for lipid bilayer formed on substrates topped with PDAC, using
the two mobile-species model given by (4.7). f0 represents the unbleached fraction
(at t=0), and m1, m2 and D1, D2 are the mobile fractions and their corresponding
diffusion coefﬁcients respectively. ....................................................................... 141

Table 4.3: Summary of model parameters obtained from ﬁt of BLM recovery on PAH to
Equation (4.6) (single mobile-species model) ....................................................... 142

Table 4.4: Summary of model parameters obtained from ﬁt of BLM recovery on PAH142

XV

NOMENCLATURE
pre—exponential factor (or amplitude) of exponential decay
index for estimating goodness of ﬁt
wall-shear rate
wavelength of incident light in free space
incident angle of light
critical angle for total internal reﬂection
ﬂuorescence lifetime of excited molecules
absorbance
bulk concentration of proteins
two-dimensional interfacial concentration of proteins
penetration depth of evanescent wave
diffusion coefﬁcient of species 1'
unbleached ﬂuorescence emission after pattern photobleaching step
total ﬂuorescence as a function of time, prior to photobleachin g
1(t) bulk contribution to the total ﬂuorescence

depth of channel
incident light intensity at depth z=0
evanescent wave intensity at depth 2

radiative decay rate

xvi

 

non-radiative decay rate

1 aqueous depth of ﬂow cell

L molar labeling ratio of proteins (dye/protein)
m.- fraction of total population that is mobile with diffusion coefﬁcient D,-
n; refractive index of denser medium

n2 refractive index of rarer medium

q quantum yield of ﬂuorophore

q, instrument constant

qb quantum yield of ﬂuorophores in solution

q, quantum yield of adsorbed ﬂuorophores

Q volumetric ﬂow rate in ﬂow cell

5 thickness of adsorbed protein layer

1 time

w width of ﬂow channel

2 depth coordinate into the ﬂow cell
Abbreviations

AFM atomic force microscopy

BLM bilayer lipid membrane

BSA bovine serum albumin

xvii

DOPA l,2—Dioleoyl-sn-Glycero-3-Phosphate (Monosodium Salt)
DOPC 1,2-Dioleoyl-sn-Glycero-3-Phosphocholine

FITC ﬂuorescein isothiocyanate

FRAP Fluorescence recovery after photobleaching

FRAPP ﬂuorescence recovery after pattern photobleachin g
HFN human plasma ﬁbronectin

HSA human serum albumin

HTS hexadecyltrichlorosilane

OTS octadecyltrichlorosilane

PEG polyethylene glycol

PEO polyethylene oxide

PEM polyelectrolyte multilayers

PAH poly(allylarrrine hydrochloride)

PDAC poly(diallyldimethyl ammonium chloride)

PMT photomultiplier tube

QCM quartz crystal microbalance

SAM self-assembled monolayer

SPS sulfonated polystyrene

2p-TCSPC 2-photon time-correlated single photon counting

TIRFM total internal reﬂection ﬂuorescence microscopy

xviii

VA-TIRFM variable angle of incidence total internal reﬂection ﬂuorescence

microscopy

xix

1 INTRODUCTION

1.1 Background

The term adsorption, coined in the 19th century, is deﬁned as the preferential
partitioning of a molecule along the boundary of two phases. This partitioning occurs due
to the afﬁnity of regions in the molecule for one phase or the other. Though classical
adsorption has its roots in catalysis, the study of adsorption of molecules at biologically
relevant interfaces (bio-adsorption) is now important to a large number of applications.
For example, protein adsorption at cellular interfaces is among the ﬁrst and most
important events that take place before the initiation of processes such as cell
differentiation, proliferation and growth, and considerable effort has been expended in

the literature to characterize the adhesion of proteins to surfaces.

In addition to examining the process of biomolecular adsorption, it is also
important to investigate the behavior of these molecules following adsorption at the
interface. For example, proteins are ﬂexible molecules that are known to undergo
relaxation after adsorption. The relaxation occurs as the protein tries to minimize its free
energy, and leads to protein denaturation that may make the adsorption irreversible.
There is considerable focus on studying the conformational changes that such molecules
undergo, during and after the adsorption process. Another area where protein adsorption
is important is in the ﬁeld of tissue engineering, particularly in the design of implants and
scaffolds. The liquid-liquid interface is an important emerging interface in the ﬁeld of

tissue engineering as it presents a “soft interface” which can be used for culturing cells.

In the area of biosensor development, current research thrusts lie in developing
sustainable biomimetic interfaces. The interfaces feature lipid bilayers deposited on them,
using vesicular adsorption or Langmuir Blodgett methods, which allows them to mimic
the cellular membrane. When liposomes adsorb to the interface, they typically tend to
fuse and rupture, thus forming lipid bilayers. The lipid bilayers have varying properties
depending on characteristics of the interface such as heterogeneity, surface charge and
roughness. Understanding the mechanisms and controlling the parameters that govern
liposome adsorption and rupture to form lipid bilayers is critical to the design of viable

biosensor systems.

This dissertation examines biomolecular adsorption and interactions using a
collage of studies. While each chapter addresses a unique issue, they all share the
underlying common theme of increasing the basic understanding of the behavior of
biomacromolecules at different types of biologically-relevant interfaces. In Chapter 2,
sequential and competitive interactions between human plasma ﬁbronectin and human
serum albumin are examined at a model oil-water interface. Chapter 3 addresses the
development and evaluation of a protocol for quantifying protein interfacial coverages at
the liquid-liquid interface. Finally, Chapter 4 examines the adsorption of liposomes to 3-
dimensional polyelectrolyte multilayers, and characterization of the resulting interfaces
using ﬂuorescence recovery after pattern photobleaching (FRAPP) measurements to
determine the lateral diffusion coefﬁcients of bilayers assembled on the biomimetic

interface.

1.2 Sequential and competitive adsorption between human serum albumin and

human plasma fibronectin at the oil-water interface

The liquid-liquid interface is of vital importance in key areas such as biomedical
engineering and food processing. The hydrocarbon-water interface is a crude ﬁrst order
approximation of a biological cell membrane]. Drug-ion transfer has been studied across
the oil-water interface as a gauge of pharmacological activityz. An examination of the
literature reveals that a majority of the existing studies on protein adsorption have
concentrated on the solid-liquid interface, with relatively few studies targeting the liquid-
liquid interface. However, a number of techniques have been used to probe the oil-water
interface. These include atomic force microscopy to probe orogenic displacement of
proteins from the oil-water interface3, neutron reﬂectivity to probe buried oil-water
interfaces‘, total internal reﬂection ﬂuorescence microscopy to study protein
adsorption“, optical second harmonic generation studies7, quasi-elastic surface light
scattering to study interfacial tensions at the heptane-water interface8 and radiolabeling to

study surfactant adsorption to the oil-water interfaceg.

Recently, the use of perﬂuorcarbon liquids as oxygen carriers and as soft interfaces
for cell-growth has come into focus”. Proteins play a major role in cell attachment,
differentiation and growth at these interfaces. With recent advances in tissue engineering,
there is a need for better understanding of several aspects of protein adsorption in order to
better design such biomimetic interfaces. In particular, effects of adsorption time scales,
interfacial competition, and extent of adsorption irreversibility are among some issues of

key interest. Hence studies of protein adsorption at such ﬂuid interfaces can aid in better

designs for tissue engineering at the liquid-liquid interface as well as shed light on the

underlying mechanisms of protein unfolding.

Chapter 2 of this dissertation presents a study of sequential and competitive
interactions between human plasma ﬁbronectin and human serum albumin at a model oil-
water interface. Protein adsorption experiments were conducted under transport-limited
conditions, characterizing both sequential adsorption and competitive adsorption. These
studies of the interactions between human plasma ﬁbronectin and human serum albumin
have revealed trends consistent with those predicted for protein adsorption in other works
reported in the literature. This study is unique in that it is among the ﬁrst studies of its

kind that uses TIRFM to probe such interactions at the liquid-liquid interface.
1.3 Estimating protein interfacial coverage at the liquid-liquid interface

Quantifying the amounts of protein adsorbed to the. solid-liquid interface is
accomplished easily using techniques such as ellipsometry, radiolabeling, neutron
reﬂectivity, quartz crystal microbalance, total internal reﬂection ﬂuorescence
microscopy, surface plasmon resonance, and optical waveguide lightrnode spectroscopy.
The interfacial signals detected using these methods can be converted into surface
coverages with models that rely on the evolving refractive index of the adsorbing ﬁlm, by
monitoring changes in the oscillation frequency or by using other internal or external
calibration methods. Protein adsorption to interfaces is governed by several factors that
inﬂuence interfacial saturation coverages. For example, electrostatic interactions between

the protein and the surface play a critical role in determining how much protein is

deposited at the interface. These electrostatic interactions can also inﬂuence other factors

such as conformation and structure of the evolving adsorbate layer.

Proteins undergo lateral relaxations during denaturation at interfaces that
eventually increase the total area occupied per molecule. The rate. of transport of proteins
to the interface is another important factor that can signiﬁcantly inﬂuence interfacial
coverages. Surface heterogeneity and roughness can also inﬂuence protein adsorption,
though a recent research article has shown that nanoscale roughness has little inﬂuence
on the amount or structure of protein adsorption1 1. The amount of protein adsorbed at an
interface also depends on the type of the protein adsorbing. For example, typical values
for interfacial coverages for albumin are'in the range of 0.3-3 mg/m2 at the solid-liquid

interface, depending on the bulk protein concentration and surface properties”'”.

Estimates of protein interfacial coverages at the liquid-liquid interface have been
done primarily using radiolabeling15 . It is also possible to estimate interfacial coverages
from tensiometry data by using the Gibbs isotherm. In an article published by Graham et
al. in 19759, the authors commented on the fact that one of the outstanding problems with
respect to liquid-liquid interfaces was the direct determination of interfacial coverages.
Indeed, with the exception of radio-labeling, there does not appear to be an easy, non-
invasive technique that can produce reliable estimates of interfacial coverages at the oil-

water interface to date.

Chapter 3 of this dissertation presents a protocol to quantify adsorbed amounts of

proteins at the liquid-liquid interface using a method that has been previously applied to

making these estimates at the solid liquid interface”. Included in these studies are

examinations of the lifetimes of the dye-labeled proteins at the liquid-liquid interface.
The ﬂuorescence lifetime of a molecule is a measure of the amount of time a molecule
spends in the excited state before its eventual decay to the ground state. Lifetimes of
ﬂuorescent molecules are strongly dependent on the surrounding environment, and can be
related to the quantum yield of the dye. Lifetime studies can, therefore, provide us with
information about the quantum yields of bulk and interface-associated ﬂuorophores. The

inﬂuence of mobile interfaces on interfacial coverages is also discussed.

1.4 Fabrication and characterization of 3D arrays of lipid-bilayers on

polyelectrolyte multilayers

In the last decade, a considerable amount of research has been conducted on the
development of biosensors. Biosensors are molecular devices that combine a biological

7, and are used for

recognition mechanism with a physical transduction techniquel
detection of compounds that may be present in small concentrations. They have been
used for the detection of analytes such as glucose18 and fructose”, among others.
Biosensors use a variety of methods for detecting the analyte of interest, including
electrochemical, optical and colorimetric methods. Biosensors are engineered for
detection of signals based on their ability to respond to different molecular mechanisms
such as ion-channel activation (sensing of current)”, ﬂuorescence resonance energy

transfer (FRET)20 and redox phenomena (cyclic voltammetry).21 These represent only a

small sampling of the available mechanisms on which biosensors can be based.

Biosensors that rely on ion-channel activity and/or use membrane bound proteins

or enzymes, incorporate lipid bilayers in their design. The lipid-bilayers are necessary in

order to preserve the functional characteristics of the proteins that are bound to them. For
biosensors based on activation of ion-channels, the channel proteins are typically
reconstituted into liposomes which are then deposited on the sensing substrate. Lipid
bilayers are generally deposited on the biosensor substrate using Langmuir-Blodgett
deposition methods or liposome adsorption. The choice of lipids used to construct
supported lipid bilayers is important as the characteristics of the resulting bilayer have a
very strong inﬂuence on the activity of the subsequently reconstituted transmembrane
protein/enzyme or ion-channel. Properties such as membrane ﬂuidity, which is an
indicator of the ability of lipid molecules to diffuse freely through the bilayer, play
important roles in maintaining the correct conformation of the biomolecules embedded
within them. It is also important to have bilayers that are highly insulating, since the

presence of large defects will render such sensors ineffective”.

There is considerable research on biomimetic interfaces, where the ultimate goal
is to develop viable, sustainable bilayer interfaces with the smallest number of defects. In
recent years, developments have been made in the area of polymeric cushions upon
which lipid bilayers can be deposited23. In addition to being able to provide ionic
reservoirs, polymeric cushions allow transmembrane proteins to attain their necessary
conformation and to retain biological activity, which would otherwise be impeded by
proximity of the protein to the substrate surface. The use of polyelectrolyte multilayers

has been proposed as one such viable method for a polymeric cushion.

Chapter 4 of this dissertation presents a study to fabricate and characterize
biomimetic interfaces that are based on polyelectrolyte multilayers. This work was done
in collaboration with Neeraj Kohli (co-advised by Dr. R. Mark Worden and Dr. Ilsoon

7

Lee) in the Department of Chemical Engineering and Materials Science. The work
consists of equal contributions from Neeraj Kohli and the author of this dissertation. The
biomimetic interfaces employed in this study consist of arrays of lipid bilayers that have
been deposited on polyelectrolyte multilayer (PEM) substrates composed of positively
and negatively charged polyelectrolyte layers. The bilayers were deposited on the
substrate by preferential adsorption of charged liposomes on oppositely charged
polyelectrolytes. Such arrays can. be used in applications that may require high
throughout screening. The deposition of liposomes on the PEMs was monitored using
total internal reﬂection ﬂuorescence microscopy, ﬂuorescence microscopy and quartz
crystal microbalance. The resulting bilayers werecharacterized using FRAPP. Bilayer
formation on two different kinds of PEM substrates was studied. Possible interactions
between the lipid bilayer and the PEM surface are discussed, based on the diffusion

coefﬁcients obtained in the two cases.

2 ADSORPTION AND INTERACTION OF FIBRONECTIN AND

ALBUMIN AT THE LIQUID-LIQUID INTERFACE

2.1 ABSTRACT

The goal of this work was to investigate the dynamics of human plasma
ﬁbronectin (HFN) at the oil-water interface and to characterize its interactions with
human serum albumin (HSA), using total internal reﬂection ﬂuorescence microscopy
(TIRFM), along with measurements of interfacial tension. Among key results, ﬁbronectin
adsorption at the oil-water interface was observed to be rapid and essentially irreversible,
even over short time scales. For example, after washout experiments with protein-free
buffer over a 3-hour rinse cycle, ﬁbronectin desorption from the interface was found to
be negligible. This may be due to the highly ﬂexible nature of the protein, which allows
the various domains to quickly attain energetically favorable conformations. On the other
hand, HSA adsorption at the oil-water interface is relatively reversible at short times, and
the protein is readily displaced by ﬁbronectin even after HSA has been adsorbed at the
interface for as long as two hours. At longer adsorption times, HSA is able to more
effectively resist complete displacement by ﬁbronectin, although signiﬁcant ﬁbronectin
adsorption was observed even under those conditions. Displacement of adsorbed
ﬁbronectin by HSA was found to be essentially negligible under all conditions.
Fibronectin also adsorbs preferentially from a mixture of HFN and HSA, even when the
concentration of HSA is substantially higher. This study is relevant to such emerging

research thrusts as the development of biomimetic interfaces for a variety of applications,

where there is a clear need for better understanding of the effects of interfacial
competition, adsorption time scales, and extent of adsorption irreversibility on interfacial

dynamics.

2.2 INTRODUCTION

Adsorption of plasma proteins to surfaces is one of the ﬁrst and most important
events before the initiation of key cellular activities such as cell attachment, migration,
differentiation and proliferation. With recent advances in tissue engineering and the
development of synthetic bio-interfaces, there is a need for better understanding of
several aspects of protein adsorption in order to better design such biomimetic interfaces.
In particular, effects of adsorption time scales, interfacial competition, and extent of
adsorption irreversibility are among some issues of key interest. Vitronectin, laminin and
ﬁbronectin are among the more important proteins involved in cell attachment processes,
and several studies have been published on their adsorption and interactions on a variety
of substrate324'27. The present study focuses on the sequential and competitive adsorption

of ﬁbronectin and human serum albumin at a model interface.

Fibronectin belongs to the family of high molecular weight glycoproteins that are
found in the extracellular matrix and in serum, and which are responsible for key
functions including provision of a structural framework for cell attachment, migration
and differentiation, as well as cell-cell and cell-substrate adhesion through integrin
receptors”. With a molecular weight of 440-500kDa, ﬁbronectin is a relatively large

molecule in comparison to other well characterized proteins such as lysozyme (14kDa),

10

serum albumin (67 kDa) and beta-casein (24kDa). Its diffusion coefﬁcient has been
reported to be 2.1 x 10'7 cm2 lszs. Fibronectin has been reported to have globular and
ﬁlamentous forms depending on solution conditions”, and dimensions varying from an
average length of 15 nm30'31 to 60nm3233. It is a ﬂexible protein that can adopt a 'number
of conformations depending on the morphology of the surface. It can adsorb in both an
extended and a globular conﬁguration“, and has been shown to undergo conformational
changes upon binding to dextran, as the molecule goes from a compact to an elongated
conformation35'36. Bergkvist et al.37 reported end to end distances of greater than 100nm

for ﬁbronectin adsorbed onto mica and silica surfaces.

The adsorption characteristics of ﬁbronectin also depend on the hydrophobicity of
the surface to which it is adsorbing. Atomic force microscopy studies of adsorption of
HFN-coated microspheres onto hydrophobic and hydrophilic surfaces have shown that
ﬁbronectin has a strong interaction with both surfaces”. Ellipsometric measurements of
HFN adsorption on modiﬁed silica show greater HFN surface coverages and
irreversiblility on hydrophobic surfaces than on hydrophilic surfaces, which demonstrate
a slightly lower HFN coverage and some degree of reversibility”. Studies of HFN
adsorption on titanium indicate that surfaces appear more hydrophobic in the presence of
the protein27. It has been recently reported that ﬁbronectin exhibits history—dependent
adsorption behavior, i.e., the rate of adsorption depends on the structure of the evolving

adsorbed layer”.

Several techniques have been used for measuring protein adsorption, including
total internal reﬂection ﬂuorescence (TIRF)4°, optical wave-guide light spectroscopy”,

ellipsometry‘", radiolabeling42 and, more recently, atomic force microscopy”. Most of

11

these studies have focused on adsorption of biomacromolecules at the solid-liquid
interface, with relatively few studies concentrating on adsorption dynamics at the liquid-
liquid interface5’6’15’42’44'45. Yet, there is general agreement that the liquid-liquid interface
is of considerable importance in biological systems, particularly in biomedical

en gin eerin g10,46.47

, pharmaceutical sciences, and food processing. The liquid-liquid
interface also serves as an excellent model for interactions of biomacromolecules at

relatively mobile interfaces.

In an earlier articles, our laboratory reported studies of protein adsorption at a
model oil-water interface using total internal reﬂection ﬂuorescence microscopy
(TIRFM). The present study reports on dynamic interactions between ﬁbronectin and
human serum albumin (3 protein present in substantial concentrations in blood) under
both competitive and sequential adsorption scenarios. In particular, information is
presented on the inﬂuence of timescales on adsorption irreversibility, and on subsequent
replacement by a competing protein at the liquid-liquid interface. Recent studies of
competitive adsorption between albumin and ﬁbronectin have shown that albumin co-
adsorption signiﬁcantly inﬂuences the availability of ﬁbronectin on polymeric
hydrOphobic substrates“. Since human serum albumin is typically used as a blocking
agent for elimination of non-speciﬁc adsorption, its competitive interaction with

ﬁbronectin is of considerable biological interest.

12

2.3 Theory

Total internal reﬂection occurs when light propagating through an optical dense
medium of refractive index n1, arrives at the interface between the denser medium and an

optical rarer medium n2, at angle exceeding the critical angle,

I9( = sin" (3] (2.1)

Upon total internal reﬂection, an electromagnetic ﬁeld is induced in the rarer medium at
the interface. This electromagnetic ﬁeld is known as the evanescent wave. The intensity
of this ﬁeld decays exponentially with distance 2, "away from the interface. This distance-

dependent intensity is given by

1(z) = 10 exp(-di) (2.2)

P

where [(2) is the intensity at any depth z, and 10 is the intensity of the light at z=0.

The penetration depth atp represents the characteristic depth at which the light

intensity approaches 1/e of its maximum magnitude Io, and is given by

d A" (2.3)

p Minn/sin2 6 —(n2 /n1 )2

where xiv is the wavelength of the incident radiation in vacuum, and 6 is the angle of
incidence. The penetration depth in our apparatus lies within 100nm of the interface,

depending on the geometry and the properties of the denser and the rarer medium. Thus,

13

this is an interfacially selective technique as only molecules within the proximity of the

interface are illuminated.
2.4 Experimental Protocol

2.4.1 Materials:

Human plasma ﬁbronectin (HFN), at a purity greater than 95% as determined by
SDS-PAGE”, was purchased from Chemicon International (FC-OlO, Temecula, CA).
Human serum albumin (HSA) was purchased from Sigma (A8763, Sigma-Aldrich Corp.
St. Louis, MO). Both proteins were used without further puriﬁcation. The proteins were
labeled with ﬂuorescein—S-isothiocyanate (FIT C), purchased from Molecular Probes
(Catalog # F-1907, Eugene, OR). The labeling reaction was carried out in the dark, as
described in the literature50 in 0.1 M carbonate buffer at a pH of 9.2 for a period of 6
hours. The unbound labels were removed by a 2-stage dialysis process performed over a
period of 36 hours. This procedure was determined to be sufﬁcient to remove all
unreacted FIT C. The reaction mixture was dialyzed against 0.05 M phosphate buffer at a
pH of 7.4, using a molecular porous regenerated cellulose dialysis membrane
(Spectra/Per l, MWCO 6000, The Spectrum Companies, Gardena, CA). The solutions
were either divided into aliquots and frozen for later use, or stored at 4°C and used within

one week of preparation.

In the case of samples containing ﬁbronectin, no vigorous stirring or vortexing
was used; instead all mixing was done using gentle swirling, to avoid precipitation of the
protein”. Absorbance spectroscopy was used to measure the protein concentrations and

labeling ratios at 280nm and 496nm, respectively, using a diode array spectrophotometer

l4

(8452 A, Hewlett-Packard, Brielle, NJ). After subtracting the corresponding absorbance

contribution from FITC at 280 nm, protein concentrations were computed using optical

densities (A‘""-"""’ )of 1.349 and 0.535] for ﬁbronectin and albumin, respectively. It was

rm

observed that conjugation of FITC to human serum albumin causes an increase in
absorbance of the protein at A230 to an :extent greater than the contribution from .
ﬂuorescein’s computed concentration dependent absorbance. We believe that this is
because the extinction coefﬁcient for human serum albumin at 280 nm is inﬂuenced by
FIT C conjugation, which may result in an overstatement of labeled protein
concentrations. For the purpose of this study, no correction was made to account for this
and further studies are continuing in our laboratory to understand this effect. Dilutions
from protein stock solutions were made just prior to infusion into the ﬂow cell, to
eliminate or minimize the tendency of ﬁbronectin to rapidly adsorb to the sides of the
storage container. The oil used in this study was an immersion oil with a refractive index
of 1.46 at 589.3nm, largely comprising long chain aliphatic hydrocarbons (Cargille

Laboratories, Cedar Grove NJ Cat. # 19572; code 06350).

Quartz slides (n=l.46, Heraeus Quarzglas) were used as substrates onto which the
oil was deposited to form layers of ﬁlm thicknesses typically ranging from 20 to 50 pm,
as estimated using a microbalance. All slides were cleaned using the protocol described
by Cheng et al.40. The oil was deposited in a smooth ﬁlm on the quartz slides using
previously described protocols“. Prior to assembling the ﬂow cell, the bottom slide was
coated with polyethylene oxide (Sigma, St. Louis, M0) by immersion in 600 ppm

solution to reduce or eliminate protein adsorption on that surface52‘54.

15

2.4.2 TIRFM Setup

The experimental setup has been described previouslys'ss. The experimental
layout and a diagram of the ﬂow cell is depicted in Figure 2.1a and Figure 2.1b. Brieﬂy,
the apparatus consists of an inverted microscope (Zeiss Axiovert 135M, Carl Zeiss Inc.
Thomwood, NY)),‘the 488 nm line-of a 5W continuous wave argon ion laser (Lexel
Lasers Model 95, Fremont, CA), a side—on photo-multiplier tube (PMT) (Hamamatsu
R4632, Bridgewater, NJ) jacketed in a thermoelectrically cooled housing (TE 177-TSRF,
Products for Research, Danvers, MA), a CCD camera (NTI, VE1000, Dage-MTI,
Michigan City, IN), and a modular automation controller (MAC 2000, Ludl Electronic
Products, Hawthorne, NY) that regulates the voltage supply to the photo-multiplier tube.
A double syringe pump system (Model 551382, Harvard Apparatus South Natick, MA)
was used to infuse and withdraw sample solutions at precisely controlled rates from a
custom-designed ﬂow cell. The photo-multiplier tube can be operated in both analog and
digital (photon counting) mode; however, several experiments were conducted in strictly
analog mode prior to adding a photon counter to the system. In photon counting mode, an
SR400 photon counter (Stanford Research Systems, Sunnyvale, CA) is used for recording
ﬂuorescence emission, and the output signal is ampliﬁed by a fast preampliﬁer (SR445,
Stanford Research Systems, Sunnyvale, CA). Fluorescence intensities were recorded

using software written in Labview 6.0 (National Instruments, Austin, TX)‘.

 

‘ Labview ﬂowsheets are located in Appendix A: Labview ﬂowsheets. Sample Labview templates for
photon counting using the SR400 were obtained from http://www.ni.com and modiﬁed accordingly for our
application.

16

An optical chopper (OC) (SR 540, Stanford Research Systems, Sunnyvale CA)
was used to prevent unintended photobleaching of ﬂuorophores during all experiments.
Analog measurements were ﬁltered using a program written in Matlab2 (Release 12, The
Mathworks Inc., Natick, MA) to eliminate dark current measurements during the dark
- phases of the chopper cycle. For digital measurements, the'photon counter is triggered by
an output reference voltage from the chopper, so that data collection only occurs during
periods when the sample is illuminated by the laser beam, thereby enhancing the signal to

noise ratio.

An optical chopper was used to signiﬁcantly reduce the possibility of
photobleaching by the monitoring beam during long experiments. This is because
photobleaching can occur even at the low monitoring beam intensity of 4 11W used in
these experiments, and makes the accurate interpretation of experimental results difﬁcult.
The optical chopper reduces the exposure time of the sample to the laser beam; typically,
in a 17-second data collection cycle, the total time the sample cell is exposed to the laser
beam amounts to less than two seconds. Since the illumination of ﬂuorophores only
occurs during data sampling periods, this ensures that no unintended photobleaching
occurs. The intensity of the monitoring beam was carefully controlled in all experiments.
As a general protocol, no experiments were conducted with a monitoring beam intensity

exceeding 13.5 11W, even over short periods of data collection.

A system of optical ﬂats5 (F1 and F2) enables us to easily switch between a low

intensity monitoring beam for observation of adsorption dynamics and a high intensity

 

2 Matlab programs listed in Appendix B: Matlab programs for ﬁltering and averaging analog measurements

17

beam for conducting ﬂuorescence recovery after photobleaching (FRAP) experiments.
The laser beam passes through the ﬁrst optical ﬂat and is split into a high intensity
photobleaching beam and a low intensity monitoring beam. An adjustable shutter (D122,
Uniblitz, Vincent Associates, Rochester, NY) placed between the optical ﬂats, blocks the
high intensity beam in the “closed” mode while allowing the low intensity monitoring
beam to pass through. Upon triggering, the shutter allows the passage of the high
intensity beam for a controlled amount of time. The optical ﬂats are aligned so that the
high intensity and low intensity beams are recombined after passing through the second
ﬂat. A second shutter placed below the PMI‘ is synchronized with the ﬁrst shutter in
order to prevent the photomultiplier from being exposed to high intensity light during the
photobleaching step. This is done to prevent damage to the photomultiplier tube in

addition to avoiding light induced hysteresis in the PMT signal.

