. , 5...”.
in 22:. 3:...

3w

.. .3.“
:5 r3.
F.
. ..
fl:

 

fig

17

 

in .
1.1...

. “a? H.“ 5%... r.
lfit‘..«nn5. nu. 3..“in
Tflfiunnfi . .

.14 (.95

.Q
..

. .. a
.5. II.
32:13:...
$41.5
JR?

K

«At alvcxl...‘
1‘ 4 l. l.« .7!
:11

. .3 » 1v
"‘11r».

.1 ~ 5...:

A. Lit-41.x: l

E

1,. a.) :
....y_.“....;....‘w&.my..1

 

 

 

 

.2 00‘;

 

* LIBRARY
Michigan State
University

 

This is to certify that the
dissertation entitled

OPTICAL AND ELECTROCHEMICAL CHARACTERIZATION
OF BIOMACROMOLECULAR INTERACTIONS AT FLUID-
LIKE INTERFACES

presented by
LAVANYA PARTHASARATHY

has been accepted towards fulfillment
of the requirements for the

DOCTORAL degree in CHEMICAL ENGINEERING

Ma (trio/fl

MajodProfesder’s Signature
05/175? (15"

Date

 

 

MSU is an Affinnative Action/Equal Opportunity Institution

 

PLACE IN REIURN 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.

 

DATE DUE

DATE DUE

DATE DUE

 

AUG 0 3 2008

 

050419

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 chliC/DateDandd—pds

OPTICAL AND ELECTROCHEMICAL CHARACTERIZATION OF
BIOMACROMOLECULAR INTERACTIONS AT FLUID-LIKE INTERFACES

By

LAVANY A PARTHASARATHY

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

OPTICAL AND ELECTROCHEMICAL CHARACTERIZATION OF
BIOMACROMOLECULAR INTERACTIONS AT FLUID-LIKE INTERFACES

By

LAVANY A PARTHASARATHY

This thesis comprises three different studies, each aimed at understanding
biomacromolecular interactions at fluid-like interfaces. The first study is on sequential
and competitive adsorption and interactions of nisin and selected proteins at a model oil-
water interface, using total internal reflection fluorescence microscopy (TIRFM).
Experimental results showed strong evidence of significantly enhanced adsorption of
some proteins in the presence of nisin. Apparent interfacial concentrations of B-casein
were three times higher in the presence of nisin than when the pure protein adsorbed by
itself under identical conditions. Enhanced adsorption was also observed for lysozyme
and bovine serum albumin (BSA) in the presence of nisin, compared to adsorption of
each protein by itself. Very little enhanced adsorption was observed for fibronectin in the
presence of nisin. The exact mechanism for adsorption enhancement by nisin is not clear
at this time. However, the experimental evidence suggests strongly that molecular
weight, size of the adsorbing protein, and bulk protein concentration are all important

factors.

The second study concerned the development of a flow cell to enable
simultaneous optical and electrochemical characterization of biomimetic interfaces.
Modifications were made to our existing optical flow cell to incorporate the capability to

conduct cyclic voltammetry (CV) measurements using a three-electrode system. The

 

HF
_. we.

feasibility of the simultaneous optical-electrochemical set-up was tested using the dye
resorufin. On potential scans towards increasingly negative voltages, the fluorescent
resorufin was reversibly reduced to non-fluorescent dihydroresorufm, coincident with a
decrease in fluorescence emission intensity. Similarly, upon scanning towards
increasingly positive voltages, the dihydroresorufin was reversibly oxidized to resorufin,
resulting in an increase in fluorescence emission.

The third study outlines the process involved in the development and
characterization of a fructose dehydrogenase (FDH) biosensor based on indirect
bioelectrocatalysis, using coenzyme Q6 as the mediator. FDH, a naturally membrane-
bound enzyme, was incorporated along with Q6 into a bilayer lipid membrane (BLM)
deposited on indium tin oxide (ITO). The sensor was evaluated by using CV to monitor
the change in anodic current upon addition of D-fructose. However, this FDH biosensor
exhibited poor sensitivity, because electron tunneling through the BLM was very low.
To improve electron tunneling, octyltrimethoxysilane was used to form an 8-carbon
monolayer on ITO, and liposomes containing FDH and Q6 were deposited to form the
outer leaflet of a biomimetic layer. This sensor demonstrated a clear dependence of the
anodic current on D-fructose concentration, with a linear dependence over D-fructose
concentrations of 0.056mM to 1.98mM, and a plateau at higher D-fructose
concentrations. The bilayer was characterized optically using fluorescence recovery after
pattern photobleaching (FRAPP). The diffusion coefficient of the upper leaflet of the
biomimetic layer was 4.4 x 10'10 cmz/sec, with a membrane fluidity of 0.59. These

values are characteristic of what have been reported in the technical literature.

Dedicated to my parents and my brother

iv

ACKNOWLEDGEMENTS

I would like to begin with thanking my advisor, Dr. Robert Y. Ofoli. I am
fortunate to have worked with him over the past six years. Our interaction began when
he first recruited me. It has been a great pleasure working with him, and I sincerely thank
him for providing me with this opportunity. His hands-off approach and open door-policy
have taught me to be an independent thinker, and this is the most valuable attribute I will
carry forward. I would like to especially thank him for being patient with me while
reviewing my Ph.D. thesis and encouraging me all along the way — I hope I leave here a

better writer than when I joined!

I would like to thank Dr. Mark Worden for his helpful suggestions and for
serving on my committee, and also for the collaborations, and the free use of his
laboratory facilities (specifically the CV). . It has been a pleasure having Dr. Gary
Blanchard on my committee. I enjoyed taking his classes and appreciate his help with the
quantum yield experiments and for pointing me in the right direction to obtain the green
laser — Thank you. I would like to thank Dr. Christina Chan for serving on my committee

and her encouragement after my first defense.

I would like to thank Nancy, JoAnn, Jackie and Kim for always being helpful

with a bright smile. I appreciate all the things they did for me.

The completion of this dissertation would have been difficult without the support
of my friends, colleagues and family. Their constant support and encouragement during

the rough patches, and fi'iendship, guidance and companionship during other times made

this journey possible. I cannot imagine moving forward without them and look forward
to their continued association. I would like to begin with thanking Sachin Vaidya.
“Navigating the TIRFM waters” as he calls it, would have been next to impossible if we
had not worked together. The two of us started our experiments together in the lab and
he has been a great colleague and a wonderful friend. I shall always cherish the years
that I worked with him. I would like to thank Arun, Kiran, Bhaskar and Prema for always
being there for me. I would also like to thank Neeraj, Brian, Prajakt, Arnit, Sunder,

Tej as, Soumya, Giri and Mukta.

Another group of people that I am indebted to is my family here in Michigan. My
parents sent me here with the confidence and the assurance in their hearts that I would be
taken care of. I would like to begin by thanking Rengan Chithappa, Mali Chithi,
Ramesh, Subhashini, Aisha and Vijay - I cannot imagine what I would have done
without them and that they will always have a special place in my life. I would like to
thank another group of my family here (Cheen, Kannan, Kinna, Sridhar mama, Sudha,
J ayshree, Robbi, Jyothi, Preetha, Varshini, Netra, Keshav and Smriti) for supporting me

and providing me with enriching memories that I will treasure.

I would like to thank my family in India (Suja Chithi, Vardarajan Chithappa,
Balaji Mama, Amudha Manni, Srivathsan, Supraja, Vidya, Myla Athai and Sampath
Athimber) who always encouraged me and loved me unconditionally. I will always
remember and miss my grandparents and I wish my grandfather were here to share this

moment with me — he always encouraged me and took great pride even in my smallest

vi

achievements. I thank my grandmother Andalu Patty for her belief and pride in me.

Finally, I would like to acknowledge the peOple who were always there by my
side, without whom I cannot even remotely imagine coming this far in life — my parents
and my brother. It has been tough being far away from them, but they have ALWAYS
been there — taking pride in my accomplishments, encouraging me to achieve more,
always having faith in me and for all the sacrifices they have made. My mother has been
the beacon in my life. I thank my father for the unshakable faith he has always vested in
me. I end, knowing that all that I have achieved would not have been possible without

God’s grace.

vii

TABLE OF CONTENTS

INTRODUCTION .................................................................................................. 1
Enhanced adsorption of proteins in the presence of the polypeptide nisin ............. 2
Development of a protocol for simultaneous optical and electrochemical
characterization of biomacromolecules in TIR geometry .................................... 3

1.3 Development and optical-electrochemical characterization of a biosensor based on
D- fructose dehydrogenase (FDH) ....................................................................... 4
ENHANCED ADSORPTION OF PROTEINS IN THE PRESENCE OF THE
POLYPEPTIDE NISIN .......................................................................................... 7

2.1 ABSTRACT ............................................................................................................ 7
2.2 INTRODUCTION .................................................................................................. 8
2.2.1 Nisin ................................................................................................................. 8
2.2.2 Total Internal Reflection Fluorescence Microscopy (TIRFM) ...................... 10
2.2.3 The Liquid-liquid Interface ............................................................................ 11
2.3 MATERIALS AND METHODS .......................................................................... 12
2.3.1 Preparation of protein solutions ..................................................................... 12
2.3.2 Experimental Setup ........................................................................................ 13
2.4 RESULTS AND DISCUSSION ........................................................................... 15
2.5 CONCLUSIONS ................................................................................................... 26
DEVELOPMENT OF A PROTOCOL FOR SIMULTANEOUS OPTICAL AND
ELECTROCHEMICAL CHARACTERIZATION OF
BIOMACROMOLECULES IN TIR GEOMETRY ............................................. 44

3.1 ABSTRACT .......................................................................................................... 44
3.2 INTRODUCTION ................................................................................................ 45
3.2.1 Optical characterization ................................................................................. 46
3.2.1.1 Total Internal Reflection Fluorescence Microscopy (TIRFM) ................. 46

3.2.2 Electrochemical characterization ................................................................... 47
3.2.2.1 Cyclic voltammetry (CV) ......................................................................... 47

3.3 EXPERIMENTAL SET-UP ................................................................................. 52
3.3.1 TIRFM flow cell ............................................................................................ 52
3.3.2 Modifications to TIRFM set-up to conduct optical and electrochemical
experiments (first generation) ........................................................................ 53

3.3.2.1 Materials and Methods: Ferricyanide Experiments .................................. 54

3.3.3 Results and Discussion: Ferricyanide Experiments ....................................... 55
3.3.4 Problems with first generation optical — electrochemical set-up and resulting
modifications (second generation experimental set-up) ................................ 58

3.3.4.1 Problems encountered with the second generation experimental set-up .. 60
3.3.4.2 Proof-of-concept of optical/electrochemical set-up .................................. 61

3.3.5 Introduction to resazurin ................................................................................ 62

viii

4

5

3.3.5.1 Materials and methods: Resazurin Experiments ....................................... 64

3. 3. 6 Experimental details: Resazurin Experiments ............................................... 64

3. 3. 7 Results and Discussion. Resazurin Experiments ........................................... 65

3. 4 CONCLUSION ..................................................................................................... 68
3.5 FUTURE WORK .................................................................................................. 69
DEVELOPMENT AND OPTICAL-ELECTROCHEMICAL
CHARACTERIZATION OF A BIOSENSOR BASED ON D-FRUCTOSE
DEHYDROGENASE (FDH) ................................................................................ 88