The ﬂow-cell where adsorption experiments are monitored is depicted in Figure
Figure 2.1b. The ﬂow cell is formed using two microscope slides separated from one
another by a 900 um aluminium spacer with an oval cut into it. An o-ring (Parker,
Lexington, KY) inside the oval spacer between the top and bottom slide prevents leakage
of sample solution. The underside of the top microscope slide is coated with a thin layer
of oil. The bottom slide is ground to a thickness of 0.75 mm and has two holes drilled
into it (1.60 mm diameter, separated by 42.3 mm, centered) to allow for the infusion and
withdrawal of solutions. The entire assembly is coupled to a prism and is ﬁrmly screwed

on to an anodized aluminium shell that sits on top of an inverted microscope stage.

All TIRFM experiments were conducted under conditions of gentle shearing ﬂow.

In most experiments, measurements were made under continuous ﬂow conditions to

18

eliminate the possibility of bulk depletion of proteins, which is a possibility at the very
low protein concentrations used in this study. In experiments conducted over long
timescales, ﬂow was allowed into the sample cell until apparent saturation of the
interface has been achieved, which typically took approximately one hour. After that,
ﬂow was terminated, and experiments were conducted under static conditions. This was

done to prevent destabilization of the interface by continuous flow over long periods.

Interfacial tension measurements: To provide comparative data for the TIRFM
experiments, static and dynamic interfacial tension experiments were also conducted at

the liquid-liquid interface, using a Kruss K12 tensiometer (Kruss USA, Charlotte, NC).
2.5 Results and Discussion

All the experimental data are presented in terms of ﬂuorescence emission
intensity proﬁles normalized by the maximum ﬂuorescence value in each data set. It is
also possible to convert ﬂuorescence emission data to interfacial surface densities using a
transport-limited model'2'56. It is difﬁcult to use this model for these data because, to
avoid inducing instabilities at the liquid-liquid interface, there are limitations on the range
of shear rates over which the TIRFM apparatus can be operated, which prevents model
validation. For the experiments discussed in this paper, no external calibrations were used
to standardize the ﬂuorescence emission proﬁles. Also, variability in the thickness of the
oil layers formed on quartz substrates, along with potential instabilities in the oil layer
during the course of an experiment, results in variations from one experiment to another,
thus complicating comparative data analyses. The uneven nature of the oil-water interface

can introduce artifacts in the ﬂuorescence data due to light scattering. While every

19

precaution is taken to minimize scattering effects, it cannot be fully eliminated. A typical
oil-water interface can be visualized as a wavy interface ﬁlled with undulations where
one phase protrudes into the other. This makes such interfaces less reproducible than a
solid-liquid interface such as a hexadecyltrichlorosilane self assembled monolayer (HT S-

SAM) - water interface . ‘

As a result, we have refrained from assigning undue physical significance to small
observed variations in interfacial dynamics, and have concentrated instead on
characterizing general trends. This limitation made it necessary to develop an internal
calibration procedure that would allow precise determination of absolute interfacial
concentrations from ﬂuorescence emission intensity measurements at the liquid-liquid
interface. This protocol will enable more rigorous comparisons to be made between data

sets and will be discussed in the next chapter.

2.5.1 Sequential adsorption of HSA and HFN

Data on the sequential adsorption of unlabeled HFN following the pre-adsorption
of labeled human serum albumin (HSA-FITC) are shown in Figure 2.2. In this
experiment, HSA-FIT C was introduced into the ﬂow cell alone at a ﬂow rate of 0.05
mllmin, and adsorption was allowed to proceed for approximately 70 min. At t=74 min.,
a mixture of HSA-FITC (at the same bulk concentration as before) and unlabeled HFN
was infused into the ﬂow cell, and the subsequent emission intensity of HSA was
monitored. All experiments were conducted under continuous ﬂow conditions to ensure

that bulk depletion of protein did not occur. It should also be noted that HSA was infused

20

into the ﬂow cell along with HFN to ensure that any observed changes in ﬂuorescence

emission intensity would not be the result of the depletion of HSA in the bulk.

The introduction of the binary protein mixture signiﬁcantly reduced the
ﬂuorescence emission intensity from its near plateau value to a level near the baseline,
indicating that unlabeled HFN is displacing HSA-FITC from the interface in substantial
amounts, even though the bulk concentration of HSA-FITC is greater than that of HFN.
Since the bulk concentration of HSA-FITC was the same before and after HFN was
introduced, its disappearance from the interface cannot be attributed to concentration-

induced mass transfer of the protein from the interface into the bulk.

Figure 2.3a and Figure 2.3b represent data from experiments complimentary to
the one above, with the modiﬁcation that HFN was labeled with FITC (HFN-FIT C),
while HSA was unlabeled. In the experiment depicted in Figure 2.3a, HFN-FITC at a
concentration of 0.1mg/ml is allowed to adsorb for 60 minutes, followed by infusion of a
solution containing unlabeled HSA at 0.1mg/ml and HFN-FITC (at the same bulk
concentration as before). As the data show, there is negligible change in the magnitude of
the ﬂuorescence emission intensity following introduction of the mixture of proteins into
the ﬂow cell. Even when the experiment was repeated with a 10-fold increase in the bulk
concentration of HSA (1mg/ml), there is no discernible change in the ﬂuorescence

emission intensity (Figure 2.3b).

The steady ﬂuorescence emission levels in the two experiments following
introduction of the protein mixture indicate that adsorbed HFN was not displaced from

the interface by HSA, even at the much higher bulk albumin concentration. Since the

21

adsorption of non-labeled species at the interface cannot be detected, we cannot
completely rule out HSA adsorption into empty pockets due to incomplete HFN
adsorption at the interface. However, there is little discernible increase or decrease in the
HFN-FITC ﬂuorescence emission intensity, suggesting that while there may be
adsorption/desorption occurring between molecules of the same species, there is no
visible change in the total quantity of HFN at the oil-water interface. Comparison of these
results with those obtained in the experiments described earlier using HSA indicate that,
while HFN easily displaces adsorbed HSA from the interface, HSA demonstrates little
tendency to displace HFN from the interface, even at substantially higher bulk

concentrations.

These results are not surprising. Data from long term wash-out experiments
indicate that HFN adsorbs essentially irreversibly at the oil-water interface (Figure 2.30).
In the ﬁrst experiment (represented by the upper curve in Figure 2.3c), a pure buffer
solution was circulated in the ﬂow cell as HFN-FITC adsorption approached equilibrium;
in the second experiment (lower curve in Figure 2.3c), a protein solution containing only
unlabeled HSA was circulated through the system, beginning at about the same time as
the buffer wash in the previous case. These experiments were conducted to determine
whether the introduction of the HFN-free solutions would cause the protein to detach
from the interface. Both sets of data reveal very little HFN desorption and/or removal
from the interface over the period of the attempted washout, with only a small reduction
in the ﬂuorescence emission intensity over the duration of the rinse. We believe the small
reduction in ﬂuorescence emission is the result of the elimination of the bulk

ﬂuorescence contribution to the overall signal as the ﬂow cell is ﬂushed free of labeled

22

bulk proteins, rather than an indication of the removal of interface-bound species. Thus
adsorption of HFN at the oil-water interface appears to be practically irreversible. Other
researchers have also reported ﬁbronectin adsorption irreversibility on hydrophobic

surfaces and partial reversibility at hydrophilic surfaces”.

The results of the sequential adsorption experiments above were conﬁrmed in two
additional experiments: (i) adsorption of HSA-FITC following pre-adsorption of
unlabeled HFN at the interface, and (ii) adsorption of HFN-FIT C following pre-
adsorption of unlabeled HSA. The differences between these experiments and the ones
presented earlier are that a) the protein initially introduced into the ﬂow cell is unlabeled;
and b) only the competing protein of interest (rather than a mixture of the two proteins) is
introduced into the ﬂow cell following adsorption of the ﬁrst. Therefore, diffusion due to
concentration gradients can conceivably induce a depletion in the interfacial

concentration of the protein that was adsorbed at the interface.

The proﬁle in Figure 2.4a is an example of a sequential adsorption experiment of
type (i). In a typical experiment, unlabeled HFN was pre-adsorbed at the indicated
concentration for an hour, a duration long enough to obtain adsorption levels near
pseudo—equilibrium coverage in the experimental apparatus. The ﬂuorescence emission
intensity at the interface was periodically measured, with the results essentially giving the
dark current or background noise. HSA-FITC was then introduced into the ﬂow cell after
60 minutes of pre-adsorption of HFN (0.1 mg/ml), and its adsorption proﬁle was
monitored. As Figure 2.4a shows, the introduction of HSA-FITC (0.130 mg/ml) induces a
negligible change in the ﬂuorescence emission intensity, which remains very close to the
background level. After about 40 minutes of apparently negligible HSA-FITC

23

adsorption, a higher concentration of labeled HSA was introduced into the ﬂow cell.
Twenty minutes after the introduction of the more concentrated mixture, a small increase
in the ﬂuorescence emission intensity was observed, although this was signiﬁcantly lower

than what would be observed for HSA adsorption in the absence of HFN.

The experiment was repeated, with the only modiﬁcation being a ﬁve-fold
reduction in the bulk concentration of pre-adsorbed I-IFN from 0.1 to 0.02 mg/ml. The
lower curve (open circles in Figure 2.4b) represents the HSA-FITC adsorption proﬁle
after pre-adsorption of HFN at this lower bulk concentration. In this case, a small
increase in ﬂuorescence emission intensity due to HSA adsorption can be observed.
However, the maximum signal reached a plateau at a level just slightly above the
background noise. For comparison, the adsorption proﬁle of HSA-FIT C at the same
concentration (0.140 mg/ml) at a bare oil-water interface (solid circles) has been
included. The resulting ﬂuorescence emission intensities for this control experiment is
more than an order of magnitude higher than for the case where a low concentration of

unlabeled HFN was pre-adsorbed at the interface.

In Figure 2.4a, the increase in ﬂuorescence upon introduction of the higher
concentration of HSA-FITC is to be expected, since the increased bulk concentration of
labeled proteins contributes additional, albeit small, amount of ﬂuorescence emission to
the observed signal. It is likely that, upon continuation of the experiment beyond 125
minutes, the signal would have increased, though it would have been signiﬁcantly below
the levels that would be expected in the absence of HFN. Even upon decreasing the HFN
bulk concentration to 0.02 mg/ml (Figure 2.4b), little increase is observed in the

ﬂuorescence emission intensity, suggesting that even at this low ﬁbronectin

24

concentration, albumin is effectively unable to displace adsorbed HFN molecules from
the interface. It also appears that even ﬁbronectin concentrations as low as 0.02 mg/ml
are sufﬁcient to preclude HSA adsorption. Thus, we conclude that prior adsorption of
HFN at the oil-water interface effectively inhibits subsequent albumin adsorption, even at
short adsorption timescales. The data in Figure 2.4 also conﬁrm that the previous

observations were not due to labeling effects.

It is important to note that all the above experiments were conducted at a low
shear rate of approximately 0.5 s", and at ﬂow rates that did not exceed 0.05 ml/min.
This is because, unlike the solid-liquid interface, the liquid-liquid interface may become
unstable at high shear. By comparison, these types of experiments have been conducted
at the solid-liquid interface at shear rates of 1.5 s'1 or higher”. The low shear rates and
low ﬂow rates used in the experiments can reduce the speed at which proteins arrive at
the interface. Thus, once a protein arrives at the interface, it can have a signiﬁcant
amount of time to adsorb, unfold and undergo conformational changes before a second
protein arrives, all of which can contribute to greater adsorption irreversibility. The
adsorption of early arriving molecules can therefore be less reversible than that of later
arriving molecules. In addition, the low shear rates and resulting expanded relaxation
time can lead to a lower availability of open surface locations at which later arriving
molecules can adsorb. The effect is that the overall surface coverage may be lower than
expected. The results of the sequential adsorption experiments above suggest that HFN
adsorbs rapidly and uniformly at the interface, leaving few or no empty interfacial spaces

that HSA can adsorb to.

25

2.5.2 Effect of adsorption timescales

The three proﬁles in Figure 2.5 are from complimentary experiments conducted
over longer adsorption timescales. The ﬁrst ﬂuorescence emission intensity proﬁle is for
the adsorption of HFN-FITC at time t=0 onto a clean oil-water interface from a solution
containing only labeled HFN (curve 1). The second proﬁle is of ﬁbronectin. adsorption
onto an adsorbed HSA layer that has been at the interface for two hours (curve 2). The
third proﬁle is of ﬁbronectin adsorption initiated at t=6 hrs onto an adsorbed albumin
layer which has been at the interface for six hours (curve 3). As the data show,
adsorption of HFN onto a 2-hour old layer of HSA results in an apparently and
unexpectedly higher surface coverage than for the case when HFN-FITC adsorbs to a
clean oil-water interface. This higher surface coverage after two hours of HSA residence
time was initially surprising, because one would expect that the presence of a pre-
adsorbed layer of albumin would at least partially inhibit ﬁbronectin adsorption, and that
the extent of inhibition would increase with the age of the adsorbed layer. In fact, Wertz
and Santore12 reported nearly complete suppression of ﬁbrinogen adsorption after 4 hours

of albumin adsorption on to a hexadecyltrichlorosilane monolayer on quartz.

However, recent reports on ﬁbronectin adsorption to hydrophobic (methylated)
substrates37 suggest that ﬁbronectin adsorbs in both extended and compact
conformations, but that most of the adsorbed layer is in the compact conformation.
Indeed, the ability of the molecule to undergo transitions from one conformation to the
other is made possible by its ﬂexibility in domain-connecting segments and ionic
interactions. It is known that the compact conformation of ﬁbronectin is stabilized by

intrarnolecular ionic interactions”. A recent report by Michael and coworkers58 proposes

26

a mass action model comprising two states to describe ﬁbronectin adsorption to surfaces.
They speculated that ﬁbronectin initially adsorbs reversibly and then a fraction of the
reversibly bound molecules undergo a transition to the irreversibly adsorbed state,

occupying a larger area in the process.

The cartoon in Figure 2.6 depicts a possible sequence of events that provides a
plausible physical model for the observation at t=0 and t=2 hours in this study. In the case
of ﬁbronectin adsorption onto a bare oil-water interface, early arriving PEN molecules
undergo the transition from the reversibly bound state to the irreversibly bound state,
which allows them to occupy more of the interfacial area. This results in inefﬁcient
packing and hence a fewer number of molecules at the interface. The slow ﬂow rate used
in the experiments probably further enhances this effect. We speculate that, in the
experiments involving 2 hours of albumin pre-adsorption (curve 2 in Figure 2.5), PEN,
displaces loosely bound HSA from the interface upon introduction into the ﬂow cell, and
occupies the area in the spread conformation (occupying larger area). In areas where it is
unable to displace albumin that has been irreversibly bound, it adsorbs in a compact
conﬁguration into the spaces between existing albumin molecules but, due to spatial
constraints, it is unable to transition into the relaxed state. However, the compact state

results in more efﬁcient packing, thereby increasing the effective interfacial coverage.

In the third experiment (curve 3 in Figure 2.5), HSA is essentially irreversibly
adsorbed after 6 hrs”, resulting in signiﬁcantly reduced ﬁbronectin adsorption. We
believe this is due to the occurrence of two simultaneous processes: a) as albumin unfolds

and undergoes molecular relaxation, it reduces the ability of PEN to displace it; b) at the

27

same time, it also occupies more area per molecule, thereby decreasing the size of the

interfacial area where direct attachment of PEN can occur.

This hypothesis appears to be supported by the results of experiments conducted
to assess displacement of labeled HSA by unlabeled PEN. HSA-FITC (0.130 mg/ml)
was adsorbed at the oil-water interface for periods .of one (1) and three (3) hours in
separate experiments (Figure 2.7). After each adsorption proﬁle approached pseudo-
steady state, a solution of unlabeled PEN at a concentration of 0.1 mg/nrl was infused
into the ﬂow cell and the HSA ﬂuorescence emission intensity was monitored for a
period one (1) hour. In the case of albumin that had adsorbed at the oil-water interface for
one (1) hour (solid circles), a signiﬁcant portion of the adsorbed protein was removed by
the PEN wash. In contrast, albumin that had adsorbed to the oil-water interface for three
(3) hours (open triangles) appeared to more effectively resist desorption by PEN.
Therefore, while signiﬁcant amounts of HSA was displaced from the interface by PEN in
both cases, the longer HSA residence time at the interface signiﬁcantly mitigated PEN

adsorption.

It was also observed that pre-adsorption of higher concentrations of albumin (0.5
mg lml) at the bare interface followed by introduction of 0.05 mg/ml PEN-FIT C still
resulted in signiﬁcant adsorption of ﬁbronectin (Figure 2.8). A subsequent wash with
buffer also resulted in little desorption of ﬁbronectin. Sagvolden and others38 measured
interaction forces between ﬁbronectin-coated microspheres and bovine serum albumin-
covered substrates and found low interaction forces between the two surfaces, indicating
no direct binding between ﬁbronectin and alburrrin. Their experiments also suggested that

ﬁbronectin displaced albumin on hydrophilic polystyrene but was unable to do so on

28

hydrophobic polystyrene. These observations of ﬁbronectin adsorption at the oil-water
interface suggest that, even in the presence of pre-adsorbed albumin, ﬁbronectin still
adsorbs in signiﬁcant amounts. This indicates that ﬁbronectin may either displace the
adsorbed albumin or adsorb onto the albumin ﬁlm as a second layer. If the latter case
were true then, since ﬁbronectin does not directly bind to albumin”, ﬂushing of the ﬂow
cell with protein-free buffer should result in signiﬁcant quantities of ﬁbronectin being
desorbed, as such a layer would at best be only loosely bound to the prior adsorbed‘layEr.
However, the desorption of PEN-FITC after introduction of protein-free buffer is not
very signiﬁcant (at a ﬂow rate of 0.05 ml/min) as has shown in Figure 2.8. Based on
these observations, it is likely that albumin has limited afﬁnity for the oil-water interface,

even though the interface is essentially hydrophobic.

2.5.3 Competitive adsorption of HFN and HSA

The lower curve (open triangles) in Figure 2.9 represents the adsorption proﬁle of
a mixture of labeled HSA (0.125 mg/ml) and unlabeled PEN (0.05 mg/ml) adsorbing
onto a bare oil-water interface. For comparison, the closed circles show HSA alone at the
same concentration adsorbing to a bare oil-water interface (solid circles). It is clear that
albumin has very limited afﬁnity for the interface in the presence of PEN. While no
overshoots were observed in the adsorption proﬁles in any of these experiments, it
appears that, over the time scale of the experiments, ﬁbronectin preferentially adsorbs
and effectively limits HSA adsorption. HSA adsorption is visible but very low over the
entire duration of the experiment. One other group has reported similar results. Malmsten
and Lassen59 reported preferential adsorption of ﬁbrinogen (a protein similar to

ﬁbronectin) from a binary mixture of albumin/ﬁbrinogen onto methylated surfaces. They

29

observed that albumin was unable to frustrate ﬁbrinogen adsorption, in comparison to the

ease with which it inhibited gamma-globulin adsorption.

To assess the observations of ﬁbronectin’s strong interaction with the oil-water
interface in the TIRFM apparatus, some long duration experiments were conducted at the
oil-water interface, using a K12 Kruss Interfacial Tensiometer. Figure 2.10 depicts the
interfacial tension proﬁles at the oil-water interface for different compositions of a .
mixture of ﬁbronectin and albumin. In Figure 2.10a, adsorption from a solution of PEN
(0.1 mg/ml) and HSA (0.125 mg/ml) is monitored. For comparison, the curves above and
below are the interfacial tensions of the pure proteins at their respective concentrations.
The strong interaction of ﬁbronectin with the surface and its tendency to easily displace
albumin from the interface is apparent. At short times (t<5000 seconds), the interfacial
tension proﬁle of the mixture follows that of pure albumin adsorbing to the interface due
to the higher diffusion coefﬁcient of albumin (6.1 x 10'7 cm2/s6o) . At longer times,
however, the adsorption proﬁle rapidly approaches that of pure ﬁbronectin. Figure 2.10b

depicts the a similar experiment with a higher concentration of HSA (1mg/m1).

In general, it has been shown that smaller molecular weight proteins arrive earlier
at the interface (due to their higher diffusion coefﬁcients). The smaller proteins are then
preferentially displaced by the higher molecular weight (hence, later arriving) proteins, in
a sequence commonly known as the Vroman sequencem'éz. Recent modeling work on
competitive adsorption between proteins has shown that, for the Vroman sequence to
occur, the bulk solution must have an excess composition of the small protein, and a
stronger interaction between the larger protein and the surface“. This study clearly

provides strong experimental evidence of the Vroman effect.

30

2.6 Conclusions

The principal aim of this work was to study the competitive interaction of
ﬁbronectin and human serum albumin at a model oil-water interface, using TIRFM. All
the experiments reported here were conducted under low shear conditions, to prevent
destabilization of the oil-water interface. Under these conditions, it has been
_ demonstrated that it is possible to use TIRFM to study competitive and sequential
adsorption interactions at the liquid-liquid interface. The results of this work have
revealed adsorption dynamics of the two proteins consistent with previous reports in the
literature for solid-liquid interfaces. These results are among the ﬁrst TIRF measurements
describing the sequential and competitive interactions between competing proteins at a

model oil-water interface.

It was observed that adsorbed albumin at the oil-water interface is readily
displaced by ﬁbronectin, even when albumin is present at much higher bulk
concentrations. It was also observed that albumin demonstrated no tendency to displace
adsorbed ﬁbronectin at the interface, even when it is present in much higher
concentrations in the bulk. Albumin adsorption at short time scales is also reversible at
the oil-water interface. However, at longer adsorption time scales, albumin adsorption is
somewhat less reversible, and the protein is able to prevent subsequent ﬁbrbnectin
adsorption to some extent, although signiﬁcant amounts of PEN still adsorbs at the
interface. On the other hand, ﬁbronectin adsorption appears to be irreversible over all
timescales examined, and albumin adsorption is almost completely frustrated by pre-

adsorption of even small quantities of ﬁbronectin at the interface.

31

This work addresses some key issues on adsorption dynamics between two
proteins of much biological interest. The work was also accomplished at an interface that
has received relatively little attention, but is clearly relevant to biological systems. The
careful measurements of protein dynamics presented in this work were all obtained
within 85 nm of the oil-water interface. Therefore, the data represent some of the most
interfacially sensitive measurements of protein dynamics accomplished by total internal

reﬂection ﬂuorescence microscopy at a liquid-liquid interface.
2.7 Recommendations for future work

Since albumin adsorption is almost completely frustrated by pre-adsorption of
small quantities of ﬁbronectin, it would be interesting to examine this sequential
adsorption behavior under conditions where very low concentrations of ﬁbronectin are
used to pre-treat the oil-water interface. It would also be interesting to conduct
ﬂuorescence recovery after photobleaching experiments on adsorbed layers of proteins,
when a mixture of the two competing proteins is present in the solution. Some
preliminary FRAP experiments have been initiated with PEN at the oil-water interface
and further investigations are in progress to fully characterize the bulk and lateral
diffusion coefﬁcients of this ﬁbrous protein at the oil-water interface. Ongoing work in
our laboratory is targeted at investigating the inﬂuence of the oil-layer thickness on the

reversibility of HSA and PEN adsorption to the liquid-liquid interface“.

Atomic force microscopy is another useful technique for observing the adsorbed
state conformation of proteins and other bio-macromolecules. Adsorbed ﬁbronectin has

already been characterized by AFM on silica and mica surfaces”. Cluster formation

32

during the adsorption of Vitronectin, another glycoprotein similar to ﬁbronectin, has also
been reported using AFM“. PEN and HSA molecules adsorbed to HTS coated surfaces
have already been imaged. It would be interesting to image a similar surface after
sequential adsorption of HSA and PEN is complete, to visualize the conformation of the
adsorbed proteins. Finally, direct visualization of protein adsorption using AFM at the
oil-water interface can give us much insight into the topography of such interfaces.
Mackie et al.3 used an indirect method of AFM imaging to study orogenic displacement
of protein ﬁlms adsorbed at the oil-water interface by transferring the protein ﬁlms from
the oil-water interface to a mica surface using Langmuir-Blodgett methods. Given the
‘roughness’ of the oil-water interface, studies involving AFM imaging are very
challenging and, to date, direct imaging of protein adsorption at liquid-liquid interfaces

has not been reported in the literature.

33

 

:‘ii 7”. 1':

 

 

 

 

._spacer
\

Bottom slide

  

spacer

 

 

syringe

pump S

 

 

 

 

 

 

Figure 2.1: a) Experimental layout for TIRFM set-up: NDF: Neutral density ﬁlter, 0C:
optical chopper, F1,F2: optical ﬂats, Sh: shutter, M1,M2: Mirrors, C: coatings b) Layout
of ﬂow cell for TIRF experiments.

34

1.2

 
   
    

 

1
oo o o
3 0.6 o. HSA-FlTC:0.130mg/ml .
g . rm :o.1mgImI o.
- 0.4 o .0
. 9 ' o
o . ' o o
0'2 PBA-HTC:0.130mg/ml . .
o 14 1 r r r 1 r
0 20 40 50 80 100 120 140

time (minutes)

Figure 2.2: Sequential adsorption of unlabeled PEN (0.1 mg/ml) at the oil-water
interface, following adsorption of labeled HSA (0.130 mg/ml) to near interfacial
equilibrium. In this experiment, a mixture of labeled HSA and unlabeled PEN was
introduced into the ﬂow cell approximately 75 minutes after the adsorption of labeled

HSA had been initiated. The plot gives strong evidence of displacement of adsorbed HSA
at the interface.

35

1.2

 

 

 

 

1 . . O . O . u w . . O O
z. 0.8 O I
'3
g 0.6 . HFN-FITC:0.1mgImI
' 0.4 HSA :0.1mglml
0.2 .
o o...Hr'~.~riT.<=.=<I-1rwrni..,..........
0 20 40 60 80 100 120
time (min)
12 El
1 o e e M. e e e
e O '
g 0.8 ’0
5 0-6 HFN-FITC : 0.1mgrml
E 0.4 e HSA : 1 mglml
0.2 .
o ..... EMJEiI-inism'...
0 20 40 60 80 100 120
time (min)
1.2 Flow rate: 0.05 mllmln E]

1 OQ
f Ibutier..... '0 0.
30.8

C .m"
'6 0“ ‘l‘A‘AAAAAAA‘
5 0.6 ‘
5 04
' ‘ "Shadmyml oHFN-FlTC:0.075mg/ml
0.2 O A HFN-FITC : 0.05 mgtml

o rlrrrrirrrrlrtrrlrrLrLrerJrrrJIr111111141

0 30 60 90 120 150 180 210 240 270

time (min)

Figure 2.3: Sequential adsorption of PEN/HSA at the oil-water interface. In this
experiment, a mixture of labeled PEN and unlabeled HSA was introduced into the ﬂow
cell approximately 60 minutes after adsorption of labeled PEN had been initiated. The
proﬁles show that HSA had no apparent effect on PEN adsorption. In the top graph (a),
HSA and PEN are at the same bulk concentration (0.1 mg/ml); in the bottom graph (b),
the concentration of HSA is an order of magnitude higher (1 mg/ml). (c) Desorption of
PEN from the interface by interfacial rinsing. Based on this data, adsorption of PEN at
the oil-water interface is essentially irreversible. Little reduction in PEN ﬂuorescence
emission intensity is observed upon introduction of a protein-free rinse (solid circles) or
circulation of unlabeled HSA (solid triangles) after 60 minutes of PEN at the interface.

36

0.8

I

0.6
194:0.1 mglml PBA-FiTC:0.130mg/ml HSA-me: >1mglml

“I I .I .
02E... fee e

rrrrrrrrrr

Intensity

 

 

 

o L l L L i
0 50 1 00 1 50
Time (min)
1 . . C
F . 0 e HSA-Fire (0.140 mglml)
0.8 i e
i OI'FN(0.02mgImI)$I'BA-
I Q FITC (0.140lnglml)
g 0.6 :
E : ' b
5 0.4 E 9
I O
0 2 -_ HSA-FITC
. _ .
. 1 00 O O O
0
0 20 40 60 80 100

HFN 0.02 mglml Time(mln)

Figure 2.4: Sequential adsorption of HSA and PEN at the oil-water interface. Figure 2.4a
shows that pre-adsorption of PEN completely frustrates subsequent HSA adsorption. In
this experiment, unlabeled PEN was adsorbed for 60 minutes. Subsequently, labeled
HSA was introduced into the ﬂow cell. In the top graph (a) PEN is present in the bulk
solution at a concentration of 0.1 mg/ml; in the bottom graph (b), the concentration of
PEN is at a much lower bulk concentration of 0.02 mg/ml. Curve 2 in this plot (solid
circles) is for HSA adsorbing by itself at the same bulk concentration, and is given for
comparison.

37

d
N
l

 

 

 

1f— 2
08E1
3":
2062’ . 3
3':O D
:0
50.4:. D f
: A
0'25. D A
i A
o 1L1I.1_..
0 2 4 6

Time (hrs)

Figure 2.5: Effect of pre-adsorption of HSA on PEN adsorption at the oil-water interface:
Curve (1) (solid circles) shows adsorption of labeled PEN to a bare oil-water interface.
Curve (2) (open squares) shows adsorption of labeled PEN to an oil-water interface at
which unlabeled HSA had been pre-adsorbed for 2 hrs. Curve (3) (open triangles) shows
labeled PEN adsorption at an oil-water interface at which unlabeled HSA had been pre-
adsorbed for a period of 6 hrs. Bulk concentrations of HSA and PEN are 0.1 mg/ml and
0.05 mg/ml, respectively.

38

initial relaxed 0 albumin

DE: 0 O relaxed albumin

-. Fibronectin

:__, relaxed ﬁbronectin

Lﬁ __

 

 

 

B
2hr albumin adsorption ‘ introduction of ﬁbronectin

 

 

 

 

 

ﬂ
00000000

0 . o . 0
6hr albumin adsorption introduction of ﬁbronectin

00000000 Fore—OW
O O

C

 

 

 

 

 

 

 

Figure 2.6: Cartoon of a model proposed for adsorption of PEN at an oil-water interface
at which HSA has been pre-adsorbed. Curves A, B and C represent models for proposed
interfacial organization, under conditions corresponding to curves 1, 2 and 3 in Figure 5.

39

 

 

E A AA
3. 0.8 E . AAA
.3 : . A A
5 06 i ' A A A
C ' r
L . .
0.2 b r l L 1 l r r r r l l r r L l

O
N
O
b
O
O)
C

Time (min)

Figure 2.7: Displacement of labeled HSA by unlabeled PEN at the oil—water interface. In
each experiment, unlabeled PEN (0.1 mg/ml) is introduced into the ﬂow cell after HSA-
FITC (0.130 mg/ml) adsorption for 1 hour (solid circles) and 3 hours (open triangles),
respectively (t=0 rrrinutes). In each set, the data has been normalized by the pseudo-
saturation intensity obtained prior to the buffer wash. Clearly, HSA is more effective at ‘
mitigating displacement by PEN after three hours at the interface than after one hour.

40

 

1.2 -
1 ‘ e
§ 1"....

3‘ 0-8‘ 9 buffer 'e00 e....
'5 8
51:3 0.6 ~ :
C
"" 0'4 “Hart-0.5mm: .

0.2 —

 

am

I I

1 I e i-FN-FITC(0.05mglml)

I I I I

0 30 60 90120150180210240270300330

Time(m'n)

Figure 2.8: Effect of pre-adsorbed HSA concentration on subsequent PEN adsorption.
Clearly, pre-adsorption of a higher bulk concentration of albumin (0.5 mg/ml) does not
signiﬁcantly inhibit subsequent PEN (0.05 mg/ml) adsorption at the interface.

41

 

.5
N

 

 

 

 

 

 

E
1 : . O
. O
i g 0
0.8 : Q ~
5‘ : Q AQiZSmg/mIPISA-FITC+
" 06 : 0.05mglmlI-FN
5 : . e 0.125 mg I ml HSA-FITC
: _
0.4 E .
0.2 E O A 4 4 A A A a a a
I A
o W19 l 1 l 1 1 l 1 1 l 1 r 1 1
0 20 4O 60 80

Time (minutes)

 

 

 

Figure 2.9: Competitive adsorption of HSA and PEN at the oil-water interface: This plot
shows preferential adsorption of PEN from a mixture of PEN (0.05 mg/ml) and HSA
(0.125 mg/ml) to a bare oil-water interface. The bottom curve (open triangles) is
adsorption of a mixture of unlabeled PEN and labeled HSA to a bare oil-water water
interface, and is shown for comparison.

42

14

12

 

HFN (0.1mg/ml)

 

 

 

 

 

 

 

10 — HSA(O.125mg/ml)
’5 . Mix
2 8
E. e
p.

4

2

to J

0 5 10 15 20
time (hr)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

— HSA (1 mg/ml)
HFN (0.1 mg/ml)
g 0 MIX
2
£3.
)-
4 4 A . —M M's—l—
r - v
o T I I I
O 5 10 15 20

time (hr)

Figure 2.10: Interfacial tension measurements of a mixture of PEN (0.1 mg/ml) and HSA
(0.125 mg/ml) adsorbing to an oil-water interface (middle curve) are shown in Figure
10a. For comparison, interfacial tension measurements of pure HSA (top curve) and pure
PEN (bottom curve) are also presented. In Figure 10b, the concentration of HSA is
increased to 1mg/ml. Over long times, the interfacial tension of the mixture approaches
that of pure ﬁbronectin.

43

3 ESTIMATING PROTEIN INTERFACIAL COVERAGE AT THE

LIQUID-LIQUID INTERFACE: A TIRF - FRAP STUDY

3.1 ABSTRACT

The adsorption of bovine serum albumin labeled with FITC (BSA-FITC) to a
model oil-water interface has been studied, using total internal reﬂection ﬂuorescence
microscopy (TIRFM). Estimates of the interfacial coverage of proteins at the oil-water
interface were obtained using a protocol based on TIRFM—FRAP, previously proposed
by Zimmerman and Gaub (1990). Protein coverages ranged from 0.02-0.3 mg/m2 for bulk
concentrations in the range of 0.2 mg/ml to 3.5 mg/ml. These values are an order of
magnitude lower than typical estimates reported in the literature for the solid-liquid
interface. It is speculated that the lower coverages at the oil-water interface may be due to
more rapid protein relaxations at this ﬂuid interface, which would tend to inhibit late
arriving protein molecules by occupying a greater interfacial area per molecule.
Measurements of ﬂuorescence lifetimes of BSA-FITC were also made to examine if there
were differences in the lifetimes between molecules at the interface and those in the bulk
using two-photon excitation spectroscopy. No differences were found between these two
lifetimes, suggesting that the quantum yield of BSA-FITC at the interface is the same as

its quantum yield in the bulk solution.

3.2 Introduction

Proteins are among the most important molecules involved in key biological
processes. Protein adsorption is one of the ﬁrst and most important steps that must occur
before key cellular processes such as cell attachment and differentiation can occur. They
function as enzymes, which are responsible for catalyzing several important reactions.
Proteins also play important roles in the storage and transport of biologically important

molecules such as oxygen“.

Protein adsorption to interfaces has been monitored using several methods.

Techniques such as radio-labeling'5’42, tensiometry“, optical reﬂectometry“, neutron

69,70 59,71

reﬂectometry , ellipsometry and 'I‘IRFM5’6'12 have been commonly used to
characterize interfacial adsorption behavior. More recently, methods such as quartz
crystal microbalance (QCM)72, surface plasmon resonance (SPR)6', optical wave guide
light mode spectroscopy (OWLS)l3 and AFM43‘73’74 are becoming increasingly popular. It
is important to be able to accurately estimate the quantity of adsorbed protein because
this allows us to make quantitative and qualitative predictions about the structure and

characteristics of the adsorbed protein ﬁlm, including predictions on whether the

adsorption is limited to a single monolayer.

Techniques such as reﬂectometry and ellipsometry rely on the evolving refractive
index of the protein ﬁlm to estimate the surface coverage. Research groups that use
TIRFM for observation of protein adsorption conduct their experiments in regimes where
adsorption is strictly transport-limited, so that surface coverages are calculated by direct

application of the Leveque solutionlz’w. This results in conversion of ﬂuorescence

45

intensities to surface coverages by application of a simple conversion factor. Other
TIRFM-based calibrations use angle scanning methods where the penetration depth of the
evanescent wave is varied”, or use external and internal standards to calibrate the

114

ﬂuorescence signa . Adsorbed amounts can also be computed by a visual count of

individual protein molecules, from AFM images of adsorbed protein layers”.

In contrast with the considerable literature on protein adsorption at the solid-
liquid interface, there are relatively few reports on estimating interfacial coverages at the
liquid-liquid interface. Estimates of surfactant and protein interfacial coverages at the
liquid-liquid interface have been obtained by radiolabeling“15 and neutron reﬂectivity4
Because of the instability of the interface boundary, estimating coverages at liquid—liquid
interfaces is a challenging problem. For example, quantitative interfacial concentrations
of proteins at the interface can be easily obtained by scraping the adsorbed protein off the
surface. For obvious reasons, this cannot be done at liquid-liquid interfaces. Also, the

curvature of the interface presents additional problems for optical experiments.

In the experiments on sequential and competitive adsorption between human
plasma ﬁbronectin and human serum albumin at the oil-water interface, it was observed
that PEN preferentially adsorbs over HSA. It was also observed that PEN displaces pre-
adsorbed albumin over nearly all adsorption time scales. However ﬂuorescence studies at
the oil-water interface are subject to greater amounts of “noise” induced by light
scattering than at the solid-liquid interface, due to the uneven interface. Such scatter can
also occur due to contaminants such as dust particles which can stick to the oil, even after
taking the utmost experimental precautions. Electromagnetic energy from the increased

light scattering can further excite both bulk and surface associated molecules, inducing a

46

level of ﬂuorescence emission beyond what is induced by the direct laser beam. This can
result in artiﬁcial elevations in the level of ﬂuorescence emission that is collected by the
objective, which complicates direct comparison of ﬂuorescence emission data between
different experiments. Thus, a calibration technique that allows us to directly convert
ﬂuorescence emission intensities into interfacial mass coverages will eliminate such
artifacts. This will also allow us to make more rigorous comparisons between data sets

obtained under different experimental conditions.

Gajraj55 in our laboratory had previously adapted a technique for estimation of
coverages at the liquid-liquid interface, using the technique ﬁrst proposed by Zimmerman
et a]16 and is based on a combination of total internal reﬂection ﬂuorescence microscopy
(TIRFM) and ﬂuorescence recovery after'photobleaching (FRAP). The technique relies
on the ability to discriminate between ﬂuorescence recovery from bulk species and that
from surface-associated molecules. Gajraj reported interfacial coverages that were greater

than an order of magnitude lower than similar values reported in the literature.

In the present study, a calibration protocol has been developed to obtain an
adsorption isotherm for BSA at the oil-water interface. One of the important assumptions
of this intemal-standard technique was that the quantum yield of a ﬂuorophore in the bulk
solution remains unchanged upon its adsorption at the interface. The validity of this
assumption was examined by conducting two—photon excitation spectroscopy studies of
ﬂuorescence lifetimes of BSA-FITC adsorbed at the oil-water interface and in the free
solution. The inﬂuence of shear rate (ﬂow rates) on the ultimate interfacial coverage of

BSA at the liquid-liquid interface was also assessed, along with shear rate studies of

47

protein adsorption at a hydrophobic solid-liquid interface (Octadecyltrichlorosilane

(OTS) self assembled monolayers on glass) and compared.
3.3 Theory

3.3.1 Total internal reﬂection

A detailed derivation of Maxwell’s equations governing the theory of total
internal reﬂection is available elsewhere (see Reichert et al.76, for example). In brief,
when a light beam traveling through an optically dense medium of refractive index n;
approaches the interface between the dense medium and a rarer medium or refractive
index n2 (n2<n 1) at an angle of incidence exceeding the critical angle Be, it undergoes total
internal reﬂection. The critical angle at which this occurs is given by Equation (2.1).
When the beam undergoes total internal reﬂection, the reﬂected ray travels through the
same medium as the incident ray. However, an electromagnetic ﬁeld is generated in the
rarer medium beyond the interface. The intensity of this ﬁeld (called the evanescent
wave) decays exponentially with perpendicular distance 2 from the interface as given by
Equation (2.2). The characteristic penetration depth of this wave, d , is given by
Equation (2.3) and indicates the distance at which its intensity has reduced to 1/e of its
maximum value at the interface. Typical values for (1,, vary from 50 to a few hundred
nanometers. Thus, only ﬂuorophores located in the vicinity of the interface are excited by
this energy ﬁeld. This interfacial selectivity makes this optical technique ideal for

studying surface and interfacial phenomena.

48

3.3.2 The calibration protocol

The theory for estimating protein interfacial coverages using TIRFM and FRAP was
developed by Zimmerman et aim. A complete discussion of this technique can be found
in Gajraj55 and the key steps towards developing this theory has been outlined below for

completeness .The ﬂuorescence emitted by an excited ﬂuorophore is given by
F0) = j q.I(z>q(z>C<z)dz (3.1)
0

Here, [(2) is the intensity of light striking ﬂuorophores of concentration C(z)
located at a distance z from the interface, q(z) is the quantum yield of the ﬂuorophore at
z, and q, is an instrument constant. The total ﬂuorescence collected by the objective is a
function of contributions from surface-bound proteins and ﬂuorescently labeled proteins

present in the bulk that are illuminated by the tail of the evanescent wave:
F(t) = Iartiﬁce (t) + Fbulk (t) (3.2)
Incorporating the expression for the decaying evanescent wave given by Equation

(2.2) gives

1
F(t) =q,IOJ'exp _c-iz_ (z)q(z)dz (3.3)
0

P

where I is the depth of the ﬂow cell. This integral can be deconvoluted into bulk and

surface contributions and represented as

49

I
Z
17(1):}: +Fbulk =q1qsIoCs+qlqubIOIexp _ d
S

 

Z (3.4)

surface
I)
where q, and qb are the quantum yields of ﬂuorophores at the interface and in the bulk,

respectively.

In developing Equation (3.4), Zimmerman et a1.16 assumed a model where a layer
of adsorbed proteins of thickness 3 and two-dimensional concentration C, is in
equilibrium with ﬂuorescently labeled proteins in the bulk solution at a three-dimensional
concentration Cb. This equation also assumes that that plane of total internal reﬂection
occurs at the interface between the protein ﬁlm of thickness 3 and the bulk solution. This
interface can facilitate total internal reﬂection as the refractive index of the protein ﬁlm is
assumed to be close to that of the oil (n=l.51). It is important to note that the presence of
an intermediate ﬁlm of arbitrary refractive index and nanometer scale thickness between
the dense medium and the rare medium will not prevent total internal reﬂection from
occuring. For proteins such as serum alburrrin, a typical value of ﬁlm thickness estimated
using ellipsometry is approximately 5 nml6'55. This value of sis considerably smaller than
the thickness of the ﬂow cell (D. Therefore, the deﬁnite integral in Equation (3.4) can be
solved after noting that s<<l. Thus the term corresponding to FM). can be rewriten as

follows:

 

Z—S z-s l
’ iii] it]
1:11qu = qIIoqubIe dz =41qube (—dp) e

s (3.5)

: ql qubedp

50

Combining this with Equation (3.7) leads to the following relationship for the

total ﬂuorescence emission from the sample:
F(t) : quI (qus + qubdp) (3'6)

The assumption that the quantum yields of surface and bulk associated
ﬂuorophores are equal (qo) (an assumption that would be discussed in detail later) leads

IO,
F(t) = q,quo (Cs + dep) (3.7)
The bulk contribution to the total intensity is simply
Fbulk (t) = qlquOdep (33)

Thus the unknown constants can be factored out and Equations (3.7) and (3.8) can

be combined and rearranged to give

Csszdp[ PO) 4] (3,9)

 

Ft... (1)

The variables in the above equation can be easily estimated as shown in Figure
3.1. F(t) represents the total ﬂuorescence measured prior to initiation of the bleaching
pulse, and is obtained at the end of a typical adsorption experiment at the oil-water
interface. Fbuuxt) is the ﬂuorescence after a series of bleach pulses are applied. Both

values are corrected for background ﬂuorescence before incorporation into Equation (3.9)

51

This technique relies on the ability of the evanescent wave to selectively and
primarily excite surface ﬂuorophores, along with only a small number of bulk associated
ﬂuorophores. It also relies on the fact that the diffusion coefﬁcient of proteins in the bulk
is signiﬁcantly higher than that of proteins bound at the interface. As a result, surface
bound ﬂuorophores can be considered irreversibly bleached within the time scale of
observation, while bleached bulk ﬂuorophores would be instantaneously replaced by
unbleached ﬂuorophores, thus resulting in an apparently unbleachable bulk fraction. The
diffusion coefﬁcient of proteins such as albumin and ﬁbronectin is in the range of 10'7
cmzls. For a freely diffusing molecule undergoing Brownian motion, the time required to

travel through a bleached region of depth dp, based on the Stokes-Einstein equation is

t =—"— (3.10)

For a penetration depth of approximately 80 run, this results in a characteristic
time of diffusion on the order of approximately 1 to 10 milliseconds. As a result,
ﬂuorophores in the bulk that are bleached by the evanescent wave are quickly replaced by
unbleached ﬂuorophores that essentially instantaneously diffuse into the bleached zone.
Hence, bulk species can be treated as an unbleachable fraction that maintains a constant
contribution to the total ﬂuorescence emitted. A cartoon showing the sequence of
photobleaching and bulk recovery is shown in Figure 3.2. This cartoon shows a region
near the interface that is ﬁlled with ﬂuorophores that are adsorbing to the interface. The
open circles represent unbleached ﬂuorophores while the closed circles represent

bleached ﬂuorophores.

52

Prior to application of the bleach pulse, all the ﬂuorophores are unbleached and
contribute to the total emission F(t). Immediately after the application of the bleach
pulse, all the ﬂuorophores in the illuminated region would be bleached. However, after a
few milliseconds have elapsed, the bulk ﬂuorescence would recover due to the rapid
replacement of bleached ﬂuorophores in the bulk with unbleached ﬂuorophores by
diffusion as predicted by Equation (3.10). Since the immobilized ﬂuorophores at the
interface would remain bleached, the ﬂuorescence emission at this point is represented by
Fbout). This procedure is an elegant method of separating bulk and surface contributions
to the total ﬂuorescence emission intensity. In the current experimental conﬁguration, the
typical dwell time between measurements is 100 milliseconds, thus providing sufﬁcient

time for bleached bulk molecules to be replaced by unbleached ones.

The bulk and surface ﬂuorescence contributions can also be separated by rinSing
the ﬂow-cell with a protein-free buffer, which would presumably deplete the bulk of
labeled protein species, while leaving interface-associated species in place. However

there are several disadvantages to using this method at the oil-water interface.

(1) Since the interface is mobile, it is necessary to conduct such a wash at very low
ﬂow rates to prevent inducing interfacial destabilization. As a result, ensuring that all the
labeled protein present in the bulk solution has been eliminated requires a thorough wash-
out that would take approximately three hours (this corresponds to the time required to
exchange about 10 volumes of the ﬂow cell). In such an extended wash-out period, there
is no reliable method of preventing adsorbed species and the oil ﬁlm from being

entrained and removed from the interface.

53

(ii) The quantum yield of ﬂuorophores is dependent on the local environment.
Thus, eliminating labeled proteins from the bulk would change the local environment
around the surface-bound ﬂuorophores. This may result in changes in their quantum

yield, thereby introducing artifacts in the data.

Equation (3.9) was developed on the underlying assumption that the quantum
yields of the surface and bulk associated ﬂuorophores are equal. However, as noted
above, the quantum yield of ﬂuorophores is a very sensitive function of the environment
surrounding it. Thus the proximity of a tagged protein to other ﬂuorophores (which
results in quenching) or its location in a highly viscous or hydrophobic environment can
affect the quantum yield. Hence, if q5< qb, this can result in signiﬁcant underprediction of
interfacial coverages by Equationo(3.9). It is, therefore, necessary to test the validity of
this assumption at the oil-water interface. While the estimation of absolute quantum
yields is a challenging problem, it can be related to the ﬂuorescence lifetime I; a variable
that describes the length of time a molecule resides in the excited state before eventually
transitioning to the ground state. The ﬂuorescence lifetime is more readily measurable,
and comparison of the lifetimes of molecules at the surface and in the bulk can provide us
with insight into the relationship between q, and qb. This approach is discussed in greater

detail below.

3.3.3 Two-photon ﬂuorescence lifetime spectroscopy and the quantum yield

problem

3.3.3.1 Rationale

In the development of the calibration protocol outlined in the preceding section, it
was assumed that the quantum yield of surface-associated and bulk-associated

54

ﬂuorophores are identical. Although this assumption has been made to simplify the
mathematical analysis, it is a well known fact that the quantum yield of a ﬂuorophore is
highly sensitive to the surrounding environment. As noted above, direct measurement of
ﬂuorescence quantum yields is difﬁcult. In fact, the estimation of the relative quantum
yield of ﬂuorophores to within 5% precision is extremely difﬁcult and the values against
which the relative yields are standardized are themselves subject to greater than 5%
variation77’78. Estimation of the absolute quantum yield is considerably more difﬁcult and
requires sophisticated instrumentation. The two most common methods used for
measuring quantum yields are calorimetry79 and by comparison with known standards.
Therefore, estimating the quantum yield of a ﬂuorophore adsorbed to (or in the proximity

of) an interface is obviously very difﬁcult.

The quantum yield of a ﬂuorophore is deﬁned as the ratio of the number of

photons emitted to the number of photons absorbed, and is given by80

 

q= ’ (3.11)

where k, and kn, are the rates of radiative and non-radiative decay, respectively. The

lifetime rof the ﬂuorophore is given by:

1

r=-———— .
k +k (312)
r nr

The variables k, and kn, can be calculated as

55

_ q
k, -; (3.13)

1—
km = [—2] (3.14)

Thus ﬂuorophores with a large radiative decay rates have high quantum yields
and short lifetimes. In fact, the radiative decay rate k, is a'function of the oscillator
strength of the electronic transition“, which is only slightly sensitive to the environment
around the ﬂuorophore. Thus k, can be treated as a constant for a given ﬂuorophore.
Variations in ﬂuorescence lifetimes and quantum yields of ﬂuorophores resulting from
changes in non-radiative decay rates are due to events such as collisional quenching and
ﬂuorescence resonance energy transfer”. Thus, in principle, the quantum yield and the
ﬂuorescence lifetime are directly proportional to each other, assuming that there are no
changes in the radiative decay rate. Then, the lifetime and quantum yield of a ﬂuorophore
will increase or decrease together”. Thus, the quantum yield of a given ﬂuorophore can

be indirectly probed by measuring its associated lifetimes.

It is relatively easy to obtain estimates of the ﬂuorescence lifetimes of
ﬂuorophores. The lifetime of the ﬂuorophore is the length of time a molecule spends in
the excited state, prior to its eventual transition to the ground state. It is related to the
quantum yield (please see Equation (3.13)). To estimate the quantum yield, the time
interval between the application of the input pulse and the arrival of the ﬁrst emitted

photon was measured, using time correlated single photon counting (TCSPC). Since

56

ﬂuorescence emission is a random event, the distribution of these arrival times represents

the ﬂuorescence decay for a large number of photons.

Estimating the lifetime (2') for BSA-FITC molecules in the bulk solution using
conventional TCPSPC spectroscopy is relatively easy. However, estimating the
parameter for BSA-FITC molecules adsorbed at the oil-water interface is a much more
challenging problem. In addition to the constraints imposed by the geometry of the
experimental set-up and detection equipment, there is the serious challenge of isolating
the interfacial signal from the background bulk ﬂuorescence signal. Since the beam
excites a large region around the interface, a signiﬁcant portion of the emission signal can
. be considered to originate from bulk contributions. This results in “noise” that is difﬁcult
to deconvolute. Finally, the oil-water interface is not a “smooth” interface when
compared to the solid-liquid interface. There are “ridges and bumps” at the boundary of
the two phases due to the juxtaposition of one phase into another at various points along
the interface. In addition, the relative afﬁnity of the aqueous phase for quartz causes
water to creep around the oil phase at the cuvette walls, producing a curved interface. All
these factors combine to produce a more complicated optical path. Thus, using the
conventional set-up for exciting the ﬂuorophores does not produce reliable estimates of
the dynamics at the interface. Therefore, to obtain more reliable estimates of the
ﬂuorescence lifetimes of the adsorbed protein molecules, two-photon time correlated

single photon counting spectroscopy was used.

3.33.2 Theory

Two-photon ﬂuorescence occurs when a ﬂuorophore simultaneously absorbs two

photons of wavelength 271 and then undergoes an electronic transition at a wavelength

57

3.83. The number of photons absorbed in two-photon spectroscopy is directly proportional
to the square of the incident light intensity. The ﬂuorescence emission from a sample

undergoing a two-photon transition can be described as,
F = alleU, z)12(r, z)dV (3.15)

where C(r,z) is the concentration proﬁle of the ﬂuorophore undergoing the
electronic transition, I(r,z) is the intensity of the incident light, and the integration is

computed over a volume in space.

The non-linear dependence of the absorption on the square of the incident light
intensity is responsible for the spatial selectivity of this technique. Since the amount of
light absorbed is proportional to the square of the number of incident photons, very high
intensities are required in order to obtain a measurable amount of 2-photon excitation“.
This excitation occurs at the focal point of a laser beam that has been focused through a
converging lens (or an objective). As a result, ﬂuorescence is errritted only from regions
illuminated by the focal volume occupied by the beam at its waist, thereby affording
excellent spatial selectivity. An additional advantage of 2-photon spectroscopy is that the
excitation and emission wavelengths are widely separated, so there is little possibility of
data contamination due to spectral overlap. For example, for experiments discussed in

this chapter, the excitation wavelength was 700 run while the errrission wavelength was

set near the emission maximum of FITC at about 530 nm.

Two-photon excitation has primarily been used to study the excited states of

molecules, because it follows different selection rules for electronic transitions as

58

compared to single photon absorption processess4’85. However, in the last decade, it has
been used extensively for studying biological samples because of its excellent spatial
selectivity and low background ﬂuorescence86'89. In most cases, the emission spectra for
l-photon and 2-photon excitation are the same. This is an important point because, in this .
study, lifetimes were measured using 2-photon spectroscopy. These results are then
extrapolated to speculate on possible changesin quantum yield that occur during the
TIRFM experiments, which use single-photon excitation. All TIRFM studies in this
dissertation were conducted with excitation at 488nm, whereas the two-photon studies
were conducted at an excitation of 700 nm. It has been previously demonstrated that
emission spectra for most dyes such as ﬂuorescein or Cascade Blue is independent of the
excitation wavelength, which would indicate that the ﬂuorescence quantum efﬁciency is

independent of the excitation wavelength”.

3.3.3.3 Measurement of Fluorescent Lifetimes

Fluorescence emission curves can typically be described by single or multi-

exponential decay. For a ﬂuorophore undergoing exponential decay”,
_ -t / r
F(t) — Foe (3.16)

where F0 is the ﬂuorescence intensity at t=0, and ris the lifetime. For decays that
are not adequately described by a single exponential, the decay proﬁles are typically

ﬁtted to a sum of exponentials,

—/ .
F(t)=Za'ie t T‘ (3.17)

59

where a.- is the pre-exponential factor representing the fractional contribution to

decay of the ﬂuorophore with a lifetime 2;.

Decay profiles that are best ﬁt to multi-exponential models can suggest the
presence of multiple ﬂuorophore populations, or the presence of a single ﬂuorophore in
different environments. For example, the presence of two lifetime components for BSA-
FITC can be attributed to FIT C being present in two different kinds of regions on the
protein. Multi-exponential ﬁts can also be merely statistical in nature and may not imply
any underlying physical phenomenon. These decay proﬁles are usually analyzed using
the method of least squares which minimizes the sum of the weighted squares of residuals

deﬁned as

1 n
12 =;‘:;§W. ly. -f(t.- tape->12 (3.18)

where n is the total number of data points, p is the number of ﬁttable parameters

and the factor wiis the weight associated with each square residual.