4.1 ABSTRACT .......................................................................................................... 88
4.2 INTRODUCTION ................................................................................................ 89
4.3 DEVELOPMENT OF AN FDH-BASED SENSOR ............................................ 90
4.3.1 Introduction to FDH ...................................................................................... 90
4.3.2 Theory ............................................................................................................ 92
4.3.3 Introduction to FRAPP .................................................................................. 94
4.4 EXPERIMENTAL SECTION .............................................................................. 97
4.4.1 Materials ........................................................................................................ 97
4.4.2 Preparation of the proteoliposome solution ................................................... 98
4.4.3 Adsorption of proteoliposome solution onto ITO-coated glass slides ........... 99
4.4.4 Preparation of hydrophobic silane monolayers on ITO-coated glass slides 100
4.4.5 Adsorption of proteoliposomes onto silanized ITO slides .......................... 100
4.4.6 Cyclic Voltammetry Experiments ............................................................... 101
4.4.7 Set-up for FRAPP Experiments ................................................................... 101
4.4.7.1 Optical set-up for FRAPP experiments ................................................... 101
4.4.7.2 Experimental cell used for FRAPP experiments to study F DH system . 103
4.4.7.3 Preparation of biomimetic layer for FRAPP experiments ...................... 104

4.5 RESULTS AND DISCUSSION ......................................................................... 104
4.5.1 Incorporation of FDH in a biomimetic layer ............................................... 109
4.5.1.1 Optical characterization .......................................................................... 109
4.5.1.2 Electrochemical characterization ............................................................ 110

4.6 CONCLUSIONS ................................................................................................. 11 1
4.7 FUTURE WORK ................................................................................................ 112
APPENDICES .................................................................................................... 125

5.1 APPENDIX A: .................................................................................................... 125

5.1.1 MATLAB program used for filtering analog measurements to remove dark

current measurements obtained during the phases when the chopper blocked

the laser beam. ............................................................................................. 125
5.1.2 MATLAB program used for averaging data ............................................... 126
5.1.3 Labview Program for collecting fluorescence data in the analog mode. ..... 127

5.1.4 Labview Program for collecting fluorescence data in the digital (photon

counting) mode. ........................................................................................... 129

ix

REFERENCES ............................................................................................................... 132

LIST OF TABLES

Table 3-1. Input parameters used to obtain cyclic voltammetry measurements for the
ferricyanide experiments ....................................................................................... 70

Table 3-2. Input parameters used to obtain cyclic voltammetry measurements for the
resazurin and resorufin experiments. .................................................................... 71

Table 3-3. Electrochemical properties of the dyes resazurin and resorufin (Maeda et al.
2000; Maeda et al. 2001). ..................................................................................... 72

Table 3—4. Optical properties of the dyes resazurin and resorufin (Maeda et al. 2000;
Maeda et al. 2001) ................................................................................................. 73

Table 4-1. Input parameters used to conduct cyclic voltammetry measurements for the
D-Fructose dehydrogenase system. .................................................................... 113

Table 4—2. Summary of regression values obtained by fitting experimental FRAPP data

to the single species model [Equation (4.8) (Wright et al. 1988; Starr and
Thompson 2002). ................................................................................................ 114

xi

LIST OF FIGURES

Figure 2-1. Experimental set-up for TIRFM experiments ............................................... 28
Figure 2-2. Equilibrium adsorption of [Pi-casein followed by introduction of nisin. ........ 29
Figure 2-3. Adsorption of nisin followed by introduction of B-casein. ........................... 30
Figure 2-4. Assessment of B-casein adsorption reversibility. .......................................... 31
Figure 2-5. Enhanced adsorption of B-casein in the presence of nisin. ........................... 32
Figure 2-6. Enhanced adsorption of B-casein in the presence of nisin. ........................... 33
Figure 2-7. Control experiment for B-casein. .................................................................. 34
Figure 2-8. Interaction of BSA and nisin. ........................................................................ 35
Figure 2-9. Proposed scheme of the interaction between B-casein and nisin. ................. 36
Figure 2-10. Interaction of lysozyme and nisin. .............................................................. 37
Figure 2-11. Control Experiment for Lysozyme .............................................................. 38
Figure 2-12. Lysozyme adsorption. ................................................................................. 39
Figure 2-13. Interaction of fibronectin and nisin. ............................................................ 40
Figure 2-14. Interaction of lysozyme (at a lower concentration) and nisin. .................... 41
Figure 2-15. Interaction of BSA (at a lower concentration) with nisin. .......................... 42
Figure 2-16. Interaction between F ibronectin (at a lower concentration) and nisin. ....... 43
Figure 3-1. A typical cyclic voltammogram. ................................................................... 74
Figure 3-2. Diagram of flow cell used for TIRFM characterization experiments. .......... 75
Figure 3-3. Diagram of flow cell for optical and electrochemical experiments (first

generation design). ................................................................................................ 76
Figure 3-4. Cyclic Voltammogram obtained in the first generation TIRFM-CV flow cell.

Figure 3-5.

........................................................................................................................ 77

Cyclic voltammogram obtained in a conventional three-electrode set-up for

comparison. ........................................................................................................... 78

Figure 3-6.

Comparison of cyclic voltammograms obtained in our modified first

xii

generation TIRFM-CV set-up, and in a conventional three-electrode set-up ....... 79
Figure 3-7. Diagram of second generation optical — electrochemical flow cell. ............. 80

Figure 3—8. Reaction “a” is an irreversible reaction that converts resazurin to fluorescent

resorufin; reaction “b” is a reversible reaction that converts resorufrn to non-
fluorescent dihydroresorufm. ................................................................................ 81
Figure 3-9. Cyclic Voltammogram of 0.02mM resazurin in PBS buffer. ....................... 82

Figure 3-10. Fluorescence data obtained in the TIRFM-CV flow cell containing 0.02mM
resazurin. ............................................................................................................... 83

Figure 3-11. Combined plot of fluorescence vs. time and applied potential vs. time for
0.02mM resazurin. ................................................................................................ 84

Figure 3-12. Cyclic Voltammogram of 0.002mM resorufin in PBS buffer ..................... 85

Figure 3-13. Fluorescence data obtained in the TIRFM-CV flow cell for 0.002mM
resorufin. ............................................................................................................... 86

Figure 3-14. Combined plot of fluorescence versus time and applied potential versus
time for 0.002mM resorufin .................................................................................. 87

Figure 4-1. Surface chemistry used to form an 8-carbon monolayer on ITO-coated glass
surface using trimethoxyoctylsilane as the linker molecule ............................... 115

Figure 4-2. Experimental set-up for fluorescence recovery after pattern photobleaching
(FRAPP) experiments in the epi-illumination mode. ......................................... 116

Figure 4-3. Fringe pattern in illuminated region obtained by using 100 lines per inch
ronchi ruling. b) The observation area is restricted by placing an aperture in the
camera/PMT image plane, to ensure that fluorescence recovery is restricted to

fluorophores in the fringes. ................................................................................. 117
Figure 4-4. CV curves obtained for a BLM-based FDH sensor. ................................... 118
Figure 4-5. Expanded view of relevant portion of data. ................................................ 119
Figure 4-6. Catalytic response of an immobilized FDH biosensor. ............................... 120
Figure 4-7. Expanded view of the relevant portion of data. .......................................... 121
Figure 4-8. Calibration curve of current density versus D-fr'uctose concentration ........ 122
Figure 4-9. Linear part of calibration curve at 800mV. ................................................. 123

Figure 4-10. (a) Typical FRAPP curve obtained with FDH and the coenzyme

xiii

immobilized in the biomimetic layer. The solid line represents fit to the data,
using a model that describes a population containing a single mobile and
immobile fraction (Smith and McConnell 1978; Wright et al. 1988; Starr and
Thompson 2002). (b) Plot of residual versus time showing excellent goodness of
fit. ........................................................................................................................ 124

Figure 5-1. Labview flowsheet for collecting fluorescence data in the analog mode (left
hand side) (Figure 5.1 continued on next page) .................................................. 127

Figure 5-2contd. Labview flowsheet for collecting fluorescence data in the analog mode
(Right hand side of Figure 5.1) ........................................................................... 128

Figure 5—3. Labview flowsheet for collecting fluorescence data in the digital (photon
counting) mode (Figure 5-3 continued on next two pages). ............................... 129

Figure 5-4contd. Labview flowsheet for collecting fluorescence data in the digital
(photon counting) mode (Figure 5-3 continued on next page). .......................... 130

Figure 5-5contd. Labview flowsheet for collecting fluorescence data in the digital
(photon counting) mode. ..................................................................................... 131

xiv

Ag/AgCl
a.u.
BSA

BLM

C0

CV

b

hggmgmng‘mga.
'b

ABBREVIATIONS

angle of incidence of laser beam

critical angle

wavelength of the incident laser light
transfer coefficient

scan rate

electrode area

silver/silver chloride

arbitrary units

bovine serum albumin

bilayer lipid membrane

concentration

coatings

cyclic voltammetry

diffusion coefficient

penetration depth of the evanescent wave
external applied potential

formal reduction potential

anodic peak potential

cathodic peak potential

separation of the anodic and cathodic peak potential

faraday’s constant

XV

Fl , F2
FDH
FITC
F PR

FRAPP

K3Fe(CN)6
KNO3
m

M1,M2

NAD
NADH
NDF
"I

"2

[0]

pb

PBS

optical flats

fructose dehydrogenase

fluorescein iso-thiocyanate

fluorescence photobleaching recovery
fluorescence recovery after pattern photobleaching
peak current

anodic peak current

cathodic peak current

indium tin oxide

standard homogenous electron-transfer rate constant
potassium ferricyanide

potassium nitrate

monitoring beam

mirrors

number of electrons

nicotinamide adenine dinucleotide

reduced form of nicotinamide adenine dinulecotide
neutral density filter

refractive index of the denser medium

refractive index of the rarer medium

concentration of oxidized species

photobleaching beam

phosphate buffered saline

xvi

[RI

sADH
SCE

Sh

TIR

TIRFM

concentration of reduced species
universal gas constant

detergent solution

secondary alcohol dehydrogenase
saturated calomel electrode
shutter

temperature

total internal reflection

total internal reflection fluorescence microscopy

xvii

1 INTRODUCTION

The adsorption of proteins at interfaces and their behavior in the adsorbed state
are critical in controlling the function of many biological systems. For example, protein
adsorption at oil-water interfaces is important in stabilization of emulsions in food
products and cosmetics (Dickinson 1977; Dickinson 1994). Protein adsorption also plays
an important role in biosensors and in biomedical applications such as rapid diagnostic
assays and development of biocompatible devices. Protein adsorption is the first step that
occurs when a body reacts to an implanted biomaterial (Andrade 1985), and its
adsorption to biomaterials is important in modulating cell adhesion (Hynes 1992; Sanocki
2004). Understanding of competitive interactions between proteins adsorbed at interfaces
is also important in biological systems, because most processes involve adsorption of

proteins from complex media such as blood or serum which contain a variety of species.

Though each chapter of this thesis addresses a unique problem, the common
theme is that they are all geared towards understanding biomacromolecular interactions at
fluid—like interfaces. Chapter 2 presents work on the characterization of the adsorption
and interaction of the polypeptide nisin and selected proteins at the oil-water interface.
Chapter 3 outlines the process of developing an experimental protocol that enables
simultaneous optical and electrochemical characterization of biomacromolecules in total
internal reflection (TIR) geometry, along with proof-of-concept experiments to
demonstrate the feasibility of the set-up, using the dye resorufin. Following this

demonstration, the experimental tool was utilized for optical and electrochemical

characterization of a biosensor based on D-fructose dehydrogenase (FDH). Details of

this study are presented in Chapter 4. Each of these studies is briefly discussed below.
1.1 Enhanced adsorption of proteins in the presence of the polypeptide nisin

Understanding conformational changes and the behavior of proteins upon
adsorption to interfaces is important in a variety of pharmaceutical and biomedical
applications, including rapid diagnostic assays, development of scaffolds for tissue
engineering, and drug delivery. Drugs have to be delivered to intracellular sites for
effective therapeutic applications. The hydrophobic nature of the membrane does not
allow the drug to translocate across the membrane by itself. Therefore, one of the major

challenges facing drug delivery is achieving drug transport across cell membranes.