This weight is related to the standard deviation associated with each count
interval. For photon counting applications, it is reasonable to assume that the error
follows Poisson counting statistics”. Thus the variance at each interval is directly

proportional to the number of counts at that interval.
0'. = yi (3.19)

Thus, the decay curves can be weighted as,

60

w. z _ =_1_ (3.20)

3.4 Experimental Methods

3.4.1 Materials and Methods

BSA was purchased from Sigma Aldrich (St. Louis, MO). Fluorescein
isothiocyanate (F-1907) was obtained from Molecular Probes (Eugene, OR). The oil used
for the experiments was an immersion oil (Type A) obtained from Cargille (Cedar Grove
NJ). This oil had a refractive index of approximately 1.515. Octadecyltrichlorosilane

(OTS) was obtained from Sigma Aldrich (St. Louis, MO)

3.4.2 Protein labeling

BSA was labeled with FIT C in 0.1M carbonate buffer at pH 9.2 for 4-6 hours.
The solution was then passed through a PD-10 column (Amersham Biosciences,
Piscataway, NJ) that had been pre-equilibrated with 0.05M phosphate buffer (pH 7.4).
The column was then eluted with 0.05M phosphate buffer to recover the protein
conjugate. This was followed by overnight dialysis against phosphate buffer to
completely remove unbound labels from the solution and to ensure complete buffer
exchange. Labeling ratios were determined by measuring the absorbance of the conjugate
at 496nm and 280nm. In all experiments, the labeling ratio was kept below 1.3. All
protein solutions were stored at 4°C and used within a week. All adsorption experiments

were conducted in 0.05M phosphate buffer.

61

3.4.3 Preparation of surfaces

The microscope slides (Fisher Premium slides) used for TIRFM experiments were
purchased from Fisher Scientiﬁc (Pittsburgh, PA). Slides on which oil ﬁlms were
deposited were cleaned by ultrasonication in RBS-35 detergent solution. The slides were
rinsed thoroughly in deionized water and then immersed in concentrated nitric acid for 30
minutes. The slides were stored in deionized water to prevent dust accumulation on the
surface. Microscope slides on which OTS monolayers were deposited were cleaned by
ultrasonication in RBS-35 detergent solution. After rinsing the slides thoroughly with
deionized water, they were immersed for 30 minutes in piranha solution (70% sulfuric
acid and 30% hydrogen peroxide)3. The slides were subsequently rinsed with deionized
water and dried thoroughly under nitrogen. They were then immersed in 1% (vol/vol)
solution of OTS in hexadecane for approximately 15 minutes, after which they were
thoroughly rinsed with methylene chloride and chloroform”. The contact angle of the
self assembled OTS monolayers was measured using a home-built contact angle

instrument and was found to be greater than 110 degrees in all cases‘.

3.4.4 TIRFM instrumentation and experimental set-up:

The TIRFM instrumentation, ﬂow cell design and optical conﬁgurations for
FRAP were described in detail in Chapter 2. The oil layer was deposited on a clean glass

slide by rolling a clean glass rod through a drop of oil placed at one end of the glass slide.

 

3 Piranha solution is extremely corrosive. Face mask, acid-resistant gloves and protective clothing must be
used while handling piranha. Use extreme caution.

‘ Static contact angle measurements were made using a homemade contact angle instrument in Dr. Michael
Mackay’s laboratory at Michigan State University.

62

For adsorption and FRAP experiments, monitoring beam intensities ranged from 1-4 11W
and photobleaching beam intensities were at least lOmW or higher. The duration of each
photobleaching pulse varied from 350ms to 1 3. Typically, bleaching was carried out for
a total duration of 5 to 10 seconds till no further recovery was observed from the sample.
It was found that longer'bleaching times are required to completely bleach ﬂuorophores
at the oil-water interface as compared to the OTS-water interface. This may be due to a
higher fraction of reversibly associated protein at the oil-water interface than at the OTS-
water interface. In some experiments, the observation area was reduced by placing an
aperture before the photomultiplier tube. This was done to ensure that ﬂuorescence
emission was collected from only a small portion in the center of the bleached region to
counter the effects of lateral diffusion. No signiﬁcant difference in computed interfacial

coverage was observed in the presence of the aperture.

Experiments were initiated by introducing the buffer solution into the ﬂow cell
using an infusion/withdrawal syringe pump. After beam alignments and ﬁne-focusing
were completed, BSA-FITC was introduced into the ﬂow cell at a controlled ﬂow rate
and adsorption was allowed to proceed until the ﬂuorescence emission intensity reached a
stable value. A series of rapid bleach pulses were then applied to completely bleach the
interfacial ﬂuorophores. The residual ﬂuorescence emission corresponded to the
contribution of bulk-illuminated ﬂuorophores to the pre-bleach emission intensity.
Typically, three to four different sites on each interface were bleached to determine the
degree of variation in surface coverage across the interface, and to ensure that the data are

statistically relevant.

63

3.4.5 Two photon lifetime spectroscopy: Experimental set-up.

Estimates of the ﬂuorescence lifetime of FITC and BSA-FITC were obtained
using a two-photon time correlated single photon counting spectrometer (2P-TCSPC).
The experimental layout of the 2P-TCSPC spectrometer is shown in Figure 3.3. A
description of the experimental set-up can be found in DeWitt et al.68, although the
system has been modiﬁed slightly from its original implementation. Brieﬂy, the output
from a 30 W CW mode-locked Nd:YAG laser (Coherent Antares 76 S) at 1064 nm was
frequency doubled using a temperature tuned Type I LBO crystal to produce a 532 nm
beam. This light was used to excite a cavity dumped dye laser (Coherent 702-2). For
these experiments, pulses at 700nm were generated using LDS-698 laser dye (Exciton).
In all experiments, the ﬂuorescence light is collected at 90° with respect to the incident
light. The emitted light was polarization-selected using a Glan-Taylor prism. The
polarization of this light was then scrambled and wavelength selection is carried out
using a subtractive double monochromator (CVI Digikrom 112). The collected light was
focused on to a cooled two-stage nricrochannel plate photomultiplier tube (MCP-PMT,
Hamamatsu R3809U). The PMT signal was sent to one channel of the quad constant
fraction discriminator (CFD Tennelec TC454), where it is processed for input to the time-
to—amplitude converter (TAC, Tennelec TC 864). This system can measure lifetimes from
several microseconds to the instrument response function of 35 picoseconds FWHM. The
experiments were conducted in quartz cuvettes, and the oil-water interface was assembled
by depositing an oil-layer over the protein solution. After adsorption had been initiated,

the laser beam was carefully focused at the oil-water interface and the collection lens was

adjusted to maximize the collected signal. The excitation beam was at 700nm, while

emission was monitored near the emission maximum for FITC (530nm).

3.4.5.1 Data analyses for TIRFM and calibration protocol

Interfacial coverages were computed using Equation (3.9). The prebleach
ﬂuorescence was estimated by averaging a few data points before the bleaching pulse.
Typically, the interface was bleached with a series of pulses for approximately 5-10
seconds, in order to eliminate ﬂuorescence emission from interfacially-bound molecules.
An arithmetic average of the ﬁrst few points after the completion of bleaching was used
to calculate Fbulk- Typically, the ﬁrst point right after the bleach pulse is applied is
ignored because this may be inﬂuenced by a partial opening of the shutter while the
photon counter is sampling data. Each data point consisted of the photon count rate
collected for 50 milliseconds. A dwell time of 200 milliseconds was used between each
reading. Since the repeated photobleaching of interfacial species may also induce some
photobleaching of bulk associated ﬂuorophores outside the evanescent wave area, the
longer dwell times provide sufﬁcient time for complete recovery of the bulk ﬂuorescence
emission. The background counts were subtracted from both F and F buy; prior to applying
Equation (3.9). This subtraction is important, because the background contribution cannot

be ignored due to a slight ﬂuorescence emission from the oil.

3.4.5.2 2-pTCSPC lifetime measurements

Fluorescence lifetimes were estimated by ﬁtting the decay to Equations (3.16) and
(3.17), using OriginPro 7.5 (OriginLab Corporation, Northampton, MA). The data used

for ﬁtting were not normalized against the maximum value (at t=0). Also, no constraints

65

(such as 2a,. =1 ) were imposed on the ﬁt. Although normalization and constraints are

.
routinely used in ﬁtting lifetime data, we chose not to implement this because relatively
low ﬂuorescence emission intensities (counts) were obtained using 2p-TCSPC for the
protein sample. To avoid making a biased estimate of the maximum count at t=0 (there is
considerable scatter around this value), the program was allowed to ﬁt all parameters

with no imposed constraints.
3.5 Results and Discussion

3.5.1 Adsorption isotherm for BSA-FITC at the oil-water interface

The adsorption isotherm for BSA-FITC at the oil-water interface is shown in
Figure 3.4. Experiments were conducted for BSA-FITC bulk concentrations of 0.2, 0.5,
1.0, 2.0 and 3.5 mg/ml.' Interfacial coverages obtained were in the range of 0.02 to 0.3
mg/mz. The values obtained are much improved in comparison to the results of earlier
work done in our laboratory55 primarily due to (i) improved detection equipment, and (ii)
the modiﬁed bleaching protocol, where longer bleaches were utilized to ensure that all
interfacial ﬂuorophores were adequately bleached. However, the values obtained are an
order of magnitude lower than typical values that have been reported in the literature for

the solid-liquid interface.

There are very few reports in the literature on the adsorption of BSA at the liquid-
liquid interface to provide comparative data. Beaglehole et a1.91 studied BSA adsorption
to the oleyl alcohol-water interface using ellipsometry. They reported interfacial
coverages of 0.45 mg/m2 for BSA at a bulk concentration of 3.5 mg/ml, with no apparent

signs of saturation. In contrast, Sengupta and Damodaran reported much higher

66

l .

equilibrium saturation values of 4.1 mg/m2 for BSA at the triolein-water interface using

radiolabeling”. This value appears abnormally high.

Estimates of adsorbed amounts of BSA at the solid-liquid interface vary,
depending on surface properties and the technique used for making the measurements.
Hlady and co-workers reported BSA interfacial coverages ranging from 0.4 mg/m2 to 1.2
mg/m2 on hydrophilic silica using TIRFM”. Malmsten and Lassen reported HSA
coverages of 412 mg/m2 on hydrophobic surfaces using ellipsometry”. Calonder and co-
workers13 reported interfacial coverages of 1.25 mg/m2 for HSA on Si-Ti-Ox surfaces
using OWLS. Wertz and Santore used TIRFM to study BSA adsorption to HTS-SAMs
and obtained interfacial coverages in the range of 1 to 3 mg/m2 by modeling the initial
adsorption rate under transport-limited conditions”. Finally, Choi et al.‘59 reported
interfacial coverages ranging from 0.4 to 1 mg/m2 for HSA adsorption to HTS SAMs
depending on the transport conditions, and the density of the SAMs, using optical

reﬂectometry.

We believe that some of the lower interfacial coverages obtained at the oil-water
interface can be attributed to competition between slow arriving molecules and quick
relaxation of already adsorbed molecules (because relaxed molecules occupy more area
per molecule), and the slow transport of new albumin molecules to the interface.
Evaluation of this hypothesis and supporting experiments/arguments are summarized in a

later section.

67

3.5.2 Molecular relaxations and its inﬂuence on ultimate coverages at the oil-water

interface

The interfacial coverage reported above at the oil-water interface is an order of
magnitude lower than typical values reported in the literature, which are on the order of 1
mg/mz, and points to the fact that the total adsorption at this oil-water interface is limited
by other processes. A plausible hypothesis is that the reduced interfacial estimates reﬂect
the competition for interfacial space between adsorbed molecules quickly undergoing
interfacial relaxations and protein molecules arriving at the interface. As discussed
earlier, proteins are extremely complex entities that contain several hydrophobic and
hydrophilic regions. All proteins exist in energetically favorable conformations that
minimize the free energy of the system. Upon. adsorption, they undergo conformational
changes (or re-orientations) to minimize the free energy inthe changed environment
around them. This can amount to relaxation, by which the molecule undergoes
denaturation and typically expands, thus occupying a larger surface area in the process.
The rate of this process is signiﬁcantly inﬂuenced by the properties of the

surface/interface to which the protein is adsorbing, and the conditions of the experiment.

In these experiments, the ﬂow rate is held constant at 0.047 ml/min. This low
ﬂow rate was chosen to minimize the possibility of generating destabilizing forces that

can sweep away the oil-ﬁlm. The wall shear rate is given by

 

(3.21)

where Q is the ﬂow rate and w and h are the width and height of the ﬂow channel.

68

For the geometry of the ﬂow cell, yhas an approximate value of 0.27 s". Thus,
transport of proteins to the interface occurs at a very slow rate under these experimental
conditions. Figure 3.5 depicts a plausible model for the reduced interfacial coverages.
This model is similar to that proposed in Chapter 2, at short adsorption time scales.
I Initially, the albumin molecules that arrive ﬁrst at the oil-water interface adsorb and
occupy a small area. However, with time, protein unfolding and relaxations cause the
adsorbed molecules to occupy a greater surface area per molecule, thereby limiting the
space available for adsorption of later arriving molecules. It is conceivable that if this rate
of relaxation is high enough in comparison to the rate of transport of protein to the

interface, it can result in reduced interfacial coverages.

For example, Wertz and Santore have proposed that albumin can increase its
molecular area by a factor of at least ﬁve as it undergoes relaxations”. Thus the protein
molecule expands from an initial molecular area of 25 nmzlmolecule to approximately
140 nmzlmolecule during the relaxation process. In the same paper, the authors have
shown that albumin interfacial coverages are inﬂuenced by protein concentration and
wall shear rate. They further reported that the ultimate coverage is more strongly

inﬂuenced by molecular relaxations than by a fundamental isotherm shape”.

The proﬁles in Figure 3.6 depict BSA-FITC adsorption (0.2 mg/ml) to the oil-
water interface at three different ﬂow rates: (i) 0.024 ml/min (y=0.13 3"), (ii) 0.048
mllmin (F027 s"), and (iii) 0.096 ml/min (7:0.54 s"). The proﬁles indicate a signiﬁcant
increase in albumin interfacial coverage at 0.048 ml/min and 0.096 mllmin, in

comparison to the coverage at a ﬂow rate of 0.024 ml/min. An interfacial coverage of

69

approximately 0.040 mg/m2 was obtained at ﬂow rates of 0.048 ml/min and 0.096

ml/min. On the other hand, the surface coverage at 0.024 ml/min was 0.016 mg/mz. There

are two possible reasons for the lower coverage at 0.024 ml/minlz.

(i)

(ii)

If the adsorption follows the convective-diffusion model for gentle shearing
ﬂow, then the initial adsorption rate is proportional to the third power of the
shear rate (1)/3). Thus a reduction in the shear rates (or ﬂow rate) will reduce
the interfacial coverage. However, it should be noted that, at the above ﬂow
rate, the initial adsorption rates probably do not strictly follow the transport-
lirnited adsorption kinetics predicted by the Leveque equation, because of the

higher bulk protein concentrations (0.2 mg/ml)”.

With reference to the model proposed in Figure 3.5 along with the arguments
made in the preceding paragraph, we hypothesize that interfacial relaxation of
the adsorbed albumin molecules are primarily responsible for the apparently
low interfacial coverage obtained at the lowest ﬂow rate. Due to slowly
arriving molecules at this ﬂow rate, the adsorption proceeds very slowly and
continues for almost 90 minutes after the ﬁrst increase in ﬂuorescence is
detected. Thus, there is sufﬁcient time for molecules arriving earlier to undergo
the molecular relaxations that reduce the interfacial area available for the

adsorption of later arriving molecules.

The difference between the interfacial coverage for adsorption at the two ﬂow

rates of 0.048 mllmin and 0.096 ml/rrrin are not signiﬁcant. While this is initially

surprising, it should be noted that the oil ﬁlm in contact with the glass is likely to become

70

r.“ ”ﬁr

unstable at a ﬂow rate of 0.096 mllmin, and hydrodynamic shear may be sweeping some

of the oil and adsorbed proteins from the interface.

It is worthwhile to conduct a modeling exercise to compare the coverage obtained
above with that predicted by models for transport-limited adsorption from a solution in
fully developed laminar slit ﬂow“. The Leveque solution to the convective diffusion

equation is given by

1/3
15 = 0.538(1) DmC (3.22)
dt )1:
where F is the two-dimensional interfacial concentration and x is the distance along the

ﬂow channel.

The Leveque equation only applies when a steady-state concentration near the
interface is established at times well before the surface coverage reaches the point where
it starts decreasing the adsorption rate. Thus, this equation applies when the adsorption

rate does not decrease before a characteristic time given by

x2 1/3
t = 3.23
. [7,0] ( 1

 

where It is the time required to establish the steady-state concentration profile.

This corresponds to 940 seconds for F013 s", 580 seconds for F027 s“, and

360 seconds for 7:0.54 s'l. Assuming no errors in the experimental model, if the

experimental adsorption rate is lower than that predicted by Equation (3.22), then the

71

i:

experimental adsorption rate is considered to be limited by some property of the
interface69'92. The proﬁles in Figure 3.6 were analyzed to compare the initial adsorption
rates obtained using this model with those predicted by Equation (3.22). Table 3.1 shows
these comparisons. The fact that the experimental adsorption rate is orders of magnitude
lower than that predicted by the Leveque solution is indicative of surface effects that
lower the initial adsorption rate. The parameter R, which is the ratio of the initial
adsorption rate for any shear rate to that obtained at the shear rate of 0.27 s'1 (which we
arbitrarily chose as a standard), was also calculated. If it is assumed that there is a
calibration error in our protocol that causes underestimation of the surface coverage (due
to light scattering artifacts) by a constant factor, the ratio R will account for this. When
comparing experiments conducted at r-=0.13 s'1 with those at 7:027 s’], we obtained a
value of R12: 0.17, which is much lower than the corresponding theoretical value of 0.8
predicted by the Leveque solution. This is an indicator that factors other than transport
rates are strongly inﬂuencing the interfacial coverage. Rapid molecular relaxations that
reduce the available interfacial area or lower binding afﬁnities to the surface are possible

factors.

The data presented in Figure 3.6 has been re-plotted in Figure 3.7 for the cases of
y=0.13 s‘1 and y=0.27 s". A closer examination of this data indicates that in each case
protein adsorption proceeds linearly until surface crowding causes the proﬁle to deviate
from its linear proﬁle. During the initial stages of protein adsorption, the only limiting
factor controlling protein adsorption is the rate of protein transport to the interface.
However, in later stages, interfacial crowding causes the amount of available space to

limit the adsorption kinetics. The departure of the adsorption behavior from the linear

72

proﬁle as predicted by Equation (3.22) due to crowding effects is further complicated by
protein relaxations, which serves to enhance this effect. Thus, in Figure 3.7 the departure
of the adsorption proﬁle from the solid line occurs at 0.027 mg/m2 and 0.014 mg/m2 for
y=0.27 s'1 and y=0.13 s'1 respectively. In the absence of protein relaxations, these
departures would occur at the same interfacial level. The occurrence of these departures
at different levels for shear-rates that differ by a magnitude of 2, is further conﬁrmation

of the presence of protein relaxations that effectively lower pseudo-saturation coverage.

3.5.3 Relaxations at the OTS-water interface

Preliminary investigations of albumin molecular relaxations at the OTS-water
interface have been initiated, by using the TIRF-FRAP calibration protocol for
quantifying interfacial coverage. Wertz and Santore have conducted detailed
investigations of the inﬂuence of albumin molecular relaxations on ultimate coverages at
the hexadecyltrichlorosilane (HTS) SAMs by modeling the adsorption process as
transport-limitedlz’“. In their studies, they used shear rates ranging from 1.1 s'1 to 42 s".
A limited number of experiments at the OTS-water interface were run to compare the
inﬂuence of shear rate on molecular relaxations at the solid-liquid interface and the oil-
water interface. Figure 3.8 shows results of adsorption of BSA-FITC at a bulk
concentration of 0.2 mg/ml to the OTS-water interface. The OTS SAM interface is
hydrophobic. These studies were conducted at ﬂow rates of 0.47 ml/min (r—-2.7 s")
(Figure 3.8, upper curve) and 0.047 ml/min (F027 5") (Figure 3.8, lower curve).
Interfacial coverages computed using the TIRF-FRAP protocol was approximately 0.25

mg/m2 and 0.13 mg/m2 at shear rates of 2.7 s‘1 and 0.27 s'l, respectively. Choi et al

73

 

reported similar HSA coverage at the HTS-water interface at shear rates of 1.1 s'1 69. The

following observations were made:

(i)

(ii)

(iii)

The BSA-FITC interfacial coverage at the OTS SAM is 4-5 times higher than
the corresponding coverage for BSA-FITC at the oil-water interface at the

same bulk concentration (0.2 mg/ml) and shear rate (0.27 3“).

While on the same order of magnitude, the ultimate coverage obtained is lower
than the coverage reported by Zimmerman et a]. for the same concentration,

using the TIRF-FRAP technique (> 1mg/m2).

Molecular relaxations do appear to inﬂuence BSA-FITC adsorption at the
hydrophobic solid-liquid interface for the two shear rates chosen, even at the
relatively high bulk concentration of 0.2 mg/ml. Wertz and Santore reported
similar results at the HTS-water interface for a much lower bulk concentration

of 0.01 mg/ml”.

From the observations above, it is apparent that the interfacial coverage at this

particular oil-water interface is much lower than that obtained at the OTS-water interface.

Also, since molecular relaxations appear to reduce the interfacial coverage from 0.25

mg/m2 (14:27 s") to 0.13 nrg/m2 (r-=0.27 s") at the solid-liquid interface, it is certainly

plausible that the relaxations could occur at a greater rate at the relatively more mobile

oil-water interface, where the interconnecting segments of the protein are much freer to

move, thereby reducing the coverage at the liquid-liquid interface. The slow transport rate

to the interface possibly exacerbates the situation. Finally, the reason for the discrepancy

between these results and those obtained by Zimmerman and co-workers is unclear.

74

However it should be noted that an exhaustive study at varying concentrations and shear-
rates at the OTS-water interface have not been conducted. This is the focus of ongoing
work in our laboratory. Therefore, the results at the solid—liquid interface are only

preliminary.

3.5.4 Sequential adsorption experiments

In order to examine the inﬂuence of molecular relaxations of pro-adsorbed protein
molecules on subsequent protein adsorption, sequential protein adsorption studies were
conducted. In these experiments, adsorption of 0.2 mg/ml BSA-FITC was initiated and
after 60 minutes of adsorption, the ultimate interfacial coverage was estimated by
utilizing the calibration protocol. Subsequently, BSA-FITC at a concentration of 0.5
mg/ml (step 2) was introduced and the coverage upon 60 rrrinutes of adsorption was
estimated. This was then followed by the introduction of BSA-FITC at a concentration of

1 mg/ml (step 3).

In sequential adsorption experiments, molecular relaxations of the existing layer
can affect the adsorption of proteins that are introduced later. This can happen in two
principal ways: (1) Molecular relaxations of existing protein molecules result in reduction
of the available space for later arriving molecules to adsorb and (ii) These relaxations
change the structure of the adsorbate ﬁlm, thus potentially inﬂuencing the interactions of
later arriving molecules with the existing layer. Thus, it is expected that such sequential
adsorption experiments will result in lower amounts of protein adsorption occurring in
steps 2 and 3 as compared to single shot experiments, where initial protein adsorption is

occurring at a bare oil-water interface. However Figure 3.9, shows little difference

75

between the coverages obtained in the sequential adsorption experiments and those
obtained in the single-shot experiments shown in Figure 3.4. This is a surprising result,
and suggests that transport rates have a more dominating inﬂuence on protein adsorption

at the oil-water interface.

3.5.5 Fluorescence Lifetime measurements

Fluorescence decay curves obtained using 2p-TCSPC were ﬁt to Equations (3.16)
and (3.17). The results are summarized in Table 3.2 and Table 3.3. A typical ﬂuorescence
decay curve is shown in Figure 3.10. In all cases, the two-parameter exponential model
provided a better ﬁt than the single-exponential model as conﬁrmed with hypothesis
testing using the F-test statistic. The single-exponential model also provided a reasonably
good ﬁt. For the purpose of discussion, we have chosen to use the parameters obtained
from the single-exponential model to facilitate easy comparison. Figure 3.11 and Figure
3.12 depict the proﬁles of the ﬂuorescence decay and the ﬁt to the single exponential

model.

The lifetime of the interfacially bound molecules was approximately 3.37 ns,
while the lifetimes of the proteins present in the bulk solution was approximately 3.31 ns.
Since there is no signiﬁcant difference in the two lifetimes, we conclude that the quantum
yield of BSA-FITC immobilized at the interface is unchanged from its value in solution,
thus justifying the simplifying assumption made during the development of the TIRFM-
FRAP model. The similarity in ﬂuorescence lifetimes indicates that the emission

characteristics of the ﬂuorophore are not inﬂuenced by the mobile interface.

76

The similarity in lifetimes also indicates that the dye molecules adsorbed to the
interface are free of concentration quenching effects which would lower the quantum
yield. In solution, Lakowicz et al93 has reported mean decay times of 3.12 ns for HSA-
FITC conjugated at a labeling ratio of 1. In the same study the authors reported a mean
decay lifetime of 1.84 us for HSA-FITC adsorbed to quartz surfaces. Since HSA is
known to form a complete monolayer on glass, we speculate that the rigid interfacial

properties are responsible for the decreased lifetimes.

Fluorescence decay curves were also measured for unbound FITC dissolved in
phosphate buffer. The proﬁles and the single-exponential ﬁt are shown in Figure 3.13,
and the ﬁtted parameters are given in Table 3.2 and Table 3.3. The ﬂuorescence lifetime
calculated was approximately 4.96 ns. The lifetime values obtained compare reasonably

with typical values in the literature”.

3.5.6 Other sources of error

The protocol utilized in this study is an excellent method of internally calibrating
adsorbed amounts. While care has been taken to nrinirrrize all experimental errors, there
are some sources of measurement errors that can cause inﬂuence the interfacial estimates.

These are discussed below.

3.5.6.1 Possibility of incomplete bleaching of interfacial ﬂuorophores

Bleaching experiments were conducted using a beam intensity of 15 mW. At this
intensity, it takes 2-5 seconds of pulses to bleach the interface until no further recovery is
apparent from the interface. It was observed that longer durations are required to bleach

ﬂuorophores at the oil-water interface than at the solid-liquid interface. Longer bleach

77

durations are generally not advisable, because this can result in unintended
photobleaching of bulk ﬂuorophores in a zone wider than the evanescent wave proﬁle.
The interfacial coverage calculated by Equation (3.9) is especially sensitive to the value
of Fbulk- Thus, small reductions in FM (due to over-bleaching) to near the background

level can. translate to large changes in the computed interfacial coverage.

Several experiments were conducted in an attempt to check for incomplete
bleaching of ﬂuorophores by conducting buffer wash-out experiments. In these
experiments, the background ﬂuorescence at the start of the experiment was noted. After
the adsorption and photobleaching steps were completed, a buffer wash (with protein free
buffer) was conducted to remove any ﬂuorescently labeled proteins present in the bulk
solution. In theory, if all the interfacial ﬂuorophores have been bleached, the signal after
the wash should return to the background signal. In these experiments, some residual
ﬂuorescence was found after the buffer wash. However, given the restriction on low ﬂow
rates in the experiment, a thorough wash to remove all ﬂuorophores would take in excess
of 240 rrrinutes. Thus, we cannot rule out the possibility that this residual ﬂuorescence is

due to incomplete washout of the protein solution.

3.5.6.2 Elevation of the bulk-signal contribution by light scattering:

As discussed above, the background ﬂuorescence which is the number of counts
estimated at the bare oil-water interface prior to the start of the experiment, is subtracted
from all ﬂuorescence measurements. This background ﬂuorescence is due to emission
from the oil layer, which is slightly ﬂuorescent. Due to considerable scattering of light at
the oil-water interface, the background ﬂuorescence can increase as (i) ﬂuorophores ﬁll
the ﬂow-cell, and (ii) as the protein ﬁlm evolves at the interface. If this scattering is

78

signiﬁcant, it artiﬁcially elevates the Fbuu, signal by additionally exciting the labeled
ﬂuorophores outside the evanescent ﬁeld. This can result in signiﬁcant underestimation
of the interfacial coverage. Consider, for example, the following analysis of the counts of
ﬂuorescence obtained in experiments conducted at the oil-water interface and at the solid-

liquid interface.

In a typical experiment at the oil-water interface that was conducted using BSA-
FITC at a bulk concentration of 1 mg/ml, 3595 counts per sample time were obtained
prior to photobleaching (F). The number of counts obtained after photobleaching was
2135 (Four). Prior to the start of the experiment, 1070 background counts per sample time
were recorded. Thus, in theory, the ﬂuorescence from the bulk due to evanescent wave
excitation is presumed to be 1065 counts for this experiment, while the total background
corrected ﬂuorescence prior to bleaching is 2525 counts. The ratio F/Fbuu, is calculated to
be 2.37. However if the bulk ﬂuorescence has a signiﬁcant component arising from
scattered light, it is clear that this can make a signiﬁcant difference to computation of
F/Fbuu. At a bulk concentration of lmg/ml, the bulk solution is very ﬂuorescent, even at a
labeling ratio of 1.0, and it is likely that extraneous scatter from the oil-water interface is
artiﬁcially elevating the bulk contribution by a signiﬁcant amount. Thus, it is not
inconceivable for computed interfacial coverage to double in the absence of signiﬁcant

scattering.

Similarly, in an experiment conducted at the OTS-water interface, for an albumin
bulk concentration of 0.2 mg/ml, F was 5860 counts, F bug, was 846 counts and F backgmmd
was estimated to be 472. Thus, the background corrected bulk ﬂuorescence is 373 counts

and F/Fbuy, is computed to be 14. If scattering effects are elevating the bulk ﬂuorescence

79

 

count and the evanescent wave accounts for only half of the 373 counts, this ratio doubles
to 28. Hlady and coworkers14 used TIRFM to quantify protein adsorption, and used
external calibrations to show that that scattering effects can cause a signiﬁcant (up to a 6-

fold) under-prediction of interfacial coverage.

' It is our opinion that some of the protein coverages reported in the literature are
excessively high. As a simple modeling exercise, one can consider a surface that has an
interfacial coverage of 3 mg/m2 (a value that has been reported in the literature) for a bulk
concentration of 0.2 mg/ml. Based on Equation (3.9), the ratio F/Fbuu, required to obtain a
surface coverage of 3 mg/m2 at a bulk concentration of 0.2 mg/ml is approximately 195.
In other words, assuming a background-corrected total ﬂuorescence count of 5380
counts, this ratio gives us a value of Fbulk of 27 counts, 3 value which is lower than the
average noise in the system. It is thus very unlikely that the background counts could be

as low as shown above.

Deviation in qslqb from unity: Although it has been clearly demonstrated that the
lifetime of the surface and bulk-associated ﬂuorophores are similar, assuming that the
quantum yields are also similar requires the reasonable assumption that the radiative rates
are essentially unchanged. However any changes in the radiative rate k, which offsets
changes in the non-radiative decay rate could keep the lifetime constant, while reducing
or increasing the quantum yield. Changes in k, have been reported in the literature for
ﬂuorophores in the proximity of metal surfaces82 and for ﬂuorophores in different

solvents”.

80

3.6 Conclusions

Protein interfacial coverages at the oil—water interface have been estimated using
TIRFM-FRAP. Values obtained at the interface ranged from 0.02 mg/m2 to 0.3 mg/mz.
These values are a signiﬁcant improvement over those reported previously from our
laboratory, primarily due to improved protocols. and more ”sophisticated detection
equipment”. We propose that adsorbed molecules undergo relaxations that reduce the
available area where molecules arriving later can adsorb, thereby restricting the
interfacial coverage density. The interfacial coverages at the oil-water interface in this
study are lower, in comparison to the coverage obtained by us at the same bulk
concentration and shear rate at a hydrophobic solid-liquid interface. The values obtained
for the hydrophobic solid-liquid interface using the TIRFM-FRAP methodology is
consistent with values reported in the literature. Several scenarios have been presented to
account for the differences between the results for the two interfaces. 2p-TCSPC studies
indicate that the lifetime of adsorbed BSA-FITC molecules at the oil-water interface is
similar to the corresponding value in solution, indicating that the quantum yields are not

very different.
3.7 Recommendations for future work

It would be interesting to study the inﬂuence of objective working distance and
depth of focus on the computed interfacial coverages, since objectives with higher
numerical apertures will offer a shorter depth of focus, which will more accurately reject
stray light contributions to the overall signalss. An alternative method would be to use

variable angle of incidence TIRFM (VA—TIRFM), in which the penetration depth of the

81

‘7‘ 5"" " .

evanescent wave is varied by controlling the angle of incidence. This method will allow
us to obtain an estimate of scattering effects at the liquid-liquid interface. Finally, sirrrilar
measurements at other oil-water interfaces, such as the perﬂuorocarbon-water interface (a

biologically relevant interface), will help assess the validity of this method.

The use of 2-photon spectroscopy to study lifetimes of ﬂuorescently conjugated
biomolecules at the oil-water interface is among the ﬁrst of its kind in the literature. It
would be interesting to supplement the lifetime data with measurement of rotational
diffusion and reorientation dynamics of molecules that are adsorbed at the oil-water
interface. Such studies will also provide us with information about dye and protein
partitioning at the oil-water interface. Additionally, probing the adsorbed layer using the
intrinsic ﬂuorescence of the protein (say, tryptophan) will allows us to examine the
protein in its unaltered state. At the same time, this will make it possible to obtain
information about reorientation dynarrrics of the adsorbed protein in presence of an
extrinsic ﬂuorophore and compare it against measurements obtained using the intrinsic

ﬂuorescence“.

82

 

' F(t)

Fluorescence

Fbulk(t)

 

 

 

Time (seconds)

 

 

 

Figure 3.1: A typical photobleaching experiment for calibrating interfacial
coverages. F(t) represents the total ﬂuorescence prior to bleaching. FM], (1‘) is the
total ﬂuorescence measured after bleaching is terminated.

83

Prior to photobleaching

    

region of bulk —’
adjacent to interface

 

Surface bound A few milliseconds after photobleaching
ﬂuorophores

remain bleached

- Bulk ﬂuorescence
recovers quickly

due to replacement by
unbleached molecules

 

Figure 3.2: Schematic showing the photobleaching steps involved in the calibration
protocol. The black circles represent bleached ﬂuorophores while the white circles
represent unbleached ﬂuorophores. Due to the high diffusion coefﬁcient of the bulk
associated ﬂuorophores, the bleached ﬂuorophores in the bulk are rapidly replaced by
unbleached ﬂuorophores thereby resulting in an “unbleachable bulk fraction”, while
the surface bound ﬂuorophores remain bleached.

84

   

 

   
    

, , ,, . _ .. pulse
" *' ' : " ' , diagnostics
201 -. .

    
   

cavity dumped dye laser

.,._ _.

 
 

 

ﬁber optic
560 nm - 900 nm

5. switchable
delay

     
   
 
      

polarization
control

    
 

diOde CFD delay
scrambler ﬂ

  

ratemeter

/
G-T MCP PMT
prism monochromator

Figure 3.3: Experimental layout for 2-photon time correlated single photon counting. G-

T: Glan-Taylor prism; MCP-PMT: rrricrochannel-plate photomultiplier tube; CFD:
constant fraction discriminator; TAC: Time to amplitude converter

85

0.30 -

0.25 - " ..‘

 

Cs (mg/m2)
o
i‘

 

 

0.101
1

0.054 f
,y i

o-m f ' I ' I T I ' I r I ' I ' I ' I
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

Cbulk (1118/ ml)

Figure 3.4: Adsorption isotherm of BSA-FITC at the oil-water interface using TIRFM-
FRAP protocol.

86

Initial adsorption

 

 

readin ,
Sp L

ﬂow . .

 

V

Figure 3.5: Proposed model of interfacial relaxations for albumin at an oil-water
interface. Molecules that arrive early, occupy a small area per molecule. However with

time, they spread and occupy larger areas, thereby restricting available area for molecules
that arrive much later, due to slow transport rates.

87

0.05 1

   
  

-" 0.048 mllmin
0.04 - '

. . 0.096 mllmin

0.024 mllmin

 

 

 

-0.01 , . . . , . . . , . , . . . .
0 1000 2000 3000 4000 5000 6000 7000
Time (seconds)

Figure 3.6: Adsorption of BSA-FITC (0.2 mg/ml) at the oil-water interface as a function
of ﬂow rate. Three ﬂow rates were used in this experiment: (1) 0.024 mllmin (y-0. 13 s");
0.048 mllmin 0:027 s“) and (3) 0.096 mllmin (re-0.54 s") At a ﬂow rate of 0.024
mllmin, considerably less protein adsorbs to the oil-water interface presumably due to
molecular relaxations of adsorbed protein that reduce the available interfacial area.

88

 

0.05 -

     
 

0.048 mllmin (0.27 s“)
0.04-

0.03 -

mewmh
O
s

 

‘1’
I
I
l
I
I
I
|
I
I
I
I
I
I

 

 

o '1000'2000'30'00'4000'5000'6000'7000
Time(seconds)

 

 

 

Figure 3.7: Inﬂuence of molecular relaxations on the adsorption of BSA-
FITC (0.2 myml) to the oil-water interface:. The data shown above has
been re-plotted from Figure 3.6 for y=0.27 s'1 and y=0.13 S“. The departure
of the adsorption proﬁle from the straight line, at different interfacial levels
for different transport conditions suggests the presence of relaxations.

89

0.47 mllmin

 

  
 

0.047 mllmin

.0 I.- -3-‘. ..
m“
9‘-

 

 

:5
8

. , . . . , . , . T . , . ,
0 500 1000 1500 2000 2500 3000 3500
Time (seconds)

Figure 3.8: Adsorption of BSA-FIT C (0.2mg/ml) to octadecyltrichlorosilane self
assembled monolayers at ﬂow rates of 0.47 mllmin (yr—2.7 s", top curve) and 0.047
mllmin (7:0.27 s", bottom curve).

90

 

 

 

0.10- - - " ¥

 

 

I ' I ' T ' I ' I
0.2 0.4 0.6 0.8 1 .0

 

 

Figure 3.9: Sequential adsorption experiments at the oil-water interface:
BSA-FITC at a concentration of 0.2, 0.5 and 1 mg/ml was sequentially
introduced into the ﬂow-cell and in each case, interfacial coverages were
computed using the calibration protocol.

91

10—:

 

Counts
I
I

 

 

I ' ' | V I I l r l r
0 2000 4000 6000 8000 10000
Time (ps)

Figure 3.10: Typical ﬂuorescence decay proﬁles obtained for BSA-FITC adsorbing to the
oil-water interface at a concentration of 1 mg/ml using 2—photon TCSPC.

92

120-

100-

Counts

 

 

 

   

% .

 

:1;
604

2600

I W

. , .1!

‘00'00'10000

Figure 3.11: Single exponential ﬁt to ﬂuorescence decay proﬁle for BSA-FITC (1
mg/ml) adsorbed at the oil-water interface.

93

 

 

 

 

 

 

Figure 3.12: Single exponential ﬁt to ﬂuorescence decay proﬁles for BSA—FIT C
(1ngml) in phosphate buffer.

94

ii

2

Counts
$ - F .j .

i9 .

  

 

 

 

 

 

I ' I ' I

0'20'00'400060008000'10000

Figure 3.13: Single exponential ﬁt to decay proﬁles of unbound FITC (0.15mg/ml) in
phosphate buffer.

95

Table 3.1: Comparison of initial adsorption rates obtained using the photobleaching
model with that predicted by the Leveque equation.

 

 

 

 

 

 

. Initial adsorption rate Q R: (dF/ d1).- 1
Shear ., ,2 -, d‘ tar/4r). I
.- a.. (mg m s > El
-1 S . i
Y (S ) ( ) photoblea- Le photobleac- E
. veque . Leveque .5 ..
chrng —hrng
1 0.13 940 4.65e-6 0.026 0.17 0.8
2 0.27 580 2.67e-5 0.032 1 l
3 0.54 360 2.74e-5 0.041 1.03 1.28

 

 

 

 

 

 

 

 

96

Table 3.2: Results of ﬁtting of the ﬂuorescence decay to the single exponential model
given by Equation (3.16).

 

 

 

 

 

 

Molecule Location No. a) 1'; (ns) [my
BSA-FITC 1 85.73 3.38 1.34
Oil-water
adsorbed
from 1 mg/ml Interface 2 57.18 3.36 1.26
solution
Bulk 1 34.86 3.36 1.26
BSA'FITC solution
1 mg/ml (buffer) 2 36.69 3.27 1.29
Bulk
FITC -
”mm“ 1 229.13 4.96 1.22
0'15 "lg/ml (buffer)

 

 

 

 

 

 

 

 

97

Table 3.3: Results of ﬁtting of ﬂuorescence decay to multi-exponential model given by

Equation (3.17).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Molecule Location No. a, a: 13(ns) 12 (ns) [2,“,
BSA-
FITC l 53.81 47.8 1.35 4.98 1.11
Oil-water
adsorbed
froml Interface
mg/ml 2 40.7 28.32 1.38 5.45 1.05
solution
BSA- Bulk l 36.26 4.37 2.11 38.03 1.115
FITC solution
1 mg, ml (buffer) 2 30.04 14.28 36.69 1.51 6.29
Bulk
FITC .
”mm“ 1 130.7 119.42 2.31 8.40 0.98
0'2 lug/ml (buffer)

 

 

98

1' ”Au-ﬂ
1

4 3-D ARRAYS OF LIPID BILAYERS ON POLYELECTROLYTES

MULTILAYERS

4.1 ABSTRACT

This paper presents novel and robust methods to produce 3-D arrays of lipid bilayers on
polyelectrolyte multilayer surfaces. Such arrays may be useful for high-throughput
screening of compounds that interact with cell membranes, and for applications in
biosensors and biocatalysis. Liposomes composed of 1,2-Dioleoyl-sn-Glycero-3-
Phosphocholine (DOPC) and l,2-Dioleoyl-sn-Glycero-3-Phosphate (Monosodium Salt)
(DOPA) were found to adsorb preferentially on poly(dimethyldiallylammonium chloride)
(PDAC) and poly(allylamine hydrochloride) (PAPI). Liposome adsorption on sulfonated
poly(styrene) (SPS) surfaces was minimal due to electrostatic repulsion between the
negatively charged liposomes and the SPS coated surface. Surfaces coated with
poly(ethylene glycol) (m-dPEG acid) also resisted liposome adsorption. These results
allowed us to create arrays of lipid bilayers by exposing PDAC, PAH and m-dPEG
patterned substrates to DOPA/DOPC vesicles of various compositions. The patterned
substrates were created by stamping PDAC/PAH on SPS multilayers and m-dPEG acid
on PDAC multilayers, respectively. Total internal reﬂection ﬂuorescence microscopy
(TIRFM), ﬂuorescence recovery after pattern photobleaching (FRAPP) and ﬂuorescence
microscopy were used to characterize the resulting interfaces and to check the feasibility

of the desired approach.

99

4.2 Introduction

Cell membranes are complex moieties composed of lipids, with membrane
proteins and other biomacromolecules interposed between them. They represent one of
the major structural components of biological cells. Systems that can mimic biological
cell membranes and their functionality have great potential for applications such as
biosensors and can also provide a platform for fundamental investigations of
biomolecular behavior95'99. To mimic biological cell membranes, supported bilayer lipid
membranes (sBLMs) have been formed on glass, silica surfaces and unfunctionalized
metal surfacesgs'm. Such BLMs enabled researchers to probe lipid properties such as
phase transition, lateral diffusion, permeation and lipid protein interactions “4. However,
there were some limitations associated with these supported membranes: (1) there was no
cushion between the substrate and the lipid bilayer to provide space for the hydrophilic
moieties of the membrane proteins and to give lateral mobility to the membrane
components, (2) they were fragile, and (3) no ionic reservoirs were present on either side

of the bilayer, a typical characteristic of cell membranes.

To overcome these drawbacks and allow transmembrane molecular transport, the
use of hydrophilic cushions (on which lipid bilayers are deposited) is becoming popular.
Such cushions have consisted of hydrogels, polymeric tethers or polymer ﬁlms and

22.115-130

polyelectrolytes Polyelectrolytes offer the following advantages as

cushionsm'l3 1'133: (1) they are robust and easy to fabricate, (2) they can be deposited on
many surfaces, (3) they can provide a reservoir for electron mediators and cofactors for

sensor applications and (4) their porosity and ﬂexibility may allow the protein to exist in

its natural conformation inside the lipid bilayer. Lipid bilayers composed of negatively

100

charged lipids like 1,2-Dioleoyl-sn-Glycero—3-Phosphate (Monosodium Salt) (DOPA)
and l,2-Dimyristoyl-sn-Glycero-3-[Phospho-rac-(1-glycerol)] (Sodium Salt) (DMPG)
with other zwitterionic lipids have already been shown to form on poly(allylarrrine
hydrochloride) (PAH) and poly(ethyleneirrrine) (PEI) coated substrates. These studies122
indicate that lipid coverages increase and diffusion coefﬁcients decrease on increasing
the quantity of charged lipids. In another approach, polyelectrolyte multilayers were
adsorbed on melamine formaldehyde latex particles, which were soluble at low pH,
resulting in the formation of thin polyelectrolyte shells upon dissolving the core. Lipid
bilayers were then formed on the empty shells by exposing them to charged vesicles and

the properties of this new model system as an artiﬁcial cell were then evaluatedm.

Arrays of lipid bilayers have also been fabricated and characterized on glass
surfaces'28’135‘l38. A patterned polydimethyl siloxane (PDMS) stamp was brought into
contact with a supported lipid bilayer formed on a glass slide for a short time and then
removed. Approximately 90% of the lipids in areas in contact with the stamp were
transferred to the stamp surface, resulting in an array of patches of lipid bilayers
separated from one another by regions of bare glass. The same group showed that a
bilayer can be preassembled directly onto oxidized PDMS surfaces and then transferred
intact onto a glass slide. Bilayer patches in the resulting arrays were found to be fully
ﬂuid and stable under water. To date, this methodology can only be applied to glass
surfaces. Thus, techniques that can extend the approach to other surfaces and address the

previously described limitations of sBLMs are of interest.

The approach described in this chapter, is based on the ionic layer by layer
assembly technique introduced by Decherm, rrricrocontact printing by Whitesides140 and

101

the polymer-on-polymer stamping process (POPS) developed by Hammondl‘”142

. Layer
by layer assembly can be used to deposit polyelectrolyte multilayers (PEMs) on most
substrates. PEMs are thin ﬁlms139 formed by electrostatic interactions between oppositely
charged polyelectrolyte species to create alternating layers of sequentially adsorbed ions.
PEMs are effective and economical approaches to depositing ultrathin organized ﬁlms
whose uses have included functional polymersm, colloidsm'm, biomaterials146 and

selective electroless metal Idepositionm.

Microcontact printing (uCP) is a soft
lithographic technique used in physics, chemistry, materials science and biology to
transfer patterned thin organic ﬁlms to surfaces with sub-micron resolution. Unlike other
fabrication methods that merely provide topographic contrast between the feature and the
background, uCP also allows chemical contrast to be achieved via selection of an
appropriate ink. Microcontact printing offers advantages over conventional
photolithographic techniques because it is simple to perform and is not diffraction-
lirrrited. This technique has been used to make patterns of various small and large

148-150

molecules on metals and silicon substrates as well as to deposit proteins, biological

”"153 and polyelectrolyte aggregates.”4 Polymer on polymer stamping is an

cells
approach that combines LBL assembly technique and rrricrocontact printing to generate
alternating regions of different chemical functionalities on a surface by using graft

polymers, diblock copolymers or polyelectrolytes as inkm'm

In this chapter, we present a new approach for generating arrays of 3-D lipid
bilayers and liposomes on polyelectrolytes. The height of these arrays can be precisely
controlled depending on the number of polyelectrolyte layers used to form the PEMs. The

arrays offer potential advantages over previous approaches and provide solutions to some

102

of the limitations discussed earlier. They also have excellent potential as biomimetic
interfaces for hi gh-throughput screening of compounds that interact with cell membranes,
and for probing, and possibly controlling interactions between living cells and synthetic
membranes. Arrays of lipid bilayers were created by exposing
poly(dimethyldiallylammonium chloride) PDAC patterns, polyethylene glycol (m-dPEG
acid) patterns and poly(allylarrrine hydrochloride) (PAH) patterns on PEMs to liposomes
of various compositions. Total internal reﬂection ﬂuorescence microscopy (TIRFM) was
used to monitor liposome adsorption and desorption to the PEMs. Fluorescence recovery
after pattern photobleaching (FRAPP) and ﬂuorescence microscopy were used to
characterize the resulting interfaces. The fabrication of the 3-D array patterns and quartz
crystal microbalance studies was conducted by Neeraj Kohli while TIRFM adsorption
studies and FRAPP characterization studies were conducted by the author of this

dissertation. All results obtained are reported for completeness.

4.3 Experimental section

4.3.1 Materials

Sulfonated poly(styrene) (SPS) (Mw~70,000), poly(allylanrine hydrochloride)
(PAH) ( Mw~50,000) and poly(diallyldimethyl ammonium chloride) (PDAC)
(Mw~100,000) were obtained from Sigma. l,2-Dioleoyl-sn-Glycero-3-Phosphate
(Monosodium Salt) (DOPA), 1,2—Dioleoyl-sn-Glycero-3-Phosphocholine (DOPC) and 1-
Palmitoyl-2-[6-[(7-nitro-2-1 ,3-benzoxadiazol-4-yl)amino]hexanoyl]-sn-Glycero-3-
Phosphocholine(16:0-06:0 NBD PC) were purchased from Avanti Polar Lipids. 1-

tetradecanethiol, 4—(2-hydroxyethyl)piperazine-l-ethanesulfonic acid sodium salt

103

(HEPES) was obtained from Fluka. The m-dPEG molecule was purchased from Quanta
Biodesign. Sylgard 184 silicone elastomer kit (Essex Brownell) was used to prepare the
PDMS stamps used for rrricrocontact printing. Structures of PDAC, SP8 and m-dPEG are
shown in Figure 4.1. The ﬂuorosilanes were purchased from Aldrich Chemical. Ultrapure
water (18.2MQ) was supplied by a Nanopure-UV four-stage puriﬁer (Bamstead

International); the puriﬁer was equipped with a UV source and a ﬁnal 0.2 pm ﬁlter.

4.3.2 Preparation of stamps:

An elastomeric stamp was made by curing poly(dimethylsiloxane) (PDMS) on a
rrricrofabricated silicon master, which acts as a mold, to allow the surface topology of the
stamp to form a negative replica of the master. The poly(dimethylsiloxane) (PDMS)
stamps were made by pouring a 10:1 solution of elastomer and initiator over a prepared
silicon master. The silicon master was pretreated with ﬂuorosilanes to facilitate the
removal of the PDMS stamps from the silicon masters. The mixture was allowed to cure
overnight at 60 °C. The masters were prepared in the Microsystems Technology Lab at

MIT and consisted of features (parallel lines and circles) from 1 to 10 um.

4.3.3 Preparation of liposomes:

Small unilamellar vesicles were prepared by mixing appropriate amounts of
DOPC/DOPA with 1% NBD-PC in chloroform. This mixture was then dried under
nitrogen and care was taken to ensure that the lipids form a thin cake like ﬁlm on the
walls of the test tube. The residual chloroform was then removed under high vacuum.
The lipids were then reconstituted in either PIEPES, pH 7.4 (0.1M NaCl) or 18.2 M!)

water. The resulting liposome solution was then sonicated until they became clear inside

104

f.

an ice bath using a Bransonic tip sonicator or bath sonicator. In some of the initial
experiments, vesicles were made by extruding the reconstituted lipids through a rrrini-

extruder.

4.3.4 Preparation of arrays:

Figure 4.2 and Figure 4.3 show two different schemes used to prepare arrays of
lipid bilayers. Scheme 1: (Figure 4.2) A PDMS stamp was dipped in a 250mM solution
of PDAC (or 200mM solution of PAH) in 75/25 ethanol-water mixture for about 20
rrrinutes. The stamp was then washed with ethanol, dried under nitrogen and brought into
contact with a glass slide that was coated with PDAC/SPS multilayers, with SP8 forming
the uppermost layer. The stamp was removed after 15 rrrin and the resulting PDAC (or
PAH) patterns were then rinsed with water to remove unbound or loosely bound PDAC.
The slide was then exposed to DOPC/DOPA liposomes of varying compositions, and

imaged using a microscope with attached camera.

Scheme 2: (Figure 4.3) A PDMS stamp was dipped in a 10011M solution of m-

dPEG acid in 75/25 ethanol-water mixture for 30 rrrinutes.155

The stamp was then
washed, dried under nitrogen and brought into contact with the glass slide coated with
PDAC/SP8 multilayers with PDAC as the topmost layer. The stamp was removed after

20 nrinutes and the resulting PEG patterns were then rinsed with water to remove excess

PEG.

All TIRFM and FRAPP characterizations were done on non-pattemed slides (i.e.

slides having a uniform top layer).

105

 

4.3.5 Total internal reﬂection microscopy:

The experimental setup and the ﬂow cell has been described previously5 as well
as in chapter 2 (Figure 2.1) . All ﬂuorescence measurements were made using the photon
counter. As in the case of experiments described in Chapter 2 and 3, the optical chopper
was used to prevent unintended photobleaching of ﬂuorophores during all adsorption
experiments. The only modiﬁcation to the experimental procedure described in Chapter
2, was that the data was sampled continuously, rather than intermittently, since we did
not observe signiﬁcant photobleaching by the monitoring beam at the incident laser
intensities chosen for these experiments. A 500 nm long—pass band ﬁlter was used to
separate the excitation and emission intensities. The photon counter was triggered by an
output reference voltage from the chopper, so that data collection only occurs during
periods when the ﬂow cell is illuminated by the laser beam. For TIRFM experiments, the
ﬂow cell was initially ﬁlled with buffer. Subsequently, liposome solution was introduced
into the ﬂow cell at a controlled ﬂow rate for 10-12 minutes. The infusion was then
halted and adsorption was allowed to continue for 1 hour, followed by a buffer wash (for

desorption or TIRF-FRAP experiments).

4.3.6 Fluorescence recovery after photobleaching (TIRF-FRAP):

A system of optical ﬂats5 enables us to easily switch between a low intensity
monitoring beam for observation of surface dynamics and a high intensity beam for
conducting ﬂuorescence recovery after photobleaching (FRAP) experiments in TIRF
mode (Figure 2.1). After liposome adsorption was complete, TIRF-FRAP experiments
were conducted by application of one to three bleach pulses followed by monitoring of
the ﬂuorescence recovery. Experiments conducted in the TIRF-FRAP protocol utilize

106

spot diameters signiﬁcantly greater than 50 um and are used to characterize systems with
faster diffusivities. Molecules with lower diffusivities are more accurately characterized
by ﬂuorescence recovery after pattern photobleaching (discussed later) in which the
characteristic length scale for diffusion is much smaller than in TIRF-FRAP

experimentation.

X-Y stage calibrations: In TIR mode, the area illuminated by the laser beam
waist at the interface is typically characterized by an elliptical spot. The accurate
characterization of the dimensions of this spot is not a trivial undertaking due to the
location of the beam waist. Thus, classical methods that employ translation of the beam
across a knife edge in order to determine spot diameters cannot be applied here. Spot
‘ diameters were estimated by photobleaching a spot and translating the stage in the X and

Y planes across the bleach spot5(Refer to Figure 5.3 and Figure 5.4 in Appendix.)

4.3.7 Fluorescence recovery after pattern photobleaching (FRAPP):

The experimental conﬁguration used for FRAPP has been depicted in Figure 4.4.
Stripe patterns were imposed on the substrate by directing the 488 nm laser beam
(expanded through a 5X beam expander (Edmund optics, Barrington NJ) through a 50 or
100 line per inch Ronchi ruling (Edmund optics, Barrington,NJ) placed in a real image
plane. Placing a ruling in a back image plane, results in projection of a sharply focused
pattern of alternating dark and bright fringes on the substrate in the sample plane. This
back real image plane is located near the epi-port of the Axiovert 135M microscope,

coincident with the ﬁeld iris diaphragm of a ﬂuorescence light illuminator (Zeiss)

 

5 A complete description of this method and stage calibration charts is presented in Appendix C

107

mounted through the epi-port. The ﬂuorescent light illurrrinator has a lens that projects
the beam through a Zeiss ﬁlter cube (Ex: 450-490/DM: SIG/Em: 515-565), and a 32X
Zeiss objective on to the substrate. For focusing the objective, a ﬂuorescent coating was
brushed onto a glass slide using a yellow ﬂuorescent marker (Sanford). The objective
position was then adjusted till the fringes appeared in sharp focus. The fringe spacing was
estimated by using a reticule containing 10 lines per centimeter in the eyepiece. The glass
slide was then replaced by the sample slide. Prior to initiating photobleaching, minute

adjustments were made to the objective focus knob to bring the fringes into sharp focus.