Lantibiotics possess a unique physical structure which makes them highly surface
active. They are also amphiphilic, thus enhancing their potential for use as emulsifiers in
drug formulations. One of the lantibiotics of interest is nisin, which has a unique
property in its ability to translocate across the cell membrane. Its small size, surface
activity and ability to translocate across the membrane makes it a good candidate for
transporting therapeutic compounds across the membrane (Bower et al. 2001). In
particular, the potential for transporting drugs such as insulin in nisin-formulated
emulsions through the mucosa] membrane in the nasal passage is very promising, and
was a primary motivation for this study. As part of the preliminary experiments towards
achieving this goal, we characterized the adsorption and interactions of nisin at the oil-

water interface, generally considered a crude but effective approximation of a biological

membrane (Volkov 1998). For example, the oil-water interface has been used to model
drug transport across cellular membranes, as a gauge of pharmacological activity (Arai et
al. 1993). In Chapter 2 of this thesis, we present results of a study investigating the
adsorption and interactions of nisin with selected labeled proteins at the oil-water
interface, and report observations of enhanced interfacial adsorption of certain proteins in

the presence of nisin.

1.2 Development of a protocol for simultaneous optical and electrochemical

characterization of biomacromolecules in TIR geometry

Analytical devices such as biosensors are characterized by complex interactions at
their biomimetic interfaces. However, most characterization techniques use one protocol
at a time. We believe the ability to characterize interactions using more than one
technique simultaneously would be very useful in providing additional information to aid
in understanding and interpreting phenomena in these complex systems. As a result, we
have incorporated the ability to conduct electrochemical measurements in our existing
optical flow cell, to provide an important tool for situations when optical and
electrochemical characterization can be used together to corroborate occurrence of the
same event on one sample. For example, in enzyme-based biosensors which utilize
nicotinamide adenine dinucleotide (NAD), the reversible conversion of NAD to NADH
can be monitored optically by the emission of fluorescence at 450nm. The conversion
between the two redox forms NAD and NADH can also be tracked electrochemically

using cyclic voltammetry (CV).

Alternatively, this tool would also be useful in situations when optical and
electrochemical measurements can give two independent pieces of information about a
single sample. For example, in enzymatic sensors based on bilayer lipid membrane
(BLM), optical measurements can be used to confirm the formation and integrity of the
lipid bilayer, and electrochemical measurements can be used to study redox reactions

catalyzed by enzymes at the same biomimetic interface.

To this end, we have modified the flow cell used in our total internal reflection
fluorescence microscopy (TIRFM) apparatus to incorporate the capability to also conduct
electrochemical measurements. The details of this work are given in chapter 3. The
functionality of the modified set-up was tested and confirmed by monitoring the
reversible redox conversion of resorufin to dihydroresorufin. When the potential was
scanned to reduce resorufin to dihydroresorufm, a decrease in fluorescence was observed.
Similarly, when the potential was scanned to oxidize dihydroresorufin to resorufin, an
increase in fluorescence emission was observed. In addition to allowing the estimation of
interfacial concentrations, the resulting data could potentially be used to understand the

kinetics of the reaction.

1.3 Development and optical-electrochemical characterization of a biosensor

based on D- fructose dehydrogenase (FDH)

During the past decade, there has been widespread interest in the development of
biosensors for applications in areas such as food technology, agriculture, environmental

monitoring, and biomedical devices (Castillo et al. 2004; Mehrvar and Abdi 2004).

Biosensors are being increasingly used due to their ease of miniaturization, high
specificity and sensitivity, short response times, and the fact that analyte detection does
not require prior separation of the sample (Castillo et al. 2004). A biosensor is an
analytical device that detects biochemical and physiological changes. It comprises a
receptor (biological recognition element) and a transducer in close proximity to each
other. It converts a specific recognition event between the receptor and the target analyte
into a measurable signal. There are two main classes of biosensors — catalytic or
bioaffinity. Catalytic biosensors enable measurement of the steady-state concentration of
analytes formed and/or consumed during a biocatalytic reaction. Bioaffinity biosensors,
on the other hand, are used to detect specific interactions between a receptor molecule

and a target analyte (antibody-antigen interactions).

Some common modes of transduction in biosensors are electrochemical
(potentiometric, amperometric, conductometric) (Chaubey and Malhotra 2002;
Dzyadevych et al. 2002; Gerard et al. 2002), optical (absorbance, fluorescence,
luminescence, light scattering, refractive index) (Cush et al. 1993; DiazGarcia and
ValenciaGonzalez 1995; Barak et al. 1997; Karlsson and Falt 1997; Lin et al. 1997;
Mehrvar et a1. 2000), mechanical (Raiteri et al. 2001), thermal (Abel et al. 1996; Ivnitski
et al. 1999; Ramanathan and Danielsson 2001), piezoelectric (Muramatsu et al. 1987;
Davis and Leary 1989; Saini et al. 1991), and magnetic (Baselt et al. 1998). The
preferred transduction modes are usually electrochemical for catalytic biosensors, and
optical for bioaffinity biosensors. The most widely used catalytic biosensors are enzyme-

based electrochemical sensors, which are of special interest because of the high

specificity, selectivity and efficiency of the enzyme towards the target analyte.

We have developed an electrochemical catalytic biosensor based on the enzyme
fructose dehydrogenase (FDH) for the detection of D-fi'uctose, and have characterized it
both optically and electrochemically (Chapter 4). FDH, a membrane-bound enzyme was
immobilized in a lipid bilayer membrane with the mediator coenzyme Q6, also
incorporated into the bilayer. The biomimetic layer was deposited on transparent indium
tin oxide (ITO)-coated glass, allowing us to characterize the biosensor both Optically and
electrochemically. The formation of the lipid bilayer on the ITO-coated slide was
confirmed optically, using fluorescence recovery after pattern photobleaching (FRAPP).
The functionality of the biosensor was confirmed electrochemically, based on increased

catalytic current upon addition of D-fructose to the flow cell.

2 ENHANCED ADSORPTION OF PROTEINS IN THE PRESENCE OF THE
POLYPEPTIDE NISIN

2.1 ABSTRACT

We present sequential and competitive adsorption experiments on interactions
between nisin and selected proteins at a model oil-water interface. We used total internal
reflection fluorescence microscopy (TIRFM) with an evanescent wave configured to a
depth of 85 nm, to confine the area of observation to the interface and its immediate

vicinity. All proteins except nisin were labeled with fluorescein isothiocyanate (FITC).

Experimental results showed strong evidence of significantly enhanced adsorption
of some proteins in the presence of nisin. Apparent interfacial concentrations of B-casein
were nearly three times higher in the presence of nisin than when the pure protein
adsorbed by itself under identical conditions. Enhanced adsorption was also observed for
lysozyme and bovine serum albumin (BSA) in the presence of nisin, in comparison to
adsorption of each protein by itself. Very little enhanced adsorption was observed for

fibronectin in the presence of nisin.

The exact mechanism for adsorption enhancement by nisin is not fully clear.
However, the experimental data suggest strongly that molecular weight, size of the

adsorbing protein, and bulk protein concentration are all important factors.

2.2 INTRODUCTION

2.2.1 Nisin

Nisin is a small (3.5 kDa) positively charged amphipathic polypeptide, produced
by Lactococcus lactis. It contains dehydrated residues and lanthionine or thioether rings,
has antimicrobial activity against a large variety of gram-positive bacteria, and has been
widely used as a food preservative in the dairy industry (Hurst 1981). Its antimicrobial
activity stems primarily from permeabilization/pore formation of the cytoplasmic
membrane of target organisms (Breukink et al. 1998; Van Kraaij et al. 1998; Breukink
and de Kruijff 1999; Breukink et al. 1999; Enserink 1999; Van heusden et al. 2002). The
peptide initially binds to the membrane with its C-terminus through electrostatic
interactions; after that the N-terminus is inserted into the lipid phase of the membrane.
Lipid-H, a cell wall precursor, binds to nisin, resulting in a transmembrane orientation of

the polypeptide; subsequent oligomerization of nisin and lipid-H leads to pore formation.

Because of its antimicrobial properties, there is interest in using nisin in
formulations that could effectively transfer drugs across the cellular membrane (Bower et
al. 2001). To realize this promise, it is necessary to understand the interactions of nisin at
fluid-like interfaces. Thus, the goal of this study was to characterize the adsorption
dynamics and interactions of nisin at a model oil-water interface, and to assess its

potential for use in pharmaceutical formulations.

The primary tool for characterizing protein adsorption and interactions in our

laboratory is total internal reflection fluorescence microscopy (TIRFM), which requires
proteins to be labeled with fluorophores. However, nisin was not labeled in this work
because a previous study in our laboratory had demonstrated that fluorescent labels can
markedly affect both protein diffusion and interfacial dynamics (Gajraj and Ofoli 2000b).
Due to its small size (3.5 kDa) in relation to the fluorophore we commonly use to label
proteins in our laboratory (FITC, ~0.350 kDa), attaching this label to nisin had the
potential of significantly affecting its interfacial dynamics. To avoid this, we decided to
track nisin’s interfacial dynamics through its competitive interactions with selected
labeled proteins at the interface. This decision was based in part on several studies that
had reported that nisin adsorbs strongly at the hydrophobic/hydrophilic interface (Bower
1995; Giffard et al. 1996; Lakarnraju et al. 1996; Bani-Jaber et al. 2000; Lee et al. 2000).
We hypothesized that the combination of its high interfacial activity and small size
(which usually translates into high diffusivity) (Dickinson 1977) would enable nisin to
relatively easily displace other proteins from the interface. This would enable us to
monitor its interfacial behavior by measuring the dynamics of displacement of labeled

proteins from the interface.

This interfacial characterization was done in sequential and competitive
adsorption experiments, using a diverse group of labeled proteins of varying
characteristics. We measured interactions of nisin with two fibrous proteins [ll-casein
(24 kDa) and fibronectin (440 kDa)] and two globular proteins [lysozyme (14.3 kDa) and
bovine serum albumin (66 kDa)]. This group of proteins was selected to provide a range

of hydrodynamic size, molecular weight, and structure. In addition, the interfacial

9

behavior of each of these proteins has been characterized in our laboratory (Gajraj and
Ofoli 2000b; Vaidya and Ofoli 2005), so their adsorption characteristics were well

understood.

2.2.2 Total Internal Reflection Fluorescence Microscopy (TIRFM)

Protein adsorption was measured using total internal reflection fluorescence
microscopy (TIRFM), an optical technique that enables molecular level investigations of
macromolecular adsorption and interactions at a variety of interfaces. It provides the
ability to selectively illuminate macromolecules within tens of nanometers of the
interface and is a non-intrusive, in situ technique. TIRFM has been used at the liquid-
liquid interface to study protein dynamics (Tupy et a1. 1998; Gajraj and Ofoli 2000a;
Gajraj and Ofoli 2000b; Vaidya and Ofoli 2005), and at the solid-liquid interface to study
receptor-ligand interactions (Asanov et al. 1998), living cells and single molecules,
adsorption and relaxation kinetics of proteins (W ertz and Santore 1999; Wertz and
Santore 2001; Wertz and Santore 2002; Wertz and Santore 2002), diffusion, adsorption
kinetics and lateral mobility of proteins (Giffard et al. 1996; Yuan et al. 2003, Thompson
et a1. 1981, Gajraj and Ofoli 2000b; Vaidya and Ofoli 2005), and formation and
morphology of lipid bilayers (Wagner and Tamm 2000). We present here only a brief
introduction to TIRFM. Further detail can be obtained from Axelrod et al. 1983; Axelrod

1989; Axelrod 1992; Gajraj 1999; Gajraj and Ofoli 2000a.

When a beam of light traveling in a medium of higher refractive index (n1) strikes

a medium of lower refractive index (n 2) at an angle (0) greater than the critical angle (19c),

10

the beam is totally internally reflected. The critical angle is given by

BC = sin—1(2), n2 < n1 (2.1)

"1

Upon total internal reflection, an electromagnetic field or evanescent wave is
induced in the rarer medium at the interface. The amplitude of the evanescent wave
decays exponentially with depth into the rarer medium, with a penetration depth (dp) that

depends on the refractive indices of the two media, the angle of incidence (6) and the

wavelength of the incident light (’10):

 

dp = 30 2 (2.2)
MUM/sin2 t9 —[§l]
"r

The penetration depth is usually on the order of 80nm with the setup in our

laboratory, thus making this a truly surface-selective technique.
2.2.3 The Liquid-liquid Interface

Liquid-liquid interfaces are of considerable importance in biological systems,
particularly in pharmacology, food processing, biomedical engineering and
biotechnology. For example, the distribution of proteins and surfactants at the liquid-
liquid interface influences the properties of food systems such as emulsions and
microemulsions (Dickinson 1998). Liquid-liquid interfaces also provide a simple

approximation to cellular interfaces where many important biological processes such as

11

receptor-ligand interactions, electron transfer, enzymatic reactions and drug delivery

occur (Arai et al. 1993).

Thus, the study of protein adsorption and interactions at fluid-like interfaces is
relevant in understanding the complex processes taking place in biological systems. In
particular, the oil-water interface serves as a crude approximation of a biological

membrane (V olkov 1998).

2.3 MATERIALS AND METHODS

2.3.1 Preparation of protein solutions

Pure nisin was provided as a generous gift from Dr. Joseph McGuire, Department
of Food Science and Technology, Oregon State University, Corvallis, OR. Bovine B-
casein was purchased from the Hannah Research Institute (Lot No. B/6/6, Ayr, Scotland)
at 95% purity. Bovine serum albumin (A-7511) and lysozyme (L-2879) were obtained
from the Sigma Chemical Co. (St. Louis, MO), and fibronectin (FC010) was obtained
from Chemicon Intemational (Temecula, CA). All proteins were used as received,
without further modification. Sodium phosphate monobasic and sodium phosphate
dibasic, which were used in the preparation of the buffers, were purchased from the

Sigma Chemical Company (St. Louis, MO).

Nisin was dissolved in sodium phosphate monobasic (pH 4.5, 0.05M) to ensure
complete solubilization. Sodium phosphate dibasic (pH 9.1, 0.05M) was added to the

nisin solution to bring the pH up to 7.4 (Lee et a1. 2000). All proteins except nisin were

12

labeled with fluoroscein-S-isothiocyanate (FITC, F-1907, Molecular Probes, Eugene,
OR), using the protocol given by Brinkley (Brinkley 1992). In brief, the labeling reaction
for each protein was carried out at room temperature in carbonate buffer (pH 9.1, 0.05M)
for periods of four to six hours in the dark. The protein solution was then dialysed
against phosphate buffer (pH 7.40, 0.05M) to remove any unreacted dye. Dialysis was
carried out over a period of 24 to 36 hours, in a 4-6 stage continuously stirred operation
with a regenerated cellulose porous membrane (Spectra/For 1, 132655, molecular weight
cutoff of 6000 to 8000, Spectrum Laboratories, Rancho Dominguez, CA). The solutions
of labeled proteins were initially frozen in aliquots, and thawed and diluted to the
required concentration before each experiment. Absorbance spectroscopy measurements
were carried out on a diode array spectrophotometer (Model 8452A, Hewlett-Packard,
Brielle, NJ) to determine the concentrations of proteins and the labeling ratios. A
labeling ratio (dye to protein) less than unity was usually used for all experiments, to

avoid potential problems with concentration quenching.
2.3.2 Experimental Setup

Our experimental set-up (Figure 2-1) has been described in detail in an earlier
paper (Gajraj and Ofoli 2000a). An important modification to the original setup is that
we have added a photon counter to detect very low fluorescence emission intensities at
the interface. Experiments at high concentrations (concentrations higher than 10'5M)
were conducted in stop-flow mode, where the protein solution continuously flowed into

the sample cell for a period of about 30 minutes, after which flow into the cell was

13

terminated. Experiments at lower concentrations (particularly below 10'7M) were carried
out in continuous flow mode, to avoid the possibility of bulk depletion of proteins during

the experiment.

Solutions were simultaneously infused and withdrawn from the flow cell at a rate
of 0.047 ml/min (shear rate of 0.435"), using a Harvard Apparatus infusion/withdrawal
syringe pump (Model 551382, Harvard Apparatus, South Natick, MA). Fluorescence
emission was measured with a photomultiplier tube (R4632, Hamamatsu, Bridgewater,
NJ) for protein solutions at high concentrations, and the analog measurements were
filtered to remove dark current during periods where the chopper blocked the laser beam.
Data filtering was carried out using a program written in MATLAB (Mathworks inc.,
Natick, MA) (see Appendix A). Fluorescence measurements for the low concentration
experiments were collected using the same photomultiplier tube, with the signal
amplified by a SR445 pre-amplifier (Stanford Research Systems, Sunnyvale, CA). The
output of the pre-amplifier was sent to a photon counter (SR400 Dual Channel Gated
Photon Counter, Stanford Research Systems, Sunnyvale, CA), which was used to
accurately track both adsorption and desorption, even at bulk protein concentrations as
low as 10‘8 M. Labview 6.0 (National Instruments, Austin, TX) served as the
computerized data acquisition software (Figure 5-1 and Figure 5-3 in Appendix A).

Some images in this dissertation have been presented in color.

Since many of the experiments were conducted over long periods of time

(durations of several hours sometimes), we took precautions to prevent unintended

14

photobleaching of the fluorophore by the monitoring beam. First, a beam chopper
(SR540, Stanford Research Systems, Sunnyvale, CA) was continuously used during all
experiments. In addition, the photon counter was triggered by the chopper and gated
such that data were collected only when the flow cell was illuminated by the laser beam.
Secondly, the monitoring beam was blocked during long experiments, to provide
additional protection against photobleaching; the beam was allowed to illuminate the
interface only during periods of fluorescence emission measurements. Finally, the
monitoring beam was configured to a low intensity of 2-3pW, as measured by a Newport

power meter (818-SL, Newport, Irvine, CA).

All experiments were conducted at the oil-water interface, using a flow cell
described in an earlier paper (Gajraj and Ofoli 2000a). For this study, the oil-water
interface was formed by bringing a 20-60um layer of oil in contact with an 800nm layer
of water. The bottom slide was coated with polyethylene oxide (PEO, Sigma Chemical
Co., St. Louis, M0) to eliminate protein adsorption on that surface (J eon and Andrade
1991; Andrade et al. 1996). Readings acquired at different locations at the oil—water
interface were reproducible during a given experiment; data acquired on the same batch

of labeled proteins were also reproducible.

2.4 RESULTS AND DISCUSSION

We conducted competitive and sequential adsorption experiments using unlabeled
nisin and F ITC-labeled B-casein. We chose to work with B-casein because its adsorption

characteristics have been well—characterized in several studies (Graham and Phillips

15

1979; Hunter et a1. 1991; Dickinson et al. 1993; Russev et a1. 2000). To minimize shear
at the interface, very low flow rates of approximately 0.045m1/min (shear rates of 0.435")
were used to introduce protein solutions into the flow cell. The time for protein
adsorption to approach pseudo-steady state behavior in our flow cell (as indicated by the
fluorescence emission intensity profile) varied between 30 and 50 minutes. This is much
faster than the characteristic time of adsorption at interfaces of normal depth (which can
be as long as a few hours). The reduced time of adsorption is the result of the thinness of

the oil-water assembly, which is on the order of 800nm.

In initial experiments, labeled B-casein at a bulk concentration of 10'5M was
introduced into the flow cell and allowed to adsorb at the interface in a continuous flow
format. When the fluorescence emission intensity level approached a pseudo-steady state
value after about 35 minutes, nisin (at a bulk concentration of 10'5 M) was introduced into
the flow cell (Figure 2-2). As expected, we observed a continuous drop in fluorescence
emission intensity upon introduction of nisin. This behavior appeared to confirm our
initial hypothesis that nisin, being smaller and more surface-active, was displacing
adsorbed B-casein from the interface. However, there are at least two other plausible
explanations for the behavior observed in Figure 2-2. The first is that B-casein was
removed from the interface as a result of the shear induced by the continuous flow of
nisin into the flow cell as the experiment proceeded. This is unlikely, because earlier
experiments conducted in our laboratory had demonstrated that the liquid-liquid interface

remains stable at flow rates of up to 0.05ml/min. The second plausible explanation is

16

that, since the nisin solution contained no B-casein, the adsorbed B-casein is removed
from the interface by mass transfer resulting from the concentration gradient induced by
the depletion of proteins in the bulk. That is, the bulk concentration of B—casein

continuously decreased as the nisin solution flowed into the sample cell.

To investigate this When we ran the reverse experiment where nisin at the same
bulk concentration (10'5M) was adsorbed at the interface for a period of about 30
minutes, followed by the introduction of B-casein (also at a bulk concentration of IO'SM)
into the flow cell. Upon introduction of B-casein, we observed an increase in the
fluorescence emission intensity, indicating that B-casein was adsorbing at the oil-water
interface, even in the presence of pre-adsorbed nisin (Figure 2-3). This was an
unexpected result, because the presence of nisin at the interface clearly did not preclude
or even significantly hinder B-casein adsorption. In fact, this result appeared to provide
evidence that nisin is less surface-active than B-casein. However, another plausible
explanation was that, at the low bulk concentrations at which the experiments were
conducted, the interface was never saturated, allowing both nisin and B-casein to adsorb

and coexist with essentially no effective interfacial competition.

To evaluate this possibility, we conducted an experiment where we allowed for
equilibrium adsorption of B-casein (IO'SM) at the oil-water interface, after which a
protein-free buffer solution was introduced into the flow cell. Upon introduction of the
buffer solution, a significant drop in the fluorescence emission level was observed,

indicating desorption of B-casein from the interface (Figure 2-4). The extent of apparent

l7

B-casein desorption from the interface was comparable to what was observed earlier upon
interaction of [El-casein with nisin (Figure 2-2). This indicated that B-casein was removed

because it is reversibly (loosely) adsorbed at the interface, not because nisin has superior

interfacial activity by comparison.