Beam alignment was checked at the start of each experiment and objective
focusing was performed prior to initiation of every FRAPP experiment. For FRAPP
experiments, it is crucial that the monitoring beams and the photobleaching beams be
precisely coincident. Checking that the beams coincided was accomplished with the
following steps. (1) The monitoring and photobleaching beams were projected on to a
screen some distance away from the vibration isolation table, and beam recombination
was checked.(2) Occasionally, as an additional check for recombination, once the laser
beam was directed through the Ronchi ruling so that fringe pattern was formed on the
ﬂuorescently coated glass slide, a CCD camera or the microscope oculars was used to re-
conﬁrm that the beams recombined by observing the overlaying of the faint monitoring
fringes over the photobleaching fringes6 156. For all experiments, the monitoring beam
intensity ranged from 111W to 511W. For photobleaching, the beam intensity required was

approximately 0.5W. Additional neutral density ﬁlters (NDF) were required to attenuate

 

6 Precautionary note:- Laser safety glasses must be used during alignment, especially when viewing the
sample through the oculars.)

108

the monitoring beam to the desired level. The use of additional NDFs causes signiﬁcant
deviation in the path of the monitoring beam, so checking for recombination is especially
crucial. A high bleaching intensity is required to obtain a sufﬁcient bleach depth within a
reasonable time. Typical bleach times varied from 350ms to 1 second. An aperture placed
in the image plane in front of the PMT or CCD camera was used to restrict the
observation area. Thus, the illuminated area was approximately 200 um, while the
observed area was 150 um. Stripe periodicity in the sample plane was approximately 12.5

11m. A typical fringe pattern obtained using this conﬁguration is shown in Figure 4.5.

For FRAPP experiments, liposome adsorption on the polyelectrolyte coated
substrates was initiated by directly introducing the liposome solution at a ﬂow rate of
0.34 mllmin for approximately 10-12 nrinutes using the syringe pump. The infusion was
then halted and adsorption was allowed to continue for approximately 45 minutes. The
ﬂow cell was subsequently ﬂushed with 4-5 ﬂow cell volumes of buffer in order to

remove the ﬂuorescently labeled liposomes in the bulk solution.

4.4 Theory and Data Analyses for Fluorescence recovery after pattern

photobleaching

FRAP is a technique that is commonly used to obtain estimates of translational
(lateral) mobilities of proteins or lipids. There are two principal variants of this method.
The ﬁrst involves using a focused laser beam to create a small spot. This is known as spot
photobleaching and can be effected using epi- illumination as well as using TIRFM. In
the second method (FRAPP), a laser beam is passed through a Ronchi ruling placed in a

back image plane to create a pattern of alternating dark and bright stripes of well deﬁned

109

periodicity over a broadly illuminated area. The principle advantage of FRAPP over spot
photobleaching lies in the well deﬁned characteristics of the pattern obtained in the
sample plane. For spot photobleaching, the recovery kinetics and shape depend very
strongly on the shape of the focused spot at the interface. The precise shape of the spot is
difﬁcult to discern as the beam travels through several optics in order to form the spot. As
a result there is considerable uncertainty in measurements of diffusion coefﬁcient
especially for non-ideal samples. In contrast the well deﬁned periodicity of the stripe
pattern produced in the sample plane offers two distinct advantages: (i) It allows us to
measure slow as well as fast diffusion coefﬁcients (10’lo cm2/s to 10'7 cmzls), as the stripe
periodicity can easily be varied and (ii) It allows us to examine samples where multiple
populations with different mobilities coexist, using models that describe such
populations. Other variants of FRAPP involve creation of the stripe pattern using the
intersection of two beams either in the sample plane (using TIRF illumination) or in a
back image plane (using EPI illumination). The advantage of using TIRFM-FRAPP over
EPI-FRAPP is that due to the surface selectivity of TIRFM, a buffer wash to eliminate
bulk ﬂuorophores is not necessary. The trade-off is the complexity in the experimental

conﬁguration.

4.4.1 Theory

A complete mathematical analysis of ﬂuorescence recovery after pattern

photobleaching has been presented by Starr and Thompsonm

. A brief description of this
analysis is presented below. Consider a sample of mobile ﬂuorescent lipids with diffusion

coefﬁcient D. The ﬂuorescence emission from such a sample can be described as:

110

F(t)=Q j j I(x.y)C(x.y.r>dxdy (4.1)

—oe-oo

where Q is a proportionality constant that incorporates factors such as ﬂuorophore
quantum yield and the instrument constant , I(x,y) is the monitoring beam intensity and
C(x,y,t) is the concentration of unbleached molecules as a function of position and time.
The equation that describes the intensity proﬁle for a Gaussian-shaped laser beam

intersected by a Ronchi ruling placed in a back image plane is given by

nodd

2 2
10., y) = 12am, [_E%;y—)]'[l+ Z c" cos(nkx)] (4.2)

Here, 10 is the intensity at the origin, .9 is the l/ez radius of the expanded beam, k is the

spatial frequency of the pattern, superimposed on the sample plane, deﬁned by:

(—1)[n7_l] (4.3)

k=— ;c =—4—
a " n7:

and a is the spatial period of the stripe pattern. The concentration proﬁle of diffusing

ﬂuorophores is,

_0c—x)2+(y—-y)2
4Dt

 

C(x,y,t)=%Dt-I IC(x',y',0).exp[ de'dy' (4.4)

The initial concentration of unbleached ﬂuorophores in the sample is given by:

C(x, y,0) = Cexp[-K1(x, y)] (4.5)

111

where C is the total ﬂuorophore concentration and K is a constant that incorporates the

bleach pulse duration, quantum yield and absorptivity of the ﬂuorophores.

After developing Equation (4.1) using Equations (4.2), (4.3), (4.4) and (4.5), for cases
where the stripe periodicity is small compared to the illuminated area, the following

speciﬁc instances can be considered:

4.4.1.1 Samples containing a single diffusive population:

For bilayer systems containing ﬂuorophores with a single diffusion coefﬁcient, the

ﬂuorescence pattern photobleaching recovery proﬁle ﬁt) is given by158

2 2
¢(t) .—..¢(0)+%[1—¢(0)].|:1-[-8?].{exp[— 4” 21>: J+lexp[— 36” 2D ’ M (4.6)
It a 9 a

 

 

fort 20, where ¢(t) is the ratio of the postbleach ﬂuorescence (t>0 , after the bleach
pulse) to the prebleach ﬂuorescence ¢(t<0) ,,u represents the fraction of the ﬂuorophores

present that are mobile and a and D have already been deﬁned previously. This analysis

neglects rapidly decaying (higher order) terms.

4.4.1.2 Samples containing multiple diffusive populations (It 21)

For samples containing populations with multiple diffusion coefﬁcients and an immobile

fraction, the recovery proﬁle can be characterized bylsg,

1 " 8 4220.: 1 36ﬂ'2Dit
¢(t)-¢(0)+'2'[1"¢(0)]- Eﬂi[1'[;§)-ECXP[' a2 ]+§cxp[- a2 ﬁ] (4-7)

 

 

112

where ,u, is the fraction of the ﬂuorescence recovery from mobile ﬂuorophores with

diffusion coefﬁcient 0;.

4.4.2 Data Analyses and curve fitting

Data ﬁtting was accomplished using OriginPro 7.5 (OriginLab Corporation, Northampton
MA ) which uses the Levenberg-Marquardt algorithm for non-linear least squares ﬁtting.
The post-bleach ﬂuorescence emission intensity was normalized against the pre-bleach
ﬂuorescence emission intensity (t<0). All data sets were ﬁtted to the models given
above by weighting datasets with the Poisson error and an F-test of comparison was
conducted to determine whether the model using higher parameters provided a
statistically signiﬁcant improvement in the goodness of the ﬁt. This F-statistics was

deﬁned as:

(2,2 -222)(N —5)

F:
2122

 

(4.8)

where x12 and 122 are the chi-squared goodness-of-ﬁt statistics for the models described
by Equations (4.6) and (4.7) respectively. If this F statistic exceeds 3.0, the ﬁt using the

model with more parameters is considered as a signiﬁcant improvement‘sg'lm.

Equations (4.6) and (4.7) describe proﬁles where the recovery of ﬂuorescence from
diffusing ﬂuorophores comes from diffusion between the illuminated and non-

illuminated stripes. This model does not account for the possibility of recovery

113

originating from diffusion between the illuminated region and outside the illuminated

regionm.

4.5 Results and Discussions

4.5.1 TIRFM Adsorption experiments:

The curves in Figure 4.6 depict adsorption of ﬂuorescently tagged liposomes
composed of varying amounts of DOPA and DOPC on PEMs with either PDAC or SPS
as the top layer. In Figure 4.6, curves A and B depict adsorption of negatively charged
liposomes composed of 90% DOPC / 10% DOPA to PEMs with positively charged
PDAC and negatively charged SPS as the top layers respectively. The higher intensities
obtained in curve A indicate that the liposomes adsorbed onto PDAC preferentially to
SP8 (curve B) presumably due to electrostatic interactions. To investigate the role of
liposome charge on subsequent adsorption, we conducted adsorption experiments with
liposomes composed of 80% DOPC and 20% DOPA on PEMs (curves C and D) with
PDAC and SPS as the top layer respectively. (Note: Curves C and D were experiments
conducted with a separate batch of liposomes from those used to obtain curves A and B.
Due to batch to batch variability in liposome characteristics and small changes in labeling
ratios, etc., we are cautious about directly comparing the ﬂuorescence emission data
obtained from different batches.) The higher concentration of negatively charged lipids in
the liposomes further increased the adsorption selectivity for PDAC suggesting that
adsorption of liposomes on PEMs is signiﬁcantly inﬂuenced by strong electrostatic
interactions between the charged lipids and polyelectrolytes. The increasing ﬂuorescence

intensities even after 1 hour for curve C may suggest that multilayer deposition occurs on

114

increasing the amounts of DOPA in the liposomes. For the remainder of the studies in
this chapter, liposomes composed of 10% DOPA and 90% DOPC were used, to minimize

the likelihood of multilayer deposition.

To determine the reversibility of liposome adsorption to PDAC/SPS PEMs the
ﬂow cell was ﬂushed with 34 volumes of liposome-free buffer. This wash resulted in the
removal of signiﬁcant amounts of liposomes adsorbed to PEMs with SPS as the top layer
(Figure 4.7). In comparison, the binding of liposomes to PEMs with PDAC as the top
layer is relatively strong and a buffer wash resulted in a smaller reduction in total
ﬂuorescence (Figure 4.8, top curve and Figure 4.9b, top curve). Since the evanescent
”wave decays rapidly with distance from the interface, only a small contribution to the
total ﬂuorescence comes from illumination of ﬂuorophores in the bulk solution.
Therefore, while the buffer wash experiments may reﬂect, in part, depletion of liposomes
in the bulk liquid, it is much more likely that the results suggest that lipid binding to

PDAC is stronger than to SPS.

Liposome adsorption on m-dPEG tailored surfaces was also studied. The ability
of PEG to resist biomolecular adsorption has been well established in the literature'm. In
Figure 4.8, the upper curve depicts liposome adsorption onto PEMs with PDAC as the
top layer, while the bottom curve depicts liposome adsorption on m—dPEG acid layers.
Thus the m-dPEG layer resists liposome adsorption. In such experiments, buffer wash
steps are more crucial as it allows us to further examine whether the liposomes that do
adsorb are loosely or tightly bound. The results of Figure 4.8 have been re-plotted in
Figure 4.9b, to show the liposome desorption more clearly. Figure 4.9b shows proﬁles for
experiments in which a buffer wash was initiated after approximately 45 minutes of

115

liposome adsorption. At t=0 min, liposome adsorption to PDAC (top curve) and m-dPEG
(bottom curve) PEMs was halted by introducing liposome-free buffer and only the
desorption proﬁles are shown. In each curve the ﬂuorescence intensity has been
normalized against the corresponding ﬂuorescence value obtained prior to initiation of
the buffer wash in each experiment. There is a 70% decrease in the ﬂuorescence emission
intensity for the m-dPEG case as compared to a 10% decrease in the case of PDAC. A
similar TIRFM study where the adsorption/desorption of liposomes on slides coated with
lipid monolayers is monitored, has been reported in the literature'“. That study indicated
that the kinetics 0f liposome adsorption and desorption were strongly dependent on the
solution ionic strength and the lipid concentration in solution. Much research has been
conducted to determine the mechanism of PEG resistance to adsorption of biomolecules
and many theories have been proposed to explain the resistive ability of PEG molecules.
While the mechanism for biomolecular resistance is not completely clear, it appears to

stem either from steric exclusions between the biomolecule and the PEG chain'“’164 or

from long range electrostatic repulsions165 .

We have observed that in case of liposome adsorption onto multilayers with SPS
and PEG as the top layer, in spite of electrostatic or steric repulsions, some liposome
adsorption does occur. In Figure 4.9a, although the rate of liposome adsorption to the m-
dPEG surface is much diminished in comparison to that for PDAC, no saturation of
ﬂuorescence is detected, and the signal would appear to continue increasing if the
experiment had not been terminated at 45 minutes. Thus, the additional step of ﬂushing

the ﬂow-cell with liposome-free buffer results in removal of large amounts of loosely

116

bound liposomes and is an important step during formation of BLM arrays. This was

conﬁrmed with ﬂuorescence rrricrographs which have been discussed in the next section.

It is also interesting to note the shapes of the adsorption curves obtained for the
different PEM surfaces. For example in Figure 4.9a for adsorption to m—dPEG PEMs, a
sharp break in the adsorption proﬁle occurs at approximately 400 seconds, at which point
there is a change in the adsorption rate, but no apparent saturation. In the case of PDAC
(Figure 4.8), after this break, the ﬂuorescence saturates indicating no additional vesicular
adsorption. Some researchers propose that this characteristic break indicates the point at
which a single bilayer is formed'éz. Beyond this point, any further increase in
ﬂuorescence can be construed as occurring due to the reversible adsorption of vesicles to
the bilayers. Thus multiple breaks can be attributed to the formation of multilayers. It is
interesting then, to note, that in the case of liposome adsorption to m—dPEG, a second
layer of vesicles appears to adsorb in spite of the electrostatic repulsions between the

negatively charged bilayer and vesicles.

It should also be noted that direct comparison of ﬂuorescence emission intensities
in the experiments described in this section, requires the assumption that the quantum
yield of the dye is unaffected by its proximity to the charged polyelectrolyte surface.
While we have not measured ﬂuorophore lifetimes to conﬁrm this, Nollert and coworkers
measured the lifetime of NBD-PE in a POPC BLM on glass and found it to be

comparable to its value in a vesicle'“.

117

4.5.2 Arrays of lipid bilayers:

Two different schemes were used to fabricate arrays of lipid bilayers using
methods detailed in Section 4.3. Figure 4.10a and Figure 4.10b show the resulting
ﬂuorescent images of the line and circular patterns respectively. Images in this
dissertation are presented in color. Consistent with the adsorption results, as indicated by
the ﬂuorescent features of the line and circular arrays and the clean background regions,
liposomes bind preferentially to PDAC features and negligibly to the SPS background.
Similar ﬂuoresence results were obtained when we stamped a weak polyelectrolyte PAH
instead of a strong polyelectrolyte PDAC on SPS. Figure 4.10c shows the resulting
ﬂuoresence image. It can be clearly seen that the liposomes adsorb preferentially on PAH
in comparison to SPS as the lines are well deﬁned. The ﬂuorescence micrographs clearly
indicate that the system is homogeneous with the exception of few bright spots which

may be due to crystalline dye.

In another approach m-dPEG acid was stamped on a glass slide coated with
multilayers with PDAC being the topmost layer. The m-dPEG acid molecule has a
carboxylic acid group on one end; therefore at a pH above the pKa of the carboxylic
group it has a negative charge and can therefore be stamped on PDAC, resulting in
patterns of m-dPEG acid on PDAC155 . As discussed earlier, m—dPEG acid also resists
liposome adsorption. Arrays of lipid bilayers could be created in this case by exposing m-
dPEG acid patterns to liposome solution. Liposome adsorption occurred only on the
exposed PDAC while m-dPEG acid patterns resisted their adsorption. Figure 4.11a and
Figure 4.11b show the obtained line and circular patterns. In this case liposomes bound to

the background pattern rather than the circular or line features. Thus the ﬂuorescent

118

patterns seen in Figure 4.11 are a negative replica of those seen in Figure 4.10. The
ability to make either the positive or negative image of the stamp adds to the versatility of

this method.

4.5.3 TIRF spot photobleaching:

Figure 4.12 depicts TIRF-FRAP data for bilayers formed on PDAC PEMs. For
this particular experiment, the beam waist was characterized as an elliptical proﬁle 166
um along the major axis and 70 run along the nrinor axis (Refer Appendix C, Figure 5.4).
This can be roughly approximated to a circular spot of 108 um diameter. We observed no
ﬂuorescence recovery after application of the bleach pulse. Over long durations, the
ﬂuorescence signal appears to drop due to photobleaching by the monitoring beam
(Figure 4.12a). We also conducted recovery experiments, where the recovery was
intermittently monitored over longer time scales in order to prevent photobleaching
effects by the monitoring beam (Figure 4.12b). We observed no signiﬁcant ﬂuorescence

recovery under these conditions.

For bleach spots with length scales in the region of 50—100 microns, recovery
proﬁles in a reasonable short time duration (<1000s) can be obtained for lipid bilayers
with higher diffusivities (10'7 cmzls). However for bilayers with diffusivities in the region
of 10'9- 10'10 cmzls, characteristic recovery times are signiﬁcantly larger, and hence
FRAPP is more accurate for interrogating lipid diffusion in BLMs with lower diffusion
coefﬁcients. We did not obtain signiﬁcant recoveries on PDAC surfaces using
conventional TIRF photobleaching; diffusivities for these bilayers appeared to be

considerably lower and hence we used FRAPP to obtain estimates of these values.

119

4.5.4 Fluorescence recovery after pattern photobleaching:

FRAPP was used to determine if the adsorbed liposomes remain intact on the
surface or fused together to form a bilayer. Models used to describe recoveries for species
containing one or more mobile fractions were used to ﬁt the data (Equations (4.6) and
(4.7))158'159. Both recoveries on PAH and PDAC were better described by the two
mobile-species model which is shown to provide a statistically better ﬁt than the one-
mobile species model. Figure 4.13 and Figure 4.14 show typical recovery curves for
lipids deposited on PDAC and PAH respectively. The mobilities and diffusion
coefﬁcients for populations on PDAC have been summarized in Table 4.1 (one mobile
species) and Table 4.2 (two-mobile species) while the mobilities and diffusion
coefﬁcients obtained for bilayers on PAH, have been summarized in Table 4.3 (one--

mobile species) and Table 4.4 (two-mobile species).

Although the data are well represented by the two-mobile species model, it is still
unclear whether there is an actual underlying physical basis for the presence of two
diffusing components. If the data above is indeed described more accurately by a two
component model, there are two principal hypotheses which could explain the presence
of two distinct populations with different mobilities. (1) Since bilayer deposition occurs
through liposome adsorption to the PEM surface, the dye (NBD-PC) is present both in
the lower as well as the upper leaﬂet of the lipid bilayer. The bottom leaﬂet is more
strongly bound to the PEM cushion with PDAC or PAH as the topmost layer on account
of favorable electrostatic interactions between the charged lipids and the surface. If so,
diffusion in the lower leaﬂet would be slower than the upper leaﬂet. However, in that

case, assuming a random distribution of dye molecules across the two leaﬂets (since they

120

are randomly distributed in the liposomes), one would expect to obtain equal amplitudes
for the two mobile fractions, which does not hold true (Refer Table 4.2 and Table 4.4) (ii)
There may be phase separation in the lipid bilayer due to the imrrriscibility of the two
lipids. The dye is thus located in two distinct regions, a more viscous (lower mobility)
region and a less viscous (higher mobility region). However for our lipid system, the
phase transition temperature of both lipids is below -8 °C (obtained from Avanti Lipids),
and the lipids (DOPC and DOPA) can be considered to be miscible. Also, ﬁts from the
two-mobile species model show considerable variation in the parameters from one spot
on the sample to another for bilayers on PDAC as well as PAH (Table 4.2 and Table 4.4).
In addition, the error associated with the ﬁt parameters is large. The ﬁts obtained using
the single mobile-species model are much more consistent. In light of these observations,
even though the two-mobile species model offers a superior ﬁt, we chose to characterize
the sample as a system containing a single population of mobile ﬂuorophores in addition

to an immobile fraction.

The results summarized in Table 4.1 and Table 4.3 indicate that bilayers formed
on PAH have higher mobile fractions (>0.65) than those formed on PDAC substrates
(<0.37). Therefore, a substantial fraction of lipids deposited on PDAC are immobile.
Since FRAPP measures long range diffusion much greater than the diameter of a
liposome (20—100 run), this suggests that some of the adsorbed liposomes remain
unruptured. Nollert and coworkers166 have proposed two possibilities that occur when
liposomes adsorb to surfaces. In one scenario, liposomes adsorb unruptured, as supported
vesicles (which represent the immobile fraction). In the other case, they adsorb, fuse and

spread to form sBLMs. It is also possible that lipids present in the bilayer or liposomes

121

may exist in different phases that have varying degrees of mobility. Thus, lipids in phases
of low mobility could also contribute to the immobile fraction. While FRAPP can provide
indirect evidence about the presence of multiple populations, it is difﬁcult to use this

technique to distinguish between an intact liposome and an immobile lipid phase.

We expected higher mobile fractions for bilayers formed on PDAC, because it is a
strong polyelectrolyte that we expect should facilitate liposome rupture and bilayer
formation more then a weak polyelectrolyte such as PAH. However we observed higher
mobile fractions on PAH in comparison to PDAC. The following theories may explain
this observation: (1) An examination of the PDAC and PAH molecules indicates that the
charge on PDAC is shared by more atoms than PAH. This may inﬂuence the interaction
of the liposome with the PDAC PEM in a way that reduces inhibits rupture (ii) Jenkins
and coworkersm have proposed on the basis of experimental observations that the
interactions of vesicles with a hydrophilic/hydrophobic boundary causes rupture and
bilayer formation. A recent theoretical analysis168 predicted that a heterogeneous surface
(i.e. surface having both hydrophobic and hydrophilic moieties) is more likely to
facilitate bilayer formation than a completely hydrophilic surface. While both PDAC and
PAH have hydrophilic groups and hydrophobic backbones, it is possible that the bulky
side chain on PDAC, reduces accessibility of the liposome to the hydrophobic backbone.
Thus, PAH exhibits greater hydrophobicity than PDAC which may explain the higher
mobile fractions. It is also possible that differences in the extent of swelling of the PDAC

and PAH PEMs can inﬂuence liposome rupture.

The diffusion coefﬁcients obtained in this study are comparable to those reported
for bilayers containing SOPS and POPC deposited on PDAC/SPS multilayers (2 x 10'9

122

cmZ/s)'69"70. The diffusion coefﬁcients obtained in this study are lower than values
reported in the literature the literature for DMPCzDOPA (10: 1) bilayers deposited on
PSS/PAH multilayersl31. The average diffusion coefﬁcient for bilayers on PDAC was

0.0721003 umz/s, while bilayers on PAH exhibited a diffusion coefﬁcient of 0.0565 1

0.009 umzls. One possible explanation for this trend is the location of the dye molecule
with respect to the bilayer. In our study, NBD was attached to one of the hydrophobic
tails of phosphocholine; therefore, its location in the interior of the bilayer may make it
less mobile. For example, diffusion coefﬁcients for lipids have been reported to be four

times greater if the NBD molecule is attached to the head-group than to the tail of the
phosphocholine moleculem. Moreover, other groups generally use dye molecules only in
the upper leaﬂet and, as a result, they measure the diffusion coefﬁcient only of the upper

leaﬂet. The upper leaﬂet would be less inﬂuenced by the underlying substrate and would

be expected to exhibit a higher diffusion coefﬁcient than the bottom leaﬂet.

Studies of liposome adsorption and rupture on PDAC and PAH PEMs have also
been conducted using QCM. The results obtained in these studies are consistent with
trends reported using FRAPP and can be found in APPENDIX D: Quartz crystal

microbalance studies.
4.6 Conclusions

3-D arrays of liposomes and BLMs have been fabricated on PEMs. Such arrays have
potential applications in biosensors, biocatalysis and devices that may use high
throughput screening. TIRFM and ﬂuorescence microscopy results suggested that

liposomes composed of DOPA and DOPC adsorbed strongly on PDAC and PAH

123

surfaces, but weakly on SPS. Poly(ethylene glycol) (m-dPEG) coated surfaces resisted
liposome adsorption. These results allowed us to create rrricro arrays of lipid bilayers.
Lipid diffusion coefﬁcients for lipids deposited on PDAC and PAH were in the range of
10‘9 cm2/s. FRAPP results suggest that a higher fraction of liposomes rupture to form
bilayers on PAH surfaces than on PDAC, where a signiﬁcant percentage of the adsorbed

lipids are essentially immobile.

The mechanism of liposome rupture and bilayer formation is not fully understood.
However, nanoscale surface heterogeneity is known to facilitate liposome rupture. Thus,
the ability to tune the characteristics of the interface using different kinds of
polyelectrolytes may help to elucidate the underlying mechanisms for bilayer formation

and also allow BLM formation to be controlled.
4.7 Recommendations for future work

One possible study for examining liposome rupture and bilayer formation on these
charged surfaces involves making epi-FRAPP measurements on liposomes that adsorb to
the PEMs in the presence of an electric ﬁeld. The presence of the electric ﬁeld can alter
the conformation of the charged PEMs thereby altering the surface characteristics which
in turn may inﬂuence liposome rupture. This study will allow us to probe the underlying
mechanisms for liposome rupture and bilayer formation, and will also allow us to tune
the characteristics of the surface to optimize it for deposition of BLMs. Another
interesting method to study vesicular fusion and rupture is using Fluorescence Resonant
Energy Transfer (FRET). FRET is the radiationless transfer of energy from a donor

molecule to an acceptor molecule. The amount of energy transfer is dependent on the

124

extent of overlap between the emission spectrum of the donor and the absorption
spectrum of the acceptor. This rate of energy transfer varies inversely with the 6th power
of the distance between the donor and the acceptor”. FRET has been used for studying
the mechanism of initial rupture of liposomes using single vesicle ﬂuorescence assaysm.
It would be useful to probe the mechanism of liposome rupture on the highly charged

polyelectrolyte substrates using such assays.

It will also be helpful to characterize liposome adsorption under a multitude of ﬂow
rates and ionic strengths to examine the shape of the adsorption proﬁle, since the ionic
strength of the solution will have a strong inﬂuence on the structure of the highly charged
polylectrolytes. In addition, this study will allow us to examine the phenomenon of
multibilayer formation. In fact, we have already attempted to fabricate '3-D structures
composed of alternating layers of lipid-bilayers and polyelectrolytes in our ﬂow-cell, by
making sequential introductions of liposomes and polyelectrolytes and monitoring the

ﬂuorescence emission.

Another rapidly emerging thrust of research is in the area of lipid rafts. Lipid rafts
are assemblies of lipids and proteins that can agglomerate to form ordered structures.
These rafts are rich in cholesterol and sphingolipids and are considerably more viscous
than the rest of the cell membrane. It is believed that a number of important biomolecules
such as ion-channels or G-protein coupled receptors are localized within these rafts. We
have already commented on the possibility of the presence of different phases in the lipid
system used in our study. Since the ultimate goal of this project is to develop viable
biomimetic interfaces, incorporating rafts and studying their formation is a natural

progression of this research. In addition to more rigorous characterization of such

125

systems with FRAPP, studies using two-photon spectroscopy will allow us to probe

phase separation at interfaces”.