Follow-up experiments were modified to help determine if concentration
gradients play a role in the desorption of B-casein from the interface. Similar to the first
set of experiments, we allowed B-casein (10'5M) to adsorb at the oil-water interface until
the fluorescence signal approached pseudo-equilibrium. We then introduced a mixture
of nisin and B-casein (each at identical concentrations of 10'5M) into the flow cell. The
infiision of both B-casein and nisin was done to maintain the B-casein bulk concentration
at a fixed level, to prevent bulk depletion of the protein. Upon introduction of the
mixture, we observed a significant increase in fluorescence emission intensity, indicating
that B-casein was adsorbing in substantially larger quantities than was observed when it
adsorbed alone at the interface (Figure 2-5). The level of fluorescence emission intensity
represents an increase of about 3 times the level initially observed, which qualitatively
indicates that approximately three times more B-casein was adsorbing at the interface in
the presence of nisin, in comparison to B-casein adsorbing alone. In addition, the slope of
the initial adsorption profiles for B-casein adsorbing alone was essentially identical to that
for the adsorption of B-casein in the presence of nisin (Figure 2-5). The profiles also
clearly show two distinct approaches to equilibrium — one with B-casein adsorbing by

itself and a second with B-casein adsorbing in the presence of nisin. The same result was

18

observed when the concentration of B-casein was lowered by a factor of ten (from 10'5 M
to 10'6M), with the bulk concentration of nisin maintained at lO'SM. In that case, an
increase of more than 8 times the initial fluorescence emission level was observed, with

the signal not reaching an apparent steady value during the time of the experiment

(Figure 2-6).

To evaluate whether this effect is due to factors other than the presence of nisin,
we conducted a control experiment where we allowed B-casein to adsorb at a bulk
concentration of 10‘5M until a pseudo-steady state fluorescence emission level was
attained. At that time, a solution of 1.3-casein at the same concentration was introduced
into the flow cell, to assess the extent of additional protein adsorption. While a small
change in fluorescence emission intensity was observed, the increase was negligible in
comparison to that observed in the presence of nisin (Figure 2-7). These results suggest
strongly that there is increased or “enhanced” adsorption of B-casein in the presence of

nisin.

To investigate this further, we studied the adsorption and interactions of nisin
with three other proteins: bovine serum albumin (BSA), lysozyme and fibronectin. This
particular group of proteins was chosen because of their diverse size, structure and
adsorption characteristics. The experimental protocol was similar to the one used in the
experiments conducted with B—casein and nisin. After pseudo-stabilization of the BSA
adsorption profile, a solution containing BSA and nisin, each at a bulk concentration of

IO'SM, was introduced into the flow cell. We observed little or no change in the

19

fluorescence emission level, even after 60 minutes of interaction (Figure 2-8). Therefore,
while the adsorption of B-casein (a fibrous protein of molecular weight 24 kDa) was
enhanced in the presence of nisin, there was no apparent enhancement in the adsorption

of BSA (a globular protein of molecular weight 66 kDa).

To the best of our knowledge, this level of “enhanced” adsorption of one protein
in the presence of another has not been previously reported. We should, however,
present the kinetic modeling work of Fang and Szleifer on the adsorption of an equimolar
solution containing a small and a large protein (Fang and Szleifer 2001). They showed
that the two proteins initially adsorb together but, after a short time, the rate of adsorption
of the smaller protein decreases and that of the larger protein increases, because the
surface-protein interaction is stronger for the larger protein. Fang and Szleifer’s kinetic
modeling studies were conducted at a bare interface while, in the present study, the
mixture of nisin and the competing protein arrived at an interface that already had a pre-
adsorbed layer of the larger protein. It is reasonable to expect that the interaction
between a protein and a bare (clean) surface would be much stronger than that between

the same protein and a pre-adsorbed protein layer.

While we do not know the exact mechanism for its occurrence, there are several
plausible explanations for the observation. First, nisin is a pore forming protein capable
of penetrating cellular membranes (Van Kraaij et al. 1998; Bower et al. 2001; Van
heusden et al. 2002). Therefore, one possible explanation is that, as the original protein

(ll-casein, for example) adsorbs towards its equilibrium concentration at the interface, it

20

forms a film that may mimic a cellular membrane. When nisin is introduced into the
flow cell, its interaction with the adsorbed protein film “stretches” the interface, thus
creating additional adsorption sites for the interacting proteins to adsorb. We speculate
that this “stretching” may actually be more of a type of “corrugation” of the linear
adsorbed film at the interface, which results in the creation of a larger surface area over

the same linear horizontal dimension.

Another possible explanation is that nisin causes the adsorbed protein to reorient
itself at the interface, resulting in more efficient packing (Figure 2-9). For example, some
studies have reported a reorientation of lysozyme upon initial adsorption and interaction
with the surface, following which space is created for additional proteins to adsorb at the
interface (Robeson and Tilton 1996; Daly et al. 2003). Nisin apparently accentuates this
effect, because the increases in adsorption reported by these earlier studies were not as

large as the enhanced adsorption observed in our experiments.

Other factors that could influence adsorption enhancement include protein size
and/or molecular weight, and protein type (globular versus random coil). For example,
physical protein size and/or molecular weight would be a factor in adsorption
enhancement if it is assumed that nisin creates identical pore sizes in each adsorbed layer.
Then, the smaller B-casein protein molecules would be better able to pack into the pores
than larger proteins such as BSA. In that case, we would observe more significant
enhanced adsorption for B—casein than for BSA. One could also speculate that, if the

structure and the resulting adsorbed layer influence the process, then nisin may be better

21

able to create space in the adsorbed B-casein layer (which likely adsorbs in more of a

condensed layer at the oil-water interface), than in the adsorbed albumin layer (which
adsorbs in more of an expanded or spread-out configuration at the oil-water interface)

(Graham and Phillips 1979).

To evaluate these factors, we conducted similar experiments with lysozyme in the
presence of nisin. Lysozyme is a globular protein like BSA, but smaller in size (14.3
kDa) than both BSA and B-casein. As was done in the previous experiments, a solution
containing both lysozyme and nisin (each at 10'5M) was introduced into the flow cell
after lysozyme adsorption had reached pseudo-equilibrium. Upon introduction of this
solution, we observed a slight increase in fluorescence emission intensity (Figure 2-10).
Since this increase was not as large as what was observed for B-casein, a control
experiment similar to what was done with B-casein was conducted by introducing
additional lysozyme into the flow cell after pseudo—equilibrium adsorption was achieved.
As was the case for B-casein, the slight increase observed in the control experiment is
clearly less significant than was observed in the presence of nisin (Figure 2-11). Hence,
there is an indication of “enhanced” adsorption of lysozyme in the presence of nisin
(Figure 2-10), although not as significant as that observed for B-casein. This is also

evident in an additional experiment conducted in the presence of nisin as shown in Figure

2-12.

There is evidence in the literature to suggest that lysozyme achieves a higher

surface coverage than B-casein at the same bulk molar concentration (Hunter et al. 1991).

22

Therefore, a possible consequence of the higher surface coverage is that lysozyme may
form a thicker film at the interface than B-casein. Ordinarily, one would expect the
thicker film to promote formation of more stable pores, which could lead to enhanced
adsorption. However, lysozyme is also known to form a more expanded layer at the oil-
water interface than B-casein, which forms a condensed layer (Graham and Phillips
1979). As discussed above, this may tend to reduce the possibility of enhanced

adsorption.

To further evaluate the effect of nisin on adsorbed fibrous proteins, we conducted
experiments on interactions of nisin and fibronectin, a much larger fibrous protein at 440
kDa. The results showed a fairly small enhanced fibronectin adsorption in the presence
of nisin (Figure 2-13), in comparison to the protein adsorbing by itself. In all cases
examined, no “enhanced” adsorption was observed for proteins that were larger in
molecular weight than B-casein. This would tend to support the speculation that smaller

proteins are better able to adsorb effectively into nisin-induced pores than larger ones.

To assess the effect of bulk protein concentration, we conducted experiments on
lysozyme at a concentration of 10°7M, 100 times lower than in the earlier experiments,
with the bulk concentration of nisin maintained at 10'5M. The experimental protocol was
the same as that for the earlier experiments. Upon introduction of the lysozyme-nisin
mixture, we observed very pronounced enhanced adsorption. Just as significant was the
fact that the adsorption curve for lysozyme during the initial enhanced adsorption phase

was 2.6, which is identical to the initial slope of the adsorption curve in the absence of

23

nisin (Figure 2-14).

These data on lysozyme at lower concentrations would indicate that, along with
size, bulk protein concentration may be an important factor in the mechanism for
adsorption enhancement and may, perhaps, be even more important than the structure of
the protein. To evaluate this factor, we repeated the BSA and fibronectin experiments at
much lower concentrations. The BSA experiments were run at a bulk concentration of
8x10'7M, with the nisin concentration remaining at 10'5M. At this much lower BSA
concentration, diffusion became an important factor because it took a long time to attain
pseudo-steady adsorption in a reasonable amount of time. Therefore, we introduced the
mixture of nisin and BSA into the flow cell before pseudo-equilibrium BSA adsorption
was achieved. We then compared different segments of the adsorption profile on the
basis of the initial adsorption slopes before and after introducing the mixture of proteins

into the flow cell.

An examination of the data (Figure 2-15) shows clearly that the slope of the
adsorption profile increases significantly upon introduction of BSA and nisin (0.59), in
comparison to the profile of BSA adsorbing in the absence of nisin (0.21). Thus, there is
a slight increase in the rate of adsorption in the presence of nisin, although it is not as
large as that observed with B-casein or lysozyme at the lower concentrations. While
there is clear evidence of enhanced BSA adsorption at lower concentrations in the
presence of nisin, we observed only minor adsorption enhancement in fibronectin

experiments conducted at much lower concentrations (Figure 2-16). Collectively, these

24

experiments appear to indicate that the important factors in adsorption enhancement in
the presence of nisin are the size, molecular weight and bulk concentration of the
adsorbing protein. Another factor we have not yet investigated is the possibility of

ordered multilayer formation in the presence of nisin.

Increases in fluorescence emission intensities can be attributed to a) an increase in
the amount of fluorescently tagged molecules at the interface and/or b) a change in the
fluorophore quantum yield. To evaluate the quantum yield issue, we measured the
absorption and emission spectra of both pure B-casein and a mixture of B-casein and nisin
at the same concentrations. No significant differences were observed between the two
samples. In addition, lifetime studies conducted in collaboration with Dr. Gary
Blanchard of the MSU Chemistry Department using time correlated single photon
counting spectroscopy have shown that the quantum yield is the same for surface-
associated proteins as for proteins in the bulk (V aidya 2005). Thus, the increased
adsorption is most likely due to nisin-induced interfacial interactions, rather than a

change in the quantum yield of the fluorophores.

As discussed earlier, the theory offered by Fang and Szleifer (Fang and Szleifer
2001) is also plausible. However, we believe that the interfacial energetics are very
different for the two situations. In addition, similar experiments conducted in our
laboratory on the interaction between human serum albumin (66 kDa) and fibronectin
(440 kDa) did not show any evidence of enhanced adsorption of fibronectin (V aidya and

Ofoli 2005). This appears to provide confirmation that the increased adsorption is

25

specific to nisin (and perhaps other lantibiotics), rather than the result of an interaction
between a large and a small molecular weight protein as observed by Fang and Szleifer

(Fang and Szleifer 2001).
2.5 CONCLUSIONS

We have used total internal reflection fluorescence microscopy (TIRFM) to
characterize the interfacial behavior of nisin, an antimicrobial pore-forming polypeptide.
Sequential and competitive adsorption experiments on nisin and several labeled proteins
(B-casein, bovine serum albumin, lysozyme and fibronectin) were conducted at the oil-
water interface. This group of proteins were chosen because of significant variations in
molecular weight, hydrodynamic size, and structure. A key finding was the observation
of enhanced levels of adsorption of certain proteins at the oil-water interface in the
presence of nisin. The apparent surface coverage of B-casein (at a bulk concentration of
10'5 M) in the presence of nisin was nearly three times the value observed for the protein
when it adsorbed by itself at the interface. Enhanced adsorption was also observed for
lysozyme at the same bulk molar concentration. Experiments conducted at lower bulk
concentrations resulted in even more pronounced enhanced adsorption for B-casein,
bovine serum albumin and lysozyme. We also observed adsorption enhancement for the

interaction of fibronectin and nisin, although this was not significant at any concentration.

Based on the experimental results, we propose that the significant factors
influencing enhanced adsorption in the presence of nisin are molecular weight, size of the

adsorbing protein, and the bulk concentration of proteins. While the exact reason(s)

26

and/or mechanisms for this phenomenon are still being investigated, we can reasonably
conclude that the adsorption enhancement is an effect specific to nisin and perhaps other

lantibiotics, because of the ability to form pores in membranes.

27

 

 

 
  
   
  
 
  

 

c omp uter

 

 

 

 

Haunt"

 

 

 

 

inverted microscope

 

Figure 2—1:

 

Prism

Aluminum spacer TO? 8135! slide

Bottom glass slide

Teflon tubing

| o...

F igure 2—lh

To syringe pump

 

 

 

 

Figure 2-1. Experimental set-up for TIRFM experiments

a) Optical train: PMT — photomultiplier tube; M2, M1 — mirror; F1,F2 — optical flats; C —
coatings; Sh — shutter; NDF — neutral density filters; In — monitoring beam; pb —
photobleaching beam. b) Flow cell showing holes for sample infusion and withdrawal.

28

I—n
N

 

nisin (1 O'SM)

I

H

   
   

.O
W
1

Normalized fluorescence
a
a
l

 

 

 

0.4 ~
B-caseln-FITC
(10"M)
0.2 n
O I I I I I T
0 0.25 0.5 0.75 1 1.25 1.5 1.75

Time (hours)

Figure 2-2. Equilibrium adsorption of B—casein followed by introduction of nisin.

The initial adsorption profile shows equilibrium adsorption of B—casein-FITC (10'5 M) at
an oil-water interface. Upon pseudo-stabilization of the fluorescence emission signal,
unlabeled nisin (10'5M) was introduced into the flow cell. The fluorescence level
decreased upon introduction of unlabeled nisin, indicating apparent B-casein desorption
from the interface. All solutions were introduced at a shear rate of 0.43 s".

29

 

1.2

 
  
  
   

 

 

 

1 .I
8
E 0.8 -
8
S
E
0.6 ~
‘3
E
a
g 0.4 - B—caseln (10'5M)
Z
0.2
0 | I I l l T T I I
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Time (hours)

Figure 2-3. Adsorption of nisin followed by introduction of B-casein.

Following adsorption of unlabeled nisin (10'5 M) at the oil-water interface (for a period of
about 30 minutes), B-casein-FITC (10'5M) was introduced into the flow cell. Clearly, the

prior adsorption of nisin at the interface did not preclude B-casein adsorption. All
solutions were introduced at a shear rate of 0.43 s".

30

 

 

 

 

1.4
buffer
1.2 - l.’
~ ° I «u
III III
é [III I I
E 0.8 "I I . g
g I I. ' in 10"M
2 I 'l
g M, ‘ I llllIIiIlllil l l l
[-4 B-casein-FITC 10'5M
0.2 4 “H‘s
o I I I I
0 0.5 1 1.5 2 2.5

Time (horns)

Figure 2-4. Assessment of B-casein adsorption reversibility.

The initial adsorption profile shows equilibrium adsorption of B-casein-FII‘C (10'5 M) at
the oil—water interface. After the fluorescence level stabilized, a phosphate buffered
saline solution (PBS) was introduced into the flow cell. The resulting drop in
fluorescence emission is an indication of B-casein desorption from the interface.
Following stabilization of the signal, unlabeled nisin (lO'SM) was introduced into the
flow cell to check if nisin would induce further desorption of B-casein from the interface.
However, very little additional desorption was observed. All solutions were introduced at
a shear rate of 0.43 S”.

31

 

 

 

 

2.5 - a 0"" .
o

o ,9
g -5 O.
3 2 - B-casein—FITC :10 M ’°
3 . . -5 9’
‘6 nrsrn :10 M
E.
i 1 5 ' 1.08
0
E. o
g ’
Z 9’

0-5 ‘ 1.12

B-casein—FITC : 10‘5M
0 “m? I I I I I I I
0 0.5 l 1.5 2 2.5 3 3.5 4

Time (hours)

Figure 2-5. Enhanced adsorption of B—casein in the presence of nisin.

The initial profile is for the adsorption of labeled B-casein (IO'SM) at the oil-water
interface. Upon pseudo-stabilization of the fluorescence emission level, a mixture of
labeled B-casein and unlabeled nisin, each at 10'5M, was introduced into the flow cell.
The fluorescence emission intensity increased nearly three-fold, indicating very
significant additional adsorption of B-casein (“enhanced adsorption”) at the interface. All
solutions were introduced at a shear rate of 0.43 s'l.

32

 

 

 

 

 

O
8 a o
O
O
7 - o
O
0 0
O

‘5’ 61 . o
8 . ’
g 5 d o
2
IL 0
g 4 - B-casein-FITC:10"M . . .
._ O
a nisin : 10% ,.°
g 3 '1 .0
O O
z :

2 ~ :

O
3
1 - .900 .*~.
53/ p-caseln-FITC : 105M
0
0 TI“ T I l I I
o 1 2 3 4 5 a
Tlme (hours)

Figure 2-6. Enhanced adsorption of B-casein in the presence of nisin.

The initial profile is for the adsorption of labeled B-casein (10% at the oil-water
interface. Upon pseudo-stabilization of the fluorescence signal, a mixture of labeled [3-
casein (at a concentration of 106M) and unlabeled nisin (10'5M) was introduced into the
flow cell. The fluorescence emission increased several-fold, and did not attain a steady
state value during the experiment. This is an indication of very significant enhanced

adsorption of B—casein in the presence of nisin. All solutions were introduced at a shear
rate of 0.43 S”.

33

 

1.2

 

 

-casein—FITC : 10‘5M
B ..’M Q o O . . . . .
1 ~ 9’ ’9" 0 o . 0
O
r ”"
§ 0.8 — I
3 .
O
.E.
1:. _
E 0 6
E O
6
E 0.4 -
‘6
z 0
0.2 - 5
L___B—caseln-FITC :10‘ M
0 “M I I I r I
0 0.5 l 1.5 2 2.5

Time (hours)

Figure 2-7. Control experiment for B-casein.

B-casein-FIT C (IO'SM) was initially adsorbed at the oil-water interface. Following
stabilization of the fluorescence signal, additional B-casein-FITC (10'5 M) was introduced
into the flow cell. Clearly, the increased fluorescence emission is not as significant as that
observed in the presence of nisin, indicating that adsorption enhancement is due solely to
the interaction with the polypeptide nisin. All solutions were introduced at a shear rate of

0.43 s".

34

 

 

1.2
BSA-FITC : 10‘5M + nisin : 10‘5M

. .I’o I
1.. O .9. . .

o ”.00" I... 0 9
U
E o
30.8-
0
I- o
o
E-

J
130.6
0
.53
'2
“0.4“ ’
e
Z

0'2 ‘ BSA-FITC:10'5M

 

 

0 “I“ I I I I I
0 0.5 1 1.5 2 2.5 3

Time (hours)

 

Figure 2-8. Interaction of BSA and nisin.

The initial profile shows the pseudo-equilibrium adsorption of BSA-FITC (lO'SM) at the
oil-water interface. A mixture of BSA-FITC and unlabeled nisin, each at 10'5M, was
introduced into the flow cell upon stabilization of the fluorescence level. No change in
the fluorescence emission was observed, indicating negligible adsorption enhancement
for BSA at this bulk concentration. We should note the initial overshoot in the adsorption
profile prior to adding the binary protein mixture. This type of behavior has been
attributed to orientational changes within the adsorbed layer (for example, from an end-
on to side-on configuration) by which a kinetically favored orientation is replaced by a
more stable orientation that requires more surface area per protein molecule (Wertz and
Santore 2002). All solutions were introduced at a shear rate of 0.43 s".

35

 

 

 

relaxed/extended B-casein
B-casein

nisin

relaxed/ extended nisin

_:s". ‘2... $0 W

Figure 2—9. Proposed scheme of the interaction between B-casein and nisin.

The figure on the left shows a film of B-casein formed at the interface alter pseudo-
equilibrium adsorption. We hypothesize that, when the B-casein is adsorbed at the
interface, the protein initially spreads out to occupy more surface area per molecule. We
propose that, upon introduction of nisin, the polypeptide interacts with the adsorbed [3-
casein molecules to cause reorientation of the protein molecules. This reorientation
results in more efficient packing, and creates more room for additional proteins to adsorb.

36

 

l—t
ck

 

lysozyme-FITC : lO'SM
1.2 , nisin :10'5M
I ”9"..." . . .

g 1'4 ..00099 O 0 MO.
8 9’
3 J :
S 0.8
g o
B o
:5". 0.6 -
a
E o
‘5 0.4 ~
2 o

0.2 - , A

lysozyme-FITC : 10 ’M
o
0 lb I7 I I I I I

 

 

0 0.5 1 1.5 2 2.5 3 3.5

Time (hours)

Figure 2-10. Interaction of lysozyme and nisin.

Lysozyme-FITC (10'5M) was initially adsorbed at the oil-water interface till apparent
equilibrium. Then a mixture of labeled lysozyme and unlabeled nisin, each at 10'5 M, was
introduced into the flow cell. Following this, a small increase in fluorescence emission
intensity was observed, indicating some degree of adsorption enhancement. All solutions
were introduced at a shear rate of 0.43 s".

37

 

I—a
N

 

 

 

lysozyme-FITC : 10'5M
I MM. 0 o O 0
1 T O... O . . . . .
4, o
o o
I:
3 0.8 ~ ’
g; o
i-
O
a o
=1 0.6 -
1:
0
E o
E 0.4 .
I—
o
z
o
0.2 A
lysozyme-FITC : IO‘SM
0 :;:::¢ , , , , ,
0 0.5 1 1.5 2 2.5 3

Time (hours)

Figure 2-11. Control Experiment for Lysozyme.

The initial adsorption of lysozyme-FITC (10'5M) to apparent equilibrium was followed
by re-introduction of the same solution (at the same concentration) into the flow cell.
This resulted in a very small increase in fluorescence emission intensity. All solutions
were introduced at a shear rate of 0.43 s".

38

1.4 — --—- m
lysozyme-FITC : 10'5M

1,2 - lysozyme-FITC 10'5M
nisin 10'5M I lwuoamwwm ’° ‘° “‘ 0 6..
II I I I
1" ‘fl‘. 0* A

 

 

 

 

 

 

 

 

8
I: {I‘m
0
a I
3 0 8
o - 1
5
E. l
'3 0 6 q ! run 2
g ' lysozyme-FITC:10'5M 0 lrsozyme +
g a ntstn : 10'5M nrsrn
‘5' 0.4 " I I
z A ysozynre
I ‘ control
0.2 a lysogyme-FITC : IO‘M A lysozyme +
j nisin run 2
0 - r I I f T r I
0 0.5 1 1.5 2 2.5 3 3.5 4
Time (hours)

Figure 2-12. Lysozyme adsorption.

This figure shows three lysozyme adsorption profiles from separate experiments. The
data show a larger increase in fluorescence emission in the two experiments where nisin
is present, compared to the one experiment of lysozyme adsorbing alone. The
fluorescence data have been normalized by the initial pseudo-stabilization level in each
experiment. All solutions were introduced at a shear rate of 0.43 S”.