126

SOs-Na+

 

 

PDAC SPS

 

 

PAH

NH3+ NH3+

 

 

 

 

0
Figure 4.1: Structures for a) common polyelectrolytes and b) m d-PEG.

127

Blank Stamp coated with
PDAC/PAH

 

   

Stamping PDAC/PAH on
Ms

PDA
C SPS

 

 

 

 

 

Arrays of Lipid bilayers

 

Scheme 1

Figure 4.2: Illustration showing creation of arrays of lipid bilayers on PEMs
with PDAC or PAH as the topmost layer.

128

PDAC m-dPEG acid

 

Lipid bilayer

 

Scheme 2

Figure 4.3: Illustration showing formation of lipid bilayers on PEMs with m-dPEG
acid as the uppermost layer.

129

5x beam expander

l Ar-ion laser E D

 

O ptical flats

sample

I] G

objective Filter
block detector

 

 

 

Figure 4.4: Experimental set-up for Fluorescence recovery after pattern photobleaching
using EPI-illumination.

130

 

Figure 4.5: a) Fringe pattern in illuminated region obtained using 100 lines per inch
ruling. b) Observation area restricted by placing an aperture in the camera/PMT image
plane.

131

140000
120000 -

 

C: PDAC

A: PDAC

    

f: B: SPS

._ - -- VERSPS

 

 

. tutti i

 

T I I

 

I

0 500 1000 1500 2000 2500 3000 3500 4000
Tlme(eeconds)

Figure 4.6: Adsorption curves of (A) liposomes (10%DOPA, 90% DOPC) on PDAC. (B)
Liposomes (10%DOPA, 90% DOPC) on SPS. (C) Liposomes (20%DOPA, 80% DOPC)
on PDAC. (D) Liposomes (20%DOPA, 80% DOPC) on SPS.

132

 

==s=i=s=s=
thmmw
[III]

 

P
d
I
n

Relative ﬂuorescence (AU)

 

 

0 1000 2000 3000 4000
Time (se condsj

Figure 4.7: Shear induced desorption of liposomes from PDAC/SPS PEMs with SPS as
the topmost layer. Binding of liposomes to these surfaces is relatively loose and a large
fraction is removed upon introduction of buffer solution at a ﬂow rate of 0.34 mllmin.

133

 

Fluorescence (A.U)

 

 

 

 

 

Time (seconds)

Figure 4.8: Adsorption curves of liposomes (10%DOPA, 90% DOPC) on glass slide
coated with PEMs with PDAC (upper curve) and m—dPEG acid (lower curve) being the
topmost layer. As can be seen, the lower ﬂuorescence intensitites obtained are indicative
of the ability of PEG to resist liposome adsorption.

134

 

 

 

 

 

 

 

 

25000
‘5 20000 — a
s.
3 15000 -
C
8
a 10000 ~
2
8
r; 5000 - buffer
wash
0 ' I I r
0 1000 2000 3000 4000 5000

Time (seconds)

."
N

 

 

 

 

 

 

Normalized fluorescence

 

 

PPPP
ONbOO-ﬁ
llll

 

I r l

1000 1500 2000
Time (seconds)

8

Figure 4.9: a) Adsorption of liposomes (10%DOPA, 90% DOPC) on a glass slide coated
with PEMs with m-dPEG acid being the topmost layer. b) Buffer-wash experiments to
study liposome desorption from PEMs. The top and bottom curves depict desorption of
liposomes from PEMs with PDAC and m-dPEG as the top layer, respectively. At t=0,
adsorption of liposomes (which have adsorbed for at least 45 rrrinutes) is halted by
introducing liposome-free buffer. In each curve, the ﬂuorescence intensity has been
normalized by the corresponding ﬂuorescence value obtained prior to initiation of the
buffer wash in each case.

135

 

(a) (b)

    

n
(C)

Figure 4.10: F luoresence images showing (a) line patterns on a PDAC patterned
substrate (b) circular patterns on a PDAC patterned substrate (c) line patterns on a PAH
patterned substrate.

136

   

(a) (b)

Figure 4.11: Fluorescence images showing (a) line patterns on a m-dPEG acid patterned
substrate (b) circular patterns on a m-dPEG patterned substrate.

137

140000

 

 

 

 

 

 

 

 

120000 3"- ‘— pro-bleach ﬂuorescence

illlllllll .
r . El
8 810001 .
g 600004 W
5‘ 40000. . '

20000 .

o . T I I I
0 200 400 600 Bill] 1000
Time (seconds)

14411110

124000 '9 4—— pro-bleach ﬂuorescence
E 104000 J. E]
8 84100 -
8
3 64000 if i... D”... .g.“
H- 44000 -

24000 - . °

e 4-
4m l I . I T I I I
0 200 400 600 110010001200 14001”

Time (seconds)

Figure 4.12: a) TIRF-FRAP data for DOPA/DOPC liposomes adsorbed on PEMs
with PDAC as the top layer. b) TIRF-FRAP data under conditions of intermittent
monitoring to prevent monitoring beam induced photobleaching over longer time

scales.

138

 

 

 

 

 

ﬂuorescence
3
f

 

 

 

 

 

Figure 4.13: Fluoresence recovery after pattern photobleaching (EPI-FRAPP) proﬁles
on PDAC. The solid line in the plate A represent ﬁts to the recovery data set with a
model159 that describes the sample as containing containing a single mobile and
immobile fraction, while in plate B the solid line is the ﬁt to a model describing a
sample with two mobile populations (with different mobilities) and an immobile
fraction. Average values obtained with these models are summarized in Table 4.1 and
Table 4.2. Also shown below each ﬁt is a plot of the residuals vs. time as an indication
of the goodness of ﬁt.

139

0.45..
0.40;
035:
030;
025:
020;

Fluorescence

0.15:

 

 

0.10

 

0.04. Time (seconds)
0.02: 1
0'”. a ll: -_|. .’.' ‘ ' _‘ f. :l .' H ‘ 1 V» I lit 1 lll 'L.
0.02.: ‘ ' ' '
0.04:

 

 

 

0.45;
0.401.
0.35
0.30:
0.25:
020.

Fluorescence

0.15-
0.10

 

 

 

0.047 Tlme(eeconde)
0.02.. .
0'00. "‘ ' “ ' m 0' V'u"f‘;..
0.02.

0.04} . .

 

 

 

100'150'260

Figure 4.14: Fluoresence recovery after pattern photobleaching (EPI—FRAPP) proﬁles on
PAH. The solid line in the plate A represent ﬁts to the recovery data set with a model159
that describes the sample as containing containing a single mobile and immobile fraction,
while in plate B the solid line is the ﬁt to a model describing a sample with two mobile
populations (with different mobilities) and an immobile fraction. Average values
obtained with these models are summarized in Table 4.3 and Table 4.4. Also shown
below each ﬁt is a plot of the residuals vs. time as an indication of the goodness of ﬁt.

140

Table 4.1: Summary of model parameters obtained from ﬁt of BLM recovery on
PDAC to Equation (4.6) (single mobile-species model)

 

 

 

 

parameter 1 2 3 Average
f0 0.187 0.31 0.325
PDAC m 0.363 0.096 0.201 0.22:0.13
D .
0.062 0.108 0.0465 007210.03
(108 cmzls)

 

 

 

 

 

 

 

 

Table 4.2: Fit parameters for lipid bilayer formed on substrates topped with
PDAC, using the two mobile-species model given by (4.7). f0 represents the
unbleached fraction (at t=0), and m1, m2 and D1, D2 are the mobile fractions and
their corresponding diffusion coefﬁcients respectively.

 

Substrate

PDAC

 

 

 

 

 

 

 

 

 

 

Parameter 1 2 3
r0 0.14611 0.28191 0.28546
m1 0.18469 0.10216 0.13883
m2 0.35847 0.07429 0.2276
D1 (108cmzls) 0.44962 1.86424 1.01905
02((10‘cm2/s) 0.02094 0.06672 0.01802
F-statistic 210 13.16 46.37

 

 

 

141

Table 4.3: Summary of model parameters obtained from ﬁt of BLM recovery on PAH
to Equation (4.6) (single mobile-species model)

 

PAH

 

 

 

 

 

parameter 1 2 3 4 Average
to 0.15213 0.16429 0.28392 0.25053
m 0.63635 0.66456 0.65787 0.74132 0.67:0.05
D
0.0610009
(108 cmzls) 0.06066 0.06491 0.05634 0.0444

 

 

 

 

 

 

Table 4.4: Summary of model parameters obtained from ﬁt of BLM recovery on PAH
to Equation (4.7) (two mobile-species model).

 

Substrate

PAH

 

 

 

 

 

 

 

 

 

 

 

 

 

Parameter 1 2 3 4
10 0.11526 0.14067 0.24257 0.21486
ml 0.13961 0.09044 0.13076 0.16178
m2 0.58087 0.6402 0.62934 0.72163
I)l (108cm2/s) 0.75737 0.88423 2.25164 0.47875
Dz((1080m2/s) 0.04773 0.05385 0.04789 0.02906
F-statistic 114.3622 7.71 6.49 71

 

142

 

5 APPENDICES
5.1 Appendix A: Labview ﬂowsheets

5.1.1 Labview ﬂowsheets for TIRFM experiments conducted in analog mode

-nter ﬁlename:
-pen or create

       
  
 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

  
 

 

 

1°00 >buffer size; I continuous
' >scan rate>
. can rate <number of chans< M.
scans/ sec) I \ e.
[IE] ll]
# of points
2% , 1|
[“"l— im— "' ;>-_HL:J__W ail
umber of samples to average for each data . oint Hi! J
' 'II
‘ >error cluster> I
.......... ‘1 —
t 1

. n u)
- ointsf sec

1:» rim

 

 

E

i ec/polnt

- tartin time
i

m;

 

 

 

 

 

4.
V

 

 
      
 

Figure 5.1: Labview ﬂow sheets for TIRF data acquisition in analog mode. The
ﬂow sheet is continued on the next page.

143

 

 

 

     
 
  
 
  

continuous ac-j
>scan rate> ——‘

 

<error duster!
< taskID >

 

 

 

 

 

'r-alse "
lolof

E

_m

 

 

 

 

 

 

 

 

 

 

Figure 5.1 (continued): This flow sheet has been continued from the
previous page

5.1.2 Labview ﬂow sheets for photon counting

 

 

 

0 0..2

   

CONFIGURE VI. ASSIGN INPUT 1 TO COUNTER 'A',

10 MHZ TO COUNT

.; ’31- (.11., II“I‘§IMI‘I‘,LI .I;.e‘];l.I1,e,fl,,ll;1.n‘ I‘llI:I«’l‘ I‘E‘lnlﬂilr.:l_.Il'iliﬁlﬂlmlihg-

 

 

Figure 5.2: Flow sheets for computer interfacing with SR400 photon counter. These ﬂow
sheets have been continued on the succeeding pages. These ﬂowsheets have been derived
from Labview templates obtained from National Instruments, Austin, TX

145

   

mmmﬁvm m7...

 
 

 

 

in: lei...- :IEIEI. led-ﬂglililw Inn-u. I: In I... aw... h I I game! .2

           
    

 

.mOZOUmm 2H m2? ._.Zm3<>50w
O._. Emh .._>>U._ OF mDOZ Em .3 WED .4.

 

 

.,. .,. . ._..,. .,....,. ... "PE-in. . . . ..,. .,. . . .2. ....,

Figure 5.2(continued): This ﬁgure has been continued from the previous page

and depicts subroutines for operating the SR400 photon counter

146

 

1......" ....a.......1... . .................:.;l ...0............_.._........ l. . I. ........._ . ..
: . .. . . L. . . _. . . .. . .

      

 

 

 

   

 

. :E

_u 0%.. ._u

1‘

 

 

 

Figure 5.2 (continued): This flow sheet has been continued from the previous

page and depicts subroutines for operating the SR400 photon-counter

147

5.2 Appendix B: Matlab programs for filtering and averaging analog

measurements

5.2.1 Filtering program

This Matlab program discriminates the signal from the background

function p = test(data)

delete 'r:\colloid-surface—science\output.txt'
max_va1 = 0;

min_val data(1,1);

for i=1:length(data)
if (data(i,1)>max_val)
max_val=data(i,l);
end
if (data(i,1)<min_val)
min_val=data(i,l);
end
end

threshold = (0.72) * (max‘val—min_val) + min_val;
counter=1;

finaldata(l, l) = 0;

finaldata(1, 2) = 0;

for i=1:length(data);
if (data(i,1)>threshold)
finaldata(counter, l)
finaldata(counter, 2)
counter = counter+1;
end
end

data(i,1);
data(i,2);

p=fina1data;
%mean and std. dev.%

m=mean(p,1)
average=m(1)
s=std(p,1)
stdev=s(1)

counter=l;
endgame(1, 1) 0;
endgame(1, 2) = 0;

for i=1:1ength(p);
diff=p(i,1)—average;
absolute=abs(diff);
if (absolute<stdev)
endgame(counter, l) p(i,1);
endgame(counter, 2) = p(i,2);

148

counter = counter+l;
end
end
save 'r:\colloid-surface-science\output.txt‘ endgame -ASCII;
p=endgame;

5.2.2 Averaging Script

Matlab program for averaging TIRFM ﬂuorescence data obtained in analog mode

c=1,
b=l;
g=size(A,1)/8;
p=floor(g)+l
while b<p

sum=0;

d=C+7;

for j=c:d

sum=sum+A(j);

end

avg(l,b)=sum/8;

c=c+8;

b=b+l;
end;
n=avg';
delete 'r:\colloid—surface-science\my_data.txt‘
save my_data.txt n -ASCII

149

5.3 Appendix C: Protocol for estimation of spot diameters for TIRF-

photobleaching

Estimation of bleach spot diameters was determined using the following

procedures.

Calibration of the actual speed of the stage translation: The speed of
translation of the stage in the X-coordinate was determined as follows: Two scratch
marks were made at a precise distance from each other in a microscope slide. The
objective was then centered and focused on one of the marks. The motion of the stage
was then actuated and a stopwatch and reticule in the eyepiece was used to determine the
amount of time required for the second scratch to appear in the ﬁeld of view. Using this
procedure, a stage speed calibration chart was generated for different motor settings.
This procedure was repeated for speed calibrations in the Y- coordinate. These charts are

shown in Figure 5.3.

Determination of beam waist profile: After adsorption of ﬂuorescently labeled
liposomes was completed and the ﬂow cell was rinsed with buffer, a bleach pulse was
applied. Translation of the stage in the X and Y directions across the spot was then
performed and the ﬂuorescence signal was monitored. The bleached area in each scan
was detected as a momentary drop in the ﬂuorescence. Figure 5.4 shows the ﬂuorescence
collected during such a scan. Since the speed of translation of the stage is known and the

duration of the ﬂuorescence drop can be measured, the spot diameters can be easily

150

calculated. This procedure provides a reasonable accurate estimate of the elliptical spot

dimensions.

For these methods, the assumption is made that the lateral diffusivities of the
lipids in the bilayer, are considerably slower than the duration of time taken for the stage
to traverse the length of the dimension of the bleach spot. This was conﬁrmed in TIRF-
FRAP experiments where essentially no recovery was observed after photobleaching for

bilayers formed on PDAC substrates.

151

stage speed (mm/sec)

stage speed (mm/sec)

 

 

 

 

 

 

 

 

 

 

 

 

10 4 "‘°
8 u
6 - . [. X-translatlon
4 -e
2 i
0 . . .
0 400 600 800
MAC 2000 speed (A.U)
6
5 . °
4 ‘ e
3 - [e Y-transletion"
e
2 -
D
1 _
0 0 . .
0 400 600 800
MAC 2000 speed (A.U)

Figure 5.3: Stage speed calibrations displaying speed (in arbitrary units) against actual
speed for both X- and Y motion controllers.

152

25000
23000 J
21 000 J

17000 . ‘ °
60‘”
150001

 

3

Fluorescence (A.U)

§

11000 -

 

 

..---------.:..e-----------

O
S.

E

 

15.5 16 16.5 17 17.5 18 18.5
Time (seconds)

Figure 5.4: Estimating the dimensions of the elliptical spot. This picture shows a typical
scan in the x-coordinate. A spot is bleached and the specimen is translated in the x-
direction across the bleached spot. The data between the dotted lines represents the
duration during the scan when the objective collects light from the bleached area of the
sample.

153

5.4 APPENDIX D: Quartz crystal microbalance studies

 

E o—ﬁ i 0
>

a u

a E

3- 8'

E 1:

5 3.. g 52

'0’ e

a 1 1 b

 

ll] 20 30
Time (min)

(b)

10 20 30
Time (min)

(a)

Figure 5.5: Adsorption of liposomes as studied by QCM on (a) PDAC (b) PAH. Note the
presence of two distinct phases for adsorption on PAH (suggesting vesicle adsorption and
rupture), in contrast with the presence of one phase for PDAC (suggesting vesicle

adsorption without rupture).

154

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

LIST OF REFERENCES

Volkov, A. G. "Redox reactions at liquid hydrocarbon/water interfaces:
Biophysical aspects", Electrochimica Acta 1998, 44, 139-153.

Arai, K.; Ohsawa, M.; Kusu, F.; Takamura, K. "Drug Ion Transfer across on Oil-
Water Interface and Pharmacological Activity", Bioelectrochem. Bioenerg. 1993,
31, 65-76.

Mackie, A. R.; Gunning, A. P.; Wilde, P. J.; Morris, V. J. "Orogenic displacement
of protein from the oil/water interface", Langmuir 2000, 16, 2242-2247.

Bowers, J.; Zarbakhsh, A.; Webster, J. R. P.; Hutchings, L. R.; Richards, R. W.
"Neutron reﬂectivity studies at liquid-liquid interfaces: Methodology and
analysis", Langmuir 2001, 1 7, 140-145.

Gajraj, A.; Ofoli, R. Y. "Quantitative technique for investigating macromolecular
adsorption and interactions at the liquid-liquid interface", Langmuir 2000, 16,
4279-4285.

Gajraj, A.; Ofoli, R. Y. "Effect of extrinsic ﬂuorescent labels on diffusion and
adsorption kinetics of proteins at the liquid-liquid interface", Langmuir 2000, 16,
8085-8094.

Naujok, R. R.; Paul, H. J .; Corn, R. M. "Optical second harmonic generation
studies of azobenzene surfactant adsorption and photochemistry at the water/1,2-
dichloroethane interface", J Phys Chem-Us 1996, 100, 10497-10507.

Sauer, B. B.; Chen, Y. L.; Zograﬁ, 6.; Yu, H. "Static and Dynamic Interfacial-
Tensions of a Phospholipid Monolayer at the Oil-Water Interface -
Dipalmitoylphosphatidylcholine at Heptane Water", Langmuir 1986, 2, 683-685.

155

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

Graham, D. E.; Chatergoon, L.; Phillips, M. C. "Technique for Measuring
Interfacial Concentrations of Surfactants at Oil-Water Interface", J Phys E Sci
Instrum 1975, 8, 696-699.

Ando, J .; Albelda, S. M.; Levine, E. M. "Culture of Human Adult Endothelial-
Cells on Liquid-Liquid Interfaces - a New Approach to the Study of Cell-Matrix
Interactions", In Vitro Cellular & Developmental Biology 1991, 27, 525-532.

Han, M.; Sethuraman, A.; Kane, R. 8.; Belfort, G. "Nanometer-scale roughness
having little effect on the amount or structure of adsorbed protein", Langmuir

2003, 19, 9868-9872.

Wertz, C. F.; Santore, M. M. "Adsorption and relaxation kinetics of albumin and
ﬁbrinogen on hydrophobic surfaces: Single-species and competitive behavior",
Langmuir 1999, 15, 8884-8894.

Calonder, C.; Tie, Y.; Van Tassel, P. R. "History dependence of protein
adsorption kinetics", Proc. Natl. Acad. Sci. U. S. A. 2001, 98, 10664-10669.

Hlady, V.; Reinecke, D. R.; Andrade, J. D. "Fluorescence of Adsorbed Protein
Layers .1. Quantitation of Total Intemal-Reﬂection Fluorescence", J. Colloid
Interface Sci. 1986, 111, 555-569.

Sengupta, T.; Damodaran, S. "A new methodology for studying protein
adsorption at oil-water interfaces", J. Colloid Interface Sci. 1998, 206, 407-415.

Zimmermann, R. M.; Schmidt, C. F.; Gaub, H. E. "Absolute Quantities and
Equilibrium Kinetics of Macromolecular Adsorption Measured by Fluorescence
Photobleaching in Total Internal-Reflection", J. Colloid Interface Sci. 1990, 139,
268-280.

Cornell, B. A.; BraachMaksvytis, V. L. B.; King, L. G.; Osman, P. D. J .; Raguse,
B.; Wieczorek, L.; Pace, R. J. "A biosensor that uses ion-channel switches",
Nature 1997, 387, 580-583.

Palmisano, F.; Centonze, D.; Guerrieri, A.; Zambonin, P. G. "An Interference-
Free Biosensor Based on Glucose-Oxidase Electrochemically Immobilized in a
Nonconducting Poly(Pyrrole) Film for Continuous Subcutaneous Monitoring of
Glucose through Microdialysis Sampling", Biosens. Bioelectron. 1993, 8, 393-
399.

156

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

Ikeda, T.; Matsushita, F.; Senda, M. "Amperometric Fructose Sensor Based on
Direct Bioelectrocatalysis", Biosens. Bioelectron. 1991, 6, 299-304.

Chew, T. L.; Wolf, W. A.; Gallagher, P. J .; Matsumura, F.; Chisholm, R. L. "A
ﬂuorescent resonant energy transfer-based biosensor reveals transient and

regional myosin light chain kinase activation in lamella and cleavage furrows", J.
Cell Biol. 2002, 156, 543-553.

Wang, B. Q.; Li, B.; Deng, Q.; Dong, S. J. "Amperometric glucose biosensor
based on sol-gel organic-inorganic hybrid material", Analytical Chemistry 1998,
70, 3170-3174.

Raguse, B.; Braach-Maksvytis, V.; Cornell, B. A.; King, L. G.; Osman, P. D. J .;
Pace, R. J .; Wieczorek, L. "Tethered lipid bilayer membranes: Formation and
ionic reservoir characterization", Langmuir 1998, 14, 648-659.

Sackmann, E.; Tanaka, M. "Supported membranes on soft polymer cushions:
fabrication, characterization and applications", Trends in Biotechnology 2000, 18,
58—64.

Cunningham, J. J.; Nikolovski, J .; Linderrnan, J. J .; Mooney, D. J. "Quantiﬁcation
of ﬁbronectin adsorption to silicone-rubber cell culture substrates", Biotechniques
2002, 32, 876-887.

McFarland, C. D.; Thomas, C. H.; DeFilippis, C.; Steele, J. G.; Healy, K. E.
"Protein adsorption and cell attachment to patterned surfaces", J. Biomed. Mater.
Res. 2000, 49, 200-210.

Whittle, J. D.; Bullett, N. A.; Short, R. D.; Douglas, C. W. I.; Hollander, A. P.;
Davies, J. "Adsorption of Vitronectin, collagen and immunoglobulin-G to plasma

polymer surfaces by enzyme linked immunosorbent assay (ELISA)", J. Mater.
Chem. 2002, 12, 2726-2732.

MacDonald, D. E.; Markovic, B.; Allen, M.; Somasundaran, P.; Boskey, A. L.
"Surface analysis of human plasma ﬁbronectin adsorbed to cormnercially pure
titanium materials", J. Biomed. Mater. Res. 1998, 41, 120-130.

Hynes, R. F ibronectins; Springer: New York, 1990.

157

(29)

(30)

(31)

(32)

(33)

(34)

(35)

(36)

(37)

(38)

Williams, E. C.; Janmey, P. A.; Ferry, J. D.; Mosher, D. F. "Conformational
States of Fibronectin - Effects of Ph, Ionic- Strength, and Collagen Binding", J.
Biol. Chem. 1982, 25 7, 4973-4978.

Koteliansky, V. E.; Bejanian, M. V.; Smimov, V. N. "Electron-Microscopy Study
of Fibronectin Structure", FEBS Lett. 1980, 120, 283-286.

Koteliansky, V. E.; Glukhova, M. A.; Bejanian, M. V.; Smimov, V. N.;
Filimonov, V. V.; Zalite, O. M.; Venyaminov, S. Y. "A Study of the Structure of
Fibronectin", Eur. J. Biochem. 1981, 119, 619-624.

Engel, J .; Odermatt, E.; Engel, A.; Madri, J. A.; Furthmayr, H.; Rohde, H.; Timpl,
R. "Shapes, Domain Organizations and Flexibility of Laminin and Fibronectin, 2
Multifunctional Proteins of the Extracellular- Matrix", J. Mol. Biol. 1981, 150,
97-120.

Erickson, H. P.; Carrel], N .; McDonagh, J. "Fibronectin Molecule Visualized in
Electron-Microscopy - a Long, Thin, Flexible Strand", J. Cell Biol. 1981, 91, 673-
678.

Paynter, R. W.; Ratner, B. D.; Horbett, T. A.; Thomas, H. R. "Xps Studies on the
Organization of Adsorbed Protein Films on Fluoropolymers", J. Colloid Interface
Sci. 1984, 101, 233-245.

Lai, C. S.; Wolff, C. 13.; Novello, D.; Griffone, L.; Cuniberti, C.; Molina, F.;
Rocco, M. "Solution Structure of Human Plasma Fibronectin under Different
Solvent Conditions - Fluorescence Energy-Transfer, Circular- Dichroism and

Light-Scattering-Studies", J. Mol. Biol. 1993, 230, 625-640.

Wolff, C.; Lai, C. S. "Fluorescence Energy-Transfer Detects Changes in
Fibronectin Structure Upon Surface Binding", Arch. Biochem. Biophys. 1989,
268, 536-545.

Bergkvist, M.; Carlsson, J .; Oscarsson, S. "Surface-dependent conformations of
human plasma ﬁbronectin adsorbed to silica, mica, and hydrophobic surfaces,
studied with use of Atomic Force Microscopy", Journal of Biomedical Materials
Research PartA 2003, 64A, 349-356.

Sagvolden, G.; Giaever, I.; Pettersen, E. 0.; Feder, J. "Cell adhesion force
microscopy", Proc. Natl. Acad. Sci. U. S. A. 1999, 96, 471-476.

158

(39)

(40)

(41)

(42)

(43)

(44)

(45)

(46)

(47)

(48)

(49)

Jonsson, U.; Ivarsson, B.; Lundstrom, I.; Berghem, L. "Adsorption Behavior of
Fibronectin on Well-Characterized Silica Surfaces", J. Colloid Interface Sci.
1982, 90, 148-163.

Cheng, Y. L.; Darst, S. A.; Robertson, C. R. "Bovine Serum-Albumin Adsorption
and Desorption Rates on Solid— Surfaces with Varying Surface-Properties", J.
Colloid Interface Sci. 1987, 118, 212-223.

Lassen, B.; Malmsten, M. "Competitive protein adsorption studied with TIRF and
ellipsometry", J. Colloid Interface Sci. 1996, 179, 470-477.

Sengupta, T.; Damodaran, 8. "Role of dispersion interactions in the adsorption of
proteins at oil-water and air-water interfaces", Langmuir 1998, 14, 6457-6469.

Kim, D. T.; Blanch, H. W.; Radke, C. J. "Direct imaging of lysozyme adsorption
onto mica by atomic force microscopy", Langmuir 2002, 18, 5841-5850.

Beverung, C. J.; Radke, C. J .; Blanch, H. W. "Protein adsorption at the oil/water
interface: characterization of adsorption kinetics by dynamic interfacial tension
measurements", Biophys. Chem. 1999, 81, 59-80.

Dickinson, E. "Adsorbed protein layers at ﬂuid interfaces: interactions, structure
and surface rheology", Colloid Surf. B-Biointerfaces 1999, 15, 161-176.

Giaever, 1.; Keese, C. R. "Behavior of Cells at Fluid Interfaces", Proceedings of
the National Academy of Sciences of the United States of America-Biological
Sciences 1983, 80, 219-222.

Keese, C. R.; Giaever, I. "Cell-Growth on Liquid Interfaces - Role of Surface-
Active Compounds", Proceedings of the National Academy of Sciences of the
United States of America-Biological Sciences 1983, 80, 5622-5626.

Koenig, A. L.; Gambillara, V.; Grainger, D. W. "Correlating ﬁbronectin
adsorption with endothelial cell adhesion and signaling on polymer substrates",
Journal of Biomedical Materials Research Part A 2003, 64A, 20-37.