39

 

 

 

 

12 _ Fibronectin-FITC:10'7M
nisin : 5x10'5M
LO””.””..... .
o 1 ‘ . O. 9 O
3 Q
g 0.8 - ,
= o
c:
B 0.6 — t
.5
'3' o
E
O 0.4 —
z o
0.2 J o .,
flbronectln-FITC : 10' M
o
0 ”an”
0 0.5 1 1.5 2 2.5

Time (hours)

Figure 2-13. Interaction of fibronectin and nisin.

Pseudo-equilibrium adsorption of fibronectin-FITC (10’7M) is achieved in about an hour
at the oil-water interface. Upon introduction of a mixture of fibronectin-FITC (10'7M)
and nisin (10'5 M), a slight increase in fluorescence emission was observed. All solutions
were introduced at a shear rate of 0.43 s".

40

 

2.5

 

 

 

 

2‘ o ’ ’
w’
0
g lysozyme-FITC:10'7M 0
§ 1-5“ nlsln :10"M
5‘ 2.68
E O O Q 0.”
.3 l- 0 o . «o
s .»
2
0.5 « .7
lysozyme-FITC : 10 M
l 2.04
0 “M + I I I
0 0.5 1 1.5
Tlme (hours)

Figure 2-14. Interaction of lysozyme (at a lower concentration) and nisin.

2.5

This experiment is similar to the lysozyme adsorption experiment presented earlier,
except at a lower bulk concentration. After introduction of the mixture of lysozyme-
FITC (10'7M) and nisin (lO'SM), greater than a two-fold increase in fluorescence
emission intensity was observed. This trend is very similar to what was observed at
higher molar concentrations for B-casein adsorbing in the presence of nisin. All solutions

were introduced at a shear rate of 0.43 s'l.

41

1.2

 

 

 

 

 

1» 1 _ BSA-FITC : 8x10'7M . . o 9 0 ° °
0 . . .5
a nrsrn : 10 M
g 0.3 _ I, 0.59
I- .0
o Moo
:1
E‘ 0.6 -BSA-FITC : 8x10'7M , . 0 A 011
'8
:2 0°
2 0.4 - ,. °
'5 .’
Z
0.2 ~
3.27
Ir
0 mt. I l I I I I
0 0.5 1 1.5 2 2.5 3 3.5
Time (hours)

Figure 2-15. Interaction of BSA (at a lower concentration) with nisin.

A solution of BSA-FITC (8x10'7M) was allowed to adsorb to the oil-water interface.
Since the fluorescence emission level did not stabilize even after 90 min, we introduced
the mixture of BSA-FITC (8x10‘7M) and nisin (10'5M), and observed the adsorption
slopes (indicative of the rates of adsorption) before and after introduction of the mixture.
The slope of the profile (0.59) after introduction of the binary mixture of nisin and BSA-
FITC was significantly larger than the prior slope (0.21) for BSA-FITC adsorbing by
itself. All solutions were introduced at a shear rate of 0.43 s".

42

1.2

 

 

Fibronectin-FITC : 103M
0 ..
E 1 ‘ nisin : 10'5M .0..—3W”° I
8 .0 0’
0 O
:5 0.8 - “0
0.6 r o
B .
.2‘. o
'3 0.4 r .
E o
2 0'2 I f Fibronectin-FITC :10‘8M
O
0 T I I I I I I

 

 

 

0 0.5 1 1.5 2 2.5 3 3.5
Time (hours)

Figure 2-16. Interaction between Fibronectin (at a lower concentration) and nisin.

Following pseudo-equilibrium adsorption of fibronectin-FITC (at a lower concentration
of 10'8M), a mixture of fibronectin-FIT C (10'8M) and nisin (10'5 M) was introduced.
While this apparently resulted in increased fluorescence emission intensity, the increase
was not significant. All solutions were introduced at a shear rate of 0.43 s".

43

3 DEVELOPMENT OF A PROTOCOL FOR SIMULTANEOUS OPTICAL
AND ELECTROCHEMICAL CHARACTERIZATION OF
BIOMACROMOLECULES IN TIR GEOMETRY

3.1 ABSTRACT

We describe the development of a flow cell to enable simultaneous optical and
electrochemical characterization of biomimetic interfaces. The goal was to expand the
capabilities of the TIRFM apparatus so we can critically interrogate biomimetic systems
that exhibit both optical and electrochemical responses to specific or non-specific
interactions. Modifications were made to our existing optical flow cell to incorporate the
ability to conduct cyclic voltammetry (CV) measurements using a three-electrode set-up.
The top and bottom slides in our optical flow cell were replaced with indium tin oxide
(ITO)-coated glass slides. The transparency of ITO allows us to conduct optical
experiments, while its conductivity makes it possible to conduct electrochemical

measurements on the same sample.

The feasibility of the simultaneous-optical electrochemical set-up was
tested using the dye resorufin. On potential scans towards increasing negative voltages
(400mV to -400mV), the fluorescent resorufin was reversibly reduced to non-fluorescent
dihydroresorufin, coincident with a decrease in fluorescence emission. Similarly, upon
scanning towards increasing positive voltages (-400mV to 400mV), the dihydroresorufin
was reversibly oxidized to resorufin, resulting in increased fluorescence emission. The

same relationship between fluorescence emission and potential was observed in all

44

subsequent cycles, confirming that the optical-electrochemical set-up serves its intended
function. This experimental tool can also be used to characterize independent optical and

electrochemical interactions taking place in a given sample at the biomimetic interface.

3.2 INTRODUCTION

Our lab has focused on using total internal reflection fluorescence microscopy
(TIRFM) as the primary tool to optically characterize bio-macromolecular interactions at
liquid-liquid and solid-liquid interfaces. The objective of this study was to expand the
capabilities of this technique to enable characterization of more complex systems such as
biomimetic interfaces that can be used as biosensors. A biosensor is a device that
converts specific recognition events between a receptor and a target analyte into a
measurable signal. It serves as a reliable miniaturized device with the ability to perform
sophisticated detection of analytes in areas such as biology, nanotechnology,
environmental quality, and food technology, and in rapid diagnostic assays. To design
such systems, it is necessary to accurately characterize the associated biomacromolecular

interactions at the biomimetic interfaces.

In the first part of this thesis, we concentrated on understanding fundamental
protein interactions between nisin and selected proteins at the oil-water interface. In this
section, we will outline the development of an experimental set-up that enables
simultaneous use of two important characterization techniques: TIRFM and cyclic
voltammetry (CV). Electrochemical/amperometric biosensors provide continuous and

accurate measurements of the concentration of analytes in an electrical format. The need

45

for sensitivity, specificity, speed and accuracy of measurements has triggered interest in
developing electrochemical sensors as diagnostic tools in the food and pharmaceutical

industries (Prodromidis 2002).

The ability to simultaneously track biomacromolecular interactions as well as
electron-transfer reactions in-situ represents a very powerfiil tool for analyses of complex
biological systems, and would aid in the development and characterization of biosensors.
TIRFM has proven to be well-suited for Optical interrogation of interfacial biomolecular
interactions. We believe the ability to also conduct electrochemical experiments with the
same apparatus would help us identify independent events taking place in these complex
systems. In cases where a single species produces both electrochemical and optical

signals upon interaction, we will be able to obtain corroborative evidence of that event.

3.2.1 Optical characterization

3.2.1.1 Total Internal Reflection Fluorescence Microscopy (TIRFM)

A brief introduction to TIRFM was presented in Chapter 2, and will not be
repeated here. The only additional information is that the evanescent wave for the
experiments in this section was configured to a penetration depth of about 75nm. We
achieved total internal reflection (TIR) at an interface between an indium tin oxide (ITO)-
coated glass slide and the sample solution in the flow cell. The transparency of ITO
allows us to conduct optical experiments, while its conductive surface allows for its use

as an electrode for electrochemical measurements.

46

3.2.2 Electrochemical characterization

3.2.2.1 Cyclic voltammetry (CV)

CV is a useful and versatile technique used to study electroactive species. Its ease
of use enables application in various fields of electrochemistry, inorganic chemistry,
organic chemistry, and biochemistry (Kissinger and Heineman 1983). It can be used to
conduct mechanistic as well as kinetic studies of redox behavior of a biological or a
chemical compound over a wide potential range. There have been numerous studies of
biosensors utilizing CV to monitor various enzyme-catalyzed redox reactions. Common
enzymes used for biosensors include alcohol dehydrogenase (Mullor et a1. 1996; Rao et
al. 1999), lactate dehydrogenase (Chaubey et al. 2000), glucose oxidase (Motta and
Guadalupe 1994; Willner et al. 1997; Wang et al. 1998), cytochrome c oxidase (Li et al.
1996) and horseradish peroxidase (F erri et al. 1998; Munteanu et al. 2003). CV has also
been used to study DNA-mediated electron transfer (Napier et al. 1997; Palecek et al.

1998; Kelley et al. 1999; Armistead and Thorp 2000).

A CV measurement involves cycling the potential of an electrode in solution, and
measuring the resulting current (Kissinger and Heineman 1983; Bard and Faulkner;
Kissinger and Heineman 1984). The electrode is the surface/interface where the
biological reaction takes place. In our set-up, we use a three-electrode configuration
comprising a working electrode, a counter (or auxiliary) electrode and a reference
electrode. This configuration is more advantageous than the traditional two-electrode set-

up. The traditional configuration does not allow for precise control of the external

47

applied potential (which is required in cyclic voltammetry measurements), because of
potential (ohmic) drop due to solution resistance. Better control of the external applied
potential is achieved by controlling the potential of the working electrode with respect to
the reference electrode. Polarization of the counter electrode, which completes the
current-measuring circuit, is required for cyclic voltammetry. This is not available in the
two-electrode set-up (Bard and Faulkner 2000; Bioanalytical Systems 2005; Carnes and
Wilkins 2005). The function of each electrode in the three-electrode system is briefly

described below (Bard and Faulkner 2000).

The working electrode (WE) is the test electrode where the reaction that generates
the oxidation or reduction takes place. This is the only electrode where the reaction
occurs. Many different working electrodes can be used, such as glassy carbon, boron-
doped diamond, gold, and indium tin oxide. These working electrodes can be further
tailored or modified, based on the requirements of the desired application. For example,
to study reactions catalyzed by enzymes embedded in a bilayer lipid membrane (BLM),
the enzyme incorporated into either a supported or tethered BLM can be deposited on the
working electrode, and the solution in contact with the modified electrode can contain the

analyte of interest.

The counter electrode acts as a source or sink for electrons, so that the current can
be passed through the external circuit of the cell. An ideal counter/auxiliary electrode is
one which does not produce any substances by electrolysis, because these products can

reach the working electrode and interfere with its fimction.

48

The reference electrode is one whose potential stays constant, so that it can be
used as a standard against all measured potentials in the three-electrode cell (Faulkner
1983; Kissinger and Heineman 1983; Mabbott 1983; Van Benschoten et a1. 1983; Bard
and Faulkner 2000; Kissinger and Heineman 1984). The reference electrode must,
therefore, have the property of ideal nonpolarizability. Some common reference
electrodes are the saturated calomel electrode (SCE) and the silver/silver chloride

electrode (AyAgCl).

In a CV experiment, the potential is maintained between the working and
reference electrode, while the current is measured between the working and the counter
electrode. The potential applied between the working and the reference electrode is
considered the excitation signal. The potential in our experiments is a linear scan that
exhibits a triangular waveform. The potential is swept between two values (called the
switching potentials). Following the sweep in one direction, the scan direction is
reversed and the potential is scanned back to the original starting value. The parameters
of interest are the initial and final potentials, the scan direction, the scan rate, and the
number of scans, all of which can be set by the Operator. Usually, a window is chosen to
include the potential at which the analyte of interest is oxidized and reduced. The current
generated at the working electrode as a result of cycling the potential is recorded, and a
cyclic voltammogram is generated. The CV obtained is thus a plot of current versus

potential.