Technical Documentation,Chemicon International.

159

(50)

(51)

(52)

(53)

(54)

(55)

(56)

(57)

(58)

(59)

(60)

"Arnine-Reactive Probes", Molecular Probes, Eugene, OR 2003.

"Albumin product information sheet", Sigma-Aldrich, St. Louis MO.

Jeon, S. 1.; Lee, J. H.; Andrade, J. D.; Degennes, P. G. "Protein Surface
Interactions in the Presence of Polyethylene Oxide .1. Simpliﬁed Theory", J.
Colloid Interface Sci. 1991, 142, 149-158.

Jeon, S. I.; Andrade, J. D. "Protein Surface Interactions in the Presence of
Polyethylene Oxide .2. Effect of Protein Size", J. Colloid Interface Sci. 1991, 142,
159-166.

Andrade, J. D.; Hlady, V.; Icon, 8. I. In Hydrophilic Polymers, 1996; Vol. 248, pp
51-59.

Gajraj, A. "A novel total internal reﬂection ﬂuorescence ('I‘HZF) technique for
investigating macromolecular adsorption and interactions at the liquid—liquid
interface", PhD thesis, Michigan State University 1999.

Lok, B. K.; Cheng, Y. L.; Robertson, C. R. "Protein Adsorption on Crosslinked
Polydimethylsiloxane Using Total Intemal-Reﬂection Fluorescence", J. Colloid
Interface Sci. 1983, 91 , 104-116.

Johnson, K. J .; Sage, H.; Briscoe, G.; Erickson, H. P. "The compact conformation
of ﬂbronectin is determined by intramolecular ionic interactions", J. Biol. Chem.
1999, 274, 15473-15479.

Michael, K. E.; Vemekar, V. N.; Keselowsky, B. G.; Carson Meredith, J .; Latour,
R. A.; Garcy’, A. J. "Adsorption-induced conformational changes in Fibronectin
due to interactions with well-deﬁned surface chemistries", Langmuir 2003, 19,

8033-8040.

Malmsten, M.; Lassen, B. "Competitive Adsorption at Hydrophobic Surfaces
from Binary Protein Systems", Journal of Colloid and Interface Science 1994,
166, 490-498.

2nd edition ed.; Putnam, F. W., Ed.; Academic Press, New York, 1975; Vol. 1, pp
133-181.

160

(61)

(62)

(63)

(64)

(65)

(66)

(67)

(68)

(69)

(70)

(71)

Green, R. J.; Davies, M. C.; Roberts, C. J.; Tendler, S. J. B. "Competitive protein

adsorption as observed by surface plasmon resonance", Biomaterials 1999, 20,
385-391.

Vroman, L.; Adams, A. In Proteins at Interfaces: Physiochemical and
Biochemical Studies, 1987, pp 154-164.

Fang, F.; Szleifer, 1. "Kinetics and thermodynamics of protein adsorption: A
generalized molecular theoretical approach", Biophys. J. 2001, 80, 2568-2589.

Sanocki VA "Cell culture at the liquid-liquid interface: effects of ﬂuorocarbon
layer thickness on cell growth", M.S. thesis, Michigan State University 2004.

Zhang, H. L.; Bremmell, K.; Kumar, 8.; Smart, R. S. C. "Vitronectin adsorption
on surfaces visualized by tapping mode atomic force microscopy", Journal of
Biomedical Materials Research Part A 2004, 68A, 479-488.

Voet, D.; Voet, J. G. Biochemistry; 2nd ed.; J. Wiley & Sons: New York, 1995.

Green, R. J.; Su, T. J.; Lu, J. R.; Webster, J.; Penfold, J. "Competitive adsorption
of lysozyme and C12E5 at the air/liquid interface", Phys. Chem. Chem. Phys.
2000, 2, 5222-5229.

Robeson, J. L.; Tilton, R. D. "Spontaneous reconﬁguration of adsorbed lysozyme
layers observed by total internal reﬂection ﬂuorescence with a pH-sensitive
ﬂuorophore", Langmuir 1996, 12, 6104-6113.

Choi, E. J.; Foster, M. D.; Daly, S.; Tilton, R.; Przybycien, T.; Majkrzak, C. F.;
Witte, P.; Menzel, H. "Effect of ﬂow on human serum albumin adsorption to self-

assembled monolayers of varying packing density", Langmuir 2003, 19, 5464-
5474.

Fragneto, G.; Su, T. J.; Lu, J. R.; Thomas, R. K.; Rennie, A. R. "Adsorption of
proteins from aqueous solutions on hydrophobic surfaces studied by neutron
reﬂection", Phys. Chem. Chem. Phys. 2000, 2, 5214-5221.

Marsh, R. J .; Jones, R. A. L.; Sferrazza, M. "Adsorption and displacement of a
globular protein on hydrophilic and hydrophobic surfaces", Colloid Surf B-
Biointerfaces 2002, 23, 31—42.

161

(72)

(73)

(74)

(75)

(76)

(77)

(78)

(79)

(80)

(31)

(82)

Rodahl, M.; Hook, F.; Krozer, A.; Brzezinski, P.; Kasemo, B. "Quartz-Crystal
Microbalance Setup for Frequency and Q-Factor Measurements in Gaseous and
Liquid Environments", Rev Sci Instrum 1995, 66, 3924-3930.

Kidoaki, S.; Matsuda, T. "Mechanistic aspects of protein/material interactions
probed by atomic force microscopy", Colloid Surf B-Biointerfaces 2002, 23, 153-
163.

Richter, R. P.; Brisson, A. "Characterization of lipid bilayers and protein
assemblies supported on rough surfaces by atomic force microscopy", Langmuir
2003, 19, 1632-1640.

Rebar, V. A.; Santore, M. M. "History-dependent isotherms and TIRF calibrations
for homopolymer adsorption", Macromolecules 1996, 29, 6262-6272.

Reichert, W. M.; Rockhold, S.; Van Wagenen, R. A.; Andrade, J. D. In Interfacial
Aspects of biomedical polymers; Plenum Press: New York, 1984; Vol. 2.

Crosby, G. A.; Demas, J. N .; Callis, J. B. "Absolute Quantum Efﬁciencies", J Res
Nbs C Eng Inst 1972, A 76, 561-577.

Magde, D.; Wong, R.; Seybold, P. G. "Fluorescence quantum yields and their
relation to lifetimes of rhodamine 6G and ﬂuorescein in nine solvents: Improved
absolute standards for quantum yields", Photochem Photobiol 2002, 75, 327-334.

Brannon, J. H.; Magde, D. "Absolute Quantum Yield Determination by Thermal
Blooming - Fluorescein", J Phys Chem-Us 1978, 82, 705-709.

Lakowicz, J. R. Principles of ﬂuorescence spectroscopy; Plenum Press: New
York, 1983.

Strickler, S. J .; Berg, R. A. "Relationship between Absorption Intensity and
Fluorescence lifetime of molecules", Journal of Chemical Physics 1962, 37, 814-
822.

Lakowicz, J. R. "Radiative decay engineering: Biophysical and biomedical
applications", Anal Biochem 2001, 298, 1-24.

162

(83)

(84)

(85)

(86)

(87)

(88)

(89)

(90)

(91)

(92)

(93)

Rao, A.; Dhinojwala, A. "Two-photon ﬂuorescence technique to study polymer
adsorption at the toluene-water interface", Langmuir 2003, 19, 8813-8817.

Denk, W.; Svoboda, K. "Photon upmanship: Why multiphoton imaging is more
than a gimmick", Neuron 1997, 18, 351-357.

Xu, C.; Webb, W. W. "Measurement of two-photon excitation cross sections of
molecular ﬂuorophores with data from 690 to 1050 nm", Journal of the Optical
Society of America B-Optical Physics 1996, 13, 481-491.

Dunaevsky, A.; Tashiro, A.; Majewska, A.; Mason, C.; Yuste, R. "Developmental
regulation of spine motility in the mammalian central nervous system", Proc.
Natl. Acad. Sci. U. S. A. 1999, 96, 13438-13443.

Miller, M. J .; Wei, S. H.; Parker, 1.; Cahalan, M. D. "Two-photon imaging of
lymphocyte motility and antigen response in intact lymph node", Science 2002,
296, 1869-1873.

Konig, K. "Multiphoton microscopy in life sciences", Journal of Microscopy-
Oxford 2000, 200, 83-104.

Dietrich, C.; Bagatolli, L. A.; Volovyk, Z. N .; Thompson, N. L.; Levi, M.;
Jacobson, K.; Gratton, E. "Lipid rafts reconstituted in model membranes",
Biophys. J. 2001, 80, 1417-1428.

Grinvald, A.; Steinberg, I. Z. "Analysis of Fluorescence Decay Kinetics by
Method of Least-Squares", Anal Biochem 1974, 59, 583-598.

Beaglehole, D.; Lawson, F.; Harper, G.; Hossain, M. "Tryptophan, tryptophan-
1eucine, and BSA adsorption at an oil-water interface", J. Colloid Interface Sci.
1997, 192, 266-268.

Butler, S. M.; Tracy, M. A.; Tilton, R. D. "Adsorption of serum albumin to thin
ﬁlms of poly(lactide-co-glycolide)", J Control Release 1999, 58, 335-347.

Lakowicz, J. R.; Malicka, J .; D'Auria, S.; Gryczynski, 1. "Release of the self-
quenching of ﬂuorescence near silver metallic surfaces", Anal Biochem 2003,
320, 13-20.

163

(94)

(95)

(96)

(97)

(98)

(99)

(100)

(101)

(102)

(103)

Kelepouris, L.; Blanchard, G. J. "Dynamics of 7-azatryptophan and tryptophan
derivatives in micellar media. The role of ionic charge and substituent structure",
J. Phys. Chem. B 2003, 107, 1079-1087.

Heysel, S.; Vogel, H.; Sanger, M.; Sigrist, H. "Covalent Attachment of
Functionalized Lipid Bilayers to Planar Wave-Guides for Measuring Protein-
Binding to Biomimetic Membranes", Protein Science 1995, 4, 2532-2544.

McConnell, H. M.; Watts, T. H.; Weis, R. M.; Brian, A. A. "Supported Planar
Membranes in Studies of Cell-Cell Recognition in the Irnmune-System",
Biochimica Et Biophysica Acta 1986, 864, 95-106.

Nikolelis, D. P.; Siontorou, C. G.; Krull, U. J .; Katrivanos, P. L. "Ammonium ion
minisensors from self-assembled bilayer lipid membranes using Gramicidin as an

ionophore. Modulation of ammonium selectivity by platelet-activating factor",
Analytical Chemistry 1996, 68, 1735-1741.

Seifert, K.; Fendler, K.; Bamberg, E. "Charge Transport by Ion Translocating
Membrane-Proteins on Solid Supported Membranes", Biophysical Journal 1993,
64, 384-391.

Stelzle, M.; Sackmann, E. "Sensitive Detection of Protein Adsorption to

Supported Lipid Bilayers by Frequency-Dependent Capacitance Measurements
and Microelectrophoresis", Biochimica Et Biophysica Acta 1989, 981, 135-142.

Favero, G.; D'Annibale, A.; Campanella, L.; Santucci, R.; Ferri, T. "Membrane

supported bilayer lipid membranes array: preparation, stability and ion-channel
insertion", Analytica Chimica Acta 2002, 460, 23-34.

Furuike, S.; I-Iirokawa, J.; Yamada, S.; Yamazaki, M. "Atomic force microscopy
studies of interaction of the 20S proteasome with supported lipid bilayers",
Biochimica Et Biophysica Acta-Biomembranes 2003, 1615, 1-6.

Glasmastar, K.; Larsson, C.; Hook, F.; Kasemo, B. "Protein adsorption on
supported phospholipid bilayers", Journal of Colloid and Interface Science 2002,
246, 40-47.

Graneli, A.; Rydstrom, J .; Kasemo, B.; Hook, F. "Formation of supported lipid
bilayer membranes on Si02 from proteoliposomes containing transmembrane
proteins", Langmuir 2003, 19, 842-850. '

164

(104)

(105)

(106)

(107)

(108)

(109)

(110)

(111)

(112)

(113)

(114)

Groves, J. T.; Mahal, L. K.; Bertozzi, C. R. "Control of cell adhesion and growth
with micropatterned supported lipid membranes", Langmuir 2001, 17, 5129-5133.

Lahiri, J.; Kalal, P.; Frutos, A. 6.; Jonas, S. T.; Schaefﬂer, R. "Method for
fabricating supported bilayer lipid membranes on gold", Langmuir 2000, 16,
7805-7810.

Ohlsson, P. A.; Tjamhage, T.; Herbai, E.; Lofas, S.; Puu, G. "Liposome and
Proteoliposome Fusion onto Solid Substrates, Studied Using Atomic-Force
Microscopy, Quartz-Crystal Microbalance and Surface-Plasmon Resonance -
Biological-Activities of Incorporated Components", Bioelectrochemistry and
Bioenergetics 1995, 38, 137-148.

Tien, H. T.; Barish, R. H.; Gu, L. Q.; Ottova, A. L. "Supported bilayer lipid
membranes as ion and molecular probes", Analytical Sciences 1998, 14, 3-18.

Tien, H. T.; Ottova, A. L. "Supported planar lipid bilayers (s-BLMs) as
electrochemical biosensors", Electrochimica Acta 1998, 43, 3587-3610.

Ariga, K.; Okahata, Y. "Polymerized Monolayers of Single-Chain, Double-Chain,
and Triple-Chain Silane Amphiphiles and Permeation Control through the

Monolayer-Immobilized Porous-Glass Plate in an Aqueous-Solution", J. Am.
Chem. Soc. 1989, 111, 5618-5622.

Fare, T. L. "Electrical Characterization of Dipalmitoylphosphatidylethanolamine
and Cadmium Stearate Films on Platinum Surfaces in Aqueous-Solutions",
Langmuir 1990, 6, 1172-1179.

Plant, A. L. "Self-Assembled Phospholipid Alkanethiol Biomimetic Bilayers on
Gold", Langmuir 1993, 9, 2764-2767.

Plant, A. L.; Gueguetchkeri, M.; Yap, W. "Supported Phospholipid/Alkanethiol
Biomimetic Membranes - Insulating Properties", Biophys. J. 1994, 67, 1126-1133.

Stenger, D. A.; Fare, T. L.; Cribbs, D. H.; Rustin, K. M. "Voltage Modulation of a
Gated Ion Channel Admittance in Platinum-Supported Lipid Bilayers",
Biosensors and Bioelectronics 1992, 7, 11-20.

Hoppe, W.; Lohmann, W. Biophysics; Springer-Verlag: Berlin, 1983.

165

(115)

(116)

(117)

(118)

(119)

(120)

(121)

(122)

(123)

(124)

(125)

Beddow, J. A.; Peterson, 1. R.; Heptinstall, J .; Walton, D. J. "Electrochemical
characterisation of the diffusion of a biomolecule through a hydrogel", Journal of
Electroanalytical Chemistry 2003, 544, 107-112.

Cornell, B. A.; Krishna, G.; Osman, P. D.; Pace, R. D.; Wieczorek, L. "Tethered-
bilayer lipid membranes as a support for membrane-active peptides", Biochemical
Society Transactions 2001, 29, 613-617.

Fortig, A.; Jordan, R.; Graf, K. Schiavon, G.; Purrucker, 0., Tanaka, M. "Solid-
supported biomimetic membranes with tailored lipopolymer tethers",
Macromolecular Symposia 2004, 210, 329- 338.

Jin, T.; Pennefather, R; Lee, P. I. "Lipobeads: A hydrogel anchored lipid vesicle
system", F ebs Letters 1996, 397, 70-74.

Krishna, G.; Schulte, J.; Cornell, B. A.; Pace, R.; Wieczorek, L.; Osman, P. D.
"Tethered bilayer membranes containing ionic reservoirs: The interfacial
capacitance", Langmuir 2001, 1 7, 4858-4866.

Krysinski, P.; Zebrowska, A.; Michota, A.; Bukowska, J .; Becucci, L.; Moncelli,
M. R. "Tethered mono- and bilayer lipid membranes on Au and Hg", Langmuir
2001, 17, 3852-3857.

Krysinski, P.; Zebrowska, A.; Palys, B.; Lotowski, Z. "Spectroscopic and
electrochemical studies of bilayer lipid membranes tethered to the surface of
gold", Journal of the Electrochemical Society 2002, I49, E189-E194.

Kugler, R.; Knoll, W. "Polyelectrolyte-supported lipid membranes",
Bioelectrochemistry 2002, 56, 175-178.

Melzak, K. A.; Gizeli, E. "A silicate gel for promoting deposition of lipid
bilayers", Journal of Colloid and Interface Science 2002, 246, 21-28.

Moncelli, M. R.; Becucci, L.; Schiller, S. M. "Tethered bilayer lipid membranes
self-assembled on mercury electrodes", Bioelectrochemistry 2004, 63, 161-167.

Munro, J. C.; Frank, C. W. "Adsorption of lipid-functionalized poly(ethylene
glycol) to gold surfaces as a cushion for polymer-supported lipid bilayers",
Langmuir 2004, 20, 3339-3349.

166

(126)

(127)

(128)

(129)

(130)

(131)

(132)

(133)

(134)

(135)

Naumann, R.; Schiller, S. M.; Giess, F.; Grohe, B.; Hartman, K. B.; Karcher, I.;
Koper, I.; Lubben, J .; Vasilev, K.; Knoll, W. "Tethered lipid Bilayers on ultraﬂat
gold surfaces", Langmuir 2003, 19, 5435-5443.

Naumann, R.; Walz, D.; Schiller, S. M.; Knoll, W. "Kinetics of valinomycin-
mediated K+ ion transport through tethered bilayer lipid membranes", Journal of
Electroanalytical Chemistry 2003, 550, 241-252.

Proux-Delrouyre, V.; Elie, C.; Laval, J. M.; Moiroux, J.; Bourdillon, C.
"Formation of tethered and streptavidin-supported lipid bilayers on a microporous

electrode for the reconstitution of membranes of large surface area", Langmuir
2002, 18, 3263-3272.

Sinner, E. K.; Knoll, W. "Functional tethered membranes", Current Opinion in
Chemical Biology 2001, 5, 705-711.

Wagner, M. L.; Tamm, L. K. "Tethered polymer-supported planar lipid bilayers
for reconstitution of integral membrane proteins: Silane-polyethyleneglycol-lipid
as a cushion and covalent linker", Biophysical Journal 2000, 79, 1400-1414.

Cassier, T.; Sinner, A.; Offenhauser, A.; Mohwald, H. "Homogeneity, electrical
resistivity and lateral diffusion of lipid bilayers coupled to polyelectrolyte
multilayers", Colloids and Biointerfaces 1999, 15, 215-225.

Perez-Salas, U. A.; Faucher, K. M.; Majkrzak, C. F.; Berk, N. F.; Krueger, S.;
Chaikof, E. L. "Characterization of a biomimetic polymeric lipid bilayer by phase
sensitive neutron reﬂectometry", Langmuir 2003, 19, 7688-7694.

Moya, S.; Richter, W.; Leporatti, S.; Baumler, H.; Donath, E. "Freeze-Fracture
Electron Microscopy of Lipid Membranes on Colloidal Polyelectrolyte Multilayer
Coated Supports", Biomacromolecules 2003, 4, 808-814.

Tiourina, O. P.; Radtchenko, I.; Sukhorukov, S. B.; Mohwald, H. "Artiﬁcial cell
based on lipid hollow polyelectrolyte microcapsules: Channel reconstruction and

membrane potential measurement", Journal of Membrane Biology 2002, I90, 9-
16.

Hovis, J. S.; Boxer, S. G. "Patterning barriers to lateral diffusion in supported
lipid bilayer membranes by blotting and stamping", Langmuir 2000, 16, 894-897.

167

(136)

(137)

(138)

(139)

(140)

(141)

(142)

(143)

(144)

(145)

Kam, L.; Boxer, S. G. "Cell adhesion to protein-micropattemed—supported lipid
bilayer membranes", Journal of Biomedical Materials Research 2001, 55, 487-
495.

Kung, L. A.; Kam, L.; Hovis, J. 8.; Boxer, S. G. "Patterning hybrid surfaces of
proteins and supported lipid bilayers", Langmuir 2000, 16, 6773-6776.

Saccani, J .; Castano, S.; Desbat, B.; Blaudez, D. "A phospholipid bilayer
supported under a polymerized Langmuir ﬁlm", Biophysical Journal 2003, 85,
3781-3787.

Decher, G.; Hong, J. D. "Buildup of ultrathin multilayer ﬁlms by a self-assembly
process .2. Consecutive adsorption of anionic and cationic bipolar amphiphiles
and polyelectrolytes on charged surfaces", Ber Bunsen-Ges. 1991, 95, 1430-1434.

Kumar, A.; Whitesides, G. M. "Features of Gold Having Micrometer to
Centimeter Dimensions Can Be Formed through a Combination of Stamping with
an Elastomeric Stamp and an Alkanethiol Ink Followed by Chemical Etching",
Appl. Phys. Lett. 1993, 63, 2002-2004.

Jiang, X.; Hammond, P. T. "Selective deposition in layer-by-layer assembly:
Functional graft copolymers as molecular templates", Langmuir 2000, 16, 8501-
8509.

Jiang, X.; Zheng, H.; Shoshna, 6.; Hammond, P. T. "Polymer-on-polymer
stamping: Universal approaches to chemically patterned surfaces", Langmuir
2002, 18, 2607-2615.

Fou, A. C.; Rubner, M. F. "Molecular-Level Processing of Conjugated Polymers
.2. Layer-by-Layer Manipulation of in-Situ Polymerized P-Type Doped
Conducting Polymers", Macromolecules 1995, 28, 7115-7120.

Lee, 1.; Zheng, H.; Rubner, M. F.; Hammond, P. T. "Controlled cluster size in
patterned particle arrays via directed adsorption on conﬁned surfaces", Advanced
Materials 2002, 14, 572-577.

Zheng, H.; Lee, 1.; Rubner, M. F.; Hammond, P. T. "Two component particle
arrays on patterned polyelectrolyte multilayer templates", Adv. Mater 2002, 14,

. 569-572.

168

(146)

(147)

(148)

(149)

(150)

(151)

(152)

(153)

(154)

(155)

Lvov, Y.; Ariga, K.; Ichinose, 1.; Kunitake, T. "Assembly of Multicomponent
Protein Films by Means of Electrostatic Layer-by-Layer Adsorption", J. Am.
Chem. Soc. 1995, 117, 6117-6123.

Lee, 1.; Hammond, P. T.; Rubner, M. F. "Selective electroless nickel plating of
particle arrays on polyelectrolyte multilayers", Chemistry of Materials 2003, 15,
4583-4589. -

St. John, P. M.; Craighead, H. G. "Microcontact printing and pattern transfer
using trichlorosilanes on oxide substrates", Appl. Phys. Lett. 1996, 68, 1022-1024.

Xia, Y.; Kim, B; Mrksich, M.; Whitesides, G. M. "Microcontact printing of
alkanethiols on copper and its application in microfabrication", Chem. Mater.
1996, 8, 601-603.

Yang, X. M.; Tryk, D. A.; Hasimoto, K.; Fujishima, A. "Surface enhanced Raman
imaging of a patterned self-assembled monolayer formed by microcontact
printing on a silver ﬁlm", Appl. Phys. Lett. 1996, 69, 4020-4022.

Berg, M. C.; Yang, S. Y.; Hammond, P. T.; Rubner, M. F. "Controlling
mammalian cell interactions on patterned polyelectrolyte multilayer surfaces",
Langmuir 2004, 20, 1362-1368.

Kim, H.; Doh, J .; Irvine, D. J .; Cohen, R. E.; Hammond, P. T. "Large area two-
dimensional B cell arrays for sensing and cell-sorting applications",
Biomacromolecules 2004, 5, 822-827.

Yang, S. Y.; Mendelsohn, J. D.; Rubner, M. F. "New class of ultrathin, highly
cell-adhesion-resistant polyelectrolyte multilayers with micropatteming
capabilities", Biomacromolecules 2003, 4, 987-994.

Lee, 1.; Ahn, J .; Hendericks, T.; Rubner, M. F.; Hammond, P. T. "Patterned and
controlled polyelectrolyte fractal growth and aggregations", Langmuir 2004, 20,
2478-2483.

Kidambi, S.; Chan, C.; Lee, I. "Selective depositions on polyelectrolyte
multilayers: Self-assembled monolayers of m-dpeg acid as molecular template", J.
Am. Chem. Soc. 2004, 126, 4697-4703.

169

(156)

(157)

(158)

(159)

(160)

(161)

(162)

(163)

(164)

(165)

Robeson, J. L. "Application of total internal reﬂection ﬂuorescence to probe
surface diffusion and orientation of adsorbed protiens."1995

Starr, T. E.; Thompson, N. L. "Fluorescence pattern photobleaching recovery for
samples with multi—component diffusion", Biophys. Chem. 2002, 97, 2944.

Smith, B. A.; McConnell, H. M. "Determination of Molecular-Motion in
Membranes Using Periodic Pattern Photobleaching", Proc. Natl. Acad. Sci. U. S.
A. 1978, 75, 2759-2763.

Wright, L. L.; Palmer, A. G.; Thompson, N. L. "Inhomogeneous Translational
Diffusion of Monoclonal-Antibodies on Phospholipid Langmuir-Blodgett Films",
Biophys. J. 1988, 54, 463-470.

Ameloot, M.; Hendrickx, H. "Criteria for Model Evaluation in the Case of
Deconvolution Calculations", Journal of Chemical Physics 1982, 76, 4419-4432.

Zhu, B.; Eurell, T .; Gunawan, R.; Leckband, D. "Chain-length dependence of the
protein and cell resistance of oligo(ethylene glycol)-terminated self-assembled
monolayers on gold", J. Biomed. Mater. Res. 2001, 56, 406—416.

Kalb, E.; Frey, S.; Tamm, L. K. "Formation of Supported Planar Bilayers by
Fusion of Vesicles to Supported Phospholipid Monolayers", Biochimica Et
Biophysica Acta 1992, 1103, 307-316.

Harris, J. M. Poly( ethylene glycol) chemistry : biotechnical and biomedical
applications; Plenum Press: New York, 1992.

Harris, J. M.; Zalipsky, 8.; American Chemical Society. Division of Polymer
Chemistry.; American Chemical Society. Meeting Poly( ethylene glycol) .-
chemistry and biological applications; American Chemical Society: Washington,
DC, 1997.

Feldman, K.; Hahner, G.; Spencer, N. D.; Harder, P.; Grunze, M. "Probing
resistance to protein adsorption of oligo(ethylene glycol)-terminated self-
assembled monolayers by scanning force microscopy", Journal of the American
Chemical Society 1999, 121, 10134-10141.

170

(166)

(167)

(168)

(169)

(170)

(171)

Nollert, P.; Kiefer, H.; Jahnig, F. "Lipid Vesicle Adsorption Versus Formation of
Planar Bilayers on Solid-Surfaces", Biophys. J. 1995, 69, 1447-1455.

Jenkins, A. T. A.; Bushby, R. J.; Evans, S. D.; Knoll, W.; Offenhausser, A.;
Ogier, S. D. "Lipid vesicle fusion on mu CP patterned self-assembled

monolayers: Effect of pattern geometry on bilayer formation", Langmuir 2002,
18, 3176-3180.

Efremov, A.; Mauro, J. C.; Raghavan, S. "Macroscopic model of phospholipid
vesicle spreading and rupture", Langmuir 2004, 20, 5724-5731.

Zhang, L. Q.; Longo, M. L.; Stroeve, P. "Mobile phospholipid bilayers on a
polyion/alkylthiol layer pair", Langmuir 2000, 16, 5093-5099.

Ma, Q; Srinivasan, M. P.; Waring, A. J.; Lehrer, R. 1.; Longo, M. L.; Stroeve, P.
"Supported lipid bilayers lifted from the substrate by layer-by-layer polyion

cushions on self-assembled monolayers", Colloid Surf B-Biointerfaces 2003, 28,
319-329.

Johnson, J. M.; Ha, T.; Chu, S.; Boxer, S. G. "Early steps of supported bilayer
formation probed by single vesicle ﬂuorescence assays", Biophys. J. 2002, 83,
3371-3379.

171

 

 

1C NNNNNNNNNNNN L A E

lwmnhnhnl

3 1293 02736 4631