The current produced depends on two steps in the overall process: the movement

49

of electroactive material to the surface and the electron transfer reaction. The exponential

dependence of the electron transfer rate constant on the potential is given by

 

kf=k0exp(-:’;F(E—EO')) (3-1)

where k" is the standard homogenous electron-transfer rate constant, It is the number of
electrons transferred, R is the universal gas constant, T is the absolute temperature in

Kelvin, and a is the transfer coefficient.

A typical cyclic voltammogram is shown in Figure 3-1. The dependence of the
current on the electrochemical reaction during a CV experiment will be explained in
detail in section 3.3.3 for the ferricyanide system. Some of the important parameters of
interest in a cyclic voltammogram are the anodic peak current (ipa), the cathodic peak
current (ipc), the anodic peak potential (Em), and the cathodic peak potential (Epc) (Figure
3-1). The peak current obtained in a reversible electrochemical process is directly
proportional to the concentration of the analyte and is given by the Randles-Sevich

equation below:

ip = 2.69x105ACJn3Dv (3.2)

where ip is the peak current (in amperes), n is the electron stoichiometry, A is the
electrode area (cmz), D is the diffusion coefficient (cmZ/sec), C is the concentration

(mol/cm3) and u is the scan rate (V /sec).

50

For a reversible system, the applied external potential controls the ratio of the
oxidized to the reduced species at the electrode surface, according to the Nernst equation:
. 0
E = 15° + 2.303 IR—T-IlogU (3.3)
nF [R]
where E is the external applied potential, F is the Faraday constant, and [O] and [R] are

the concentrations of the oxidized and reduced species, respectively. The Nemst

equation is valid only at equilibrium.

An electrochemically reversible couple is one in which both the species (oxidized
and reduced) rapidly exchange electrons with the working electrode. The formal

reduction potential (E0 ') for a reversible couple is given by

_ Epa + Epc
2

50' (3.4)

where Epa is the peak potential at the anode and Epc is the peak potential at the cathode.

The number of electrons (n) cycled in the electrode reaction can be calculated

from the separation of the cathodic and the anodic peak potential, and is given by

0.059
AEp = Epa -Epc =

 

(3.5)
n

Thus, an ideal reversible one-electron transfer process should have a peak

separation of 59mV.

51

Slow reaction kinetics leads to slow electron transfer. This is an indication of an
irreversible reaction, and results in increased peak separation (between the cathodic and
the anodic peaks). If the peak separation is such that the anodic and the cathodic peaks
do not overlap on the potential axis, then the system is classified as irreversible. Usually,
reactions will be irreversible if chemical bonds break or if there is a loss of substituents in

solution (Mabbott 1983).

The concentration of the reactant in the bulk solution can be determined once we
know the peak current. However, CV is not ideally suited for that task; techniques such
as differential pulse voltammetry are better equipped for these types of measurements.
CV is best used to analyze homogeneous chemical reactions that are coupled to the
electron transfer process (Mabbott 1983), because it gives qualitative information about
electron transfer processes. One can quickly determine the formal potential, the number
of electrons transferred in the process, and the anodic and the cathodic peak potentials

using CV.

3.3 EXPERIMENTAL SET-UP

3.3.1 TIRFM flow cell

The flow cell used for our TIRFM studies was developed by a previous graduate
student. A detailed description can be found in the literature (Gajraj and Ofoli 2000a).
Briefly, the flow cell was formed from two glass microscope slides separated by a 900um

aluminum spacer. An oval channel (2.34 inch x 0.73 inch) cut into the aluminum spacer

52

formed the flow channel. Teflon tubes (TTF-l30-10M, Valco Instruments Co, Inc.,
Houston, TX) inserted into two holes drilled in the bottom slide allowed for the
simultaneous infusion and withdrawal of solution into and out of the flow cell (Figure
3-2), using a Harvard apparatus infusion/withdrawal syringe pump (Model 551382,

Harvard Apparatus, South Natick, MA).

3.3.2 Modifications to TIRFM set-up to conduct optical and electrochemical

experiments (first generation)

To conduct both optical and electrochemical studies in the TIRFM apparatus, we
made several modifications to our flow cell. The top glass slide was replaced by an
indium tin oxide (ITO)-coated glass microscope slide (CG-4llN-SIO7, Delta
Technologies Ltd., Stillwater, MN). Since the aluminum spacer used to form the flow
channel in the TIRFM flow cell is conductive, it was replaced with a rubber gasket of

about 800nm thickness.

We also modified our flow cell to accommodate the three-electrode set-up
described earlier. To incorporate the working and the counter electrodes on the ITO
slide, we adopted a design similar to that described by Asanov et al. (Asanov et al. 1998).
The top ITO slide was divided into three conductive bands insulated fiom each other by
etching the ITO layer in between the bands, using a strong acid such as 12M HCl. The
bands required to be conductive were masked with photoresist during the etching process.
Removal of the ITO coating in the etched bands was confirmed by ensuring a very high

resistance across any pair of conductive bands. The central region was used as the

53

working electrode (Figure 3-3b). The area of the working electrode could be controlled
by the size of the central region. Either of the other two bands could be used as the
auxiliary/counter electrode. Thin electrical wires attached to the ITO surface with
conductive epoxy (CW24OO Part A and B, Chemtronics, Kennesaw, GA) provided
electrical contact. The epoxy was allowed to cure for 15 minutes in an oven at 75°C.
This was done to ensure that the wires remained firmly attached to the electrode surface,

because weak contacts lead to increased resistance.

A silver/silver chloride (Ag/AgCl) reference electrode (Bioanalytical Systems Inc.
(BASi), MF-2078, West Lafayette, IN) was used for these measurements. The reference
electrode was placed in the beaker into which the exit solution emptied. The electrical
contact between the reference electrode and the working electrode was maintained as
long as there was solution in the flow cell, the exit tube, and the exit beaker in which the
reference electrode was placed. Potassium ferricyanide (K3Fe(CN)6) was used as a test

electroactive species to evaluate how the electrochemical cell functioned.

3.3.2.1 Materials and Methods: Ferricyanide Experiments

Potassium ferricyanide (K3Fe(CN)6) was purchased from Mallinckrodt
Laboratory Chemicals (Phillipsburg, NJ). Potassium nitrate (KNO3) was obtained fiom
the J T Baker Chemical Company (Phillipsburg, NJ). A lOmM aqueous solution of

K3Fe(CN)6 in 0.1 M KNO3 was used as the working solution.

The flow cell was assembled with a rubber spacer sandwiched between the top

54

ITO coated slide (housing the working and the counter electrode) and a glass bottom
slide. The prism was optically coupled to the top ITO-coated slide using an immersion
oil of refractive index 1.51 (Type A, 16482, Cargille Laboratories, Cedar Grove, NJ).
The whole assembly was tightened down with four screws. The potassium ferricyanide
solution was infused into the flow cell at a flow rate of 0.36 ml/min, and a BASi CV-
50W Voltammetric Analyzer (Bioanalytical Systems, BASi, West Lafayette, IN) served
as the potentiostat for the experiments. The leads from the WE, AE and the RE were
connected to the potentiostat. Prior to the start of each experiment, the flow cell was
visually inspected to make sure there were no air bubbles or leaks. The software program
obtained with the CV-SOW voltammetric analyzer (CV-50W) was used. The parameters
used to acquire the cyclic voltammogram for the ferricyanide system are given in Table

3-1.

3.3.3 Results and Discussion: Ferricyanide Experiments

The current versus voltage curve obtained with the IOmM potassium ferricyanide
test solution in the modified flow cell is shown in Figure 3-4. This result was compared
to the cyclic voltammogram obtained in a conventional set-up with all three electrodes
placed together in the same beaker (Figure 3-5). The parameters used for the scans were
the same in both cases. In the conventional set-up, a platinum working electrode (MF-
2013, Bioanalytical Systems (BASi), West Lafayette, IN), a platinum counter electrode,
and an Ag/AgCl reference electrode (catalog # MF-2078, Bioanalytical Systems Inc.

(BASi), West Lafayette, IN) were used.

55

The CV scan started at 700 mV in the direction of decreasing voltages, at a rate of
IOOmV/sec. When the potential was sufficiently negative to reduce Fe(lII), a cathodic

current was generated due to the electrode reaction given below:
Fe’” (CN)§' + e' ——> Fe” (CN)§‘ (3.6)

The cathodic current increased rapidly until the concentration of Fe"'(CN)63' at
the electrode surface diminished substantially, causing the current to peak (point “a” in ‘
Figure 3-5). The current decayed as the solution surrounding the electrode was depleted
of Fe'"(CN)63' due to its electrolytic conversion to FeH(CN)64'. The scan direction was
reversed at 0 mV, and a backward scan was initiated. The potential was still sufficiently
negative to reduce FeI"(CN)63', so the cathodic current continued, even though the
potential was now being scanned in the direction of increasing voltages. Oxidation of the
Fe"(CN)64' accumulated near the electrode started when the electrode became a strong

oxidant:
Fe” (cmg' —> 1%” (CN)§‘ + e' (3.7)

This resulted in production of the anodic current, which increased until the
concentration of FeII(CN)(,4' diminished, causing the current to peak (point “b” in Figure
3-5). The current then decayed as the solution surrounding the electrode was depleted of
Fe"(CN)64. The first cycle was completed when the potential reached 700mV. This

cycle was repeated three times.

56

To assess the working of our modified flow cell, the two voltammograms have
been combined in Figure 3-6 for ease of comparison. The following observations can be

made:

- The current range in the voltammogram obtained in our modified first generation
TIRFM-CV set-up (about 500uA) was approximately 50 times higher than that
obtained in the conventional set-up (about lOpA). This is simply because the
electrode area was much higher in the flow cell than in the conventional set-up (as

discussed earlier, current is directly proportional to electrode area).

- The maximum cathodic current in the traditional set-up was obtained at 218mV
while the anodic current peaked at 292mV. The difference in potential between
the anodic and cathodic peaks was about 74mV. As discussed earlier, a
difference of about 59mV would have indicated a completely reversible system.
In comparison, the maximum cathodic current (peak) in the TIRFM apparatus was
obtained at 194mV, with the anodic current peak at 309mV under constant flow
conditions. Under stopped flow conditions, the maximum cathodic current was
obtained at l90mV with an anodic current peak at 3l4mV. The difference in
peak potentials was 115mV for the experiment conducted under continuous flow
conditions, and 124mV for the stopped flow experiment. The greater departure
from the truly reversible case may be due to a number of factors, including
increased solution resistance, slow electron transfer, wires not providing good

electrical contact, undetected air gaps/bubbles, leaks in the flow cell, or the

57

placement of the reference electrode far away from the working electrode. These

factors are addressed below.

3.3.4 Problems with first generation optical — electrochemical set-up and resulting

modifications (second generation experimental set-up)

A common problem in most CV experiments is ohmic potential drop (or IR drop)
due to solution resistance. IR drop can cause decreased current magnitudes, increased
peak separation and shift in peak potential. The majority of the IR drop generally occurs
along the electrode surface, rather than normal to the electrode surface (Kissinger and
Heineman 1984), because the current between the working electrode and the auxiliary
electrode passes along the channel and adjacent to the electrode. As a consequence of
this, the potential is not uniform between all points on the face of the electrode surface
and the solution. This problem can be overcome by placing the auxiliary electrode right
across the working electrode, so the charge travels perpendicularly between the two
electrodes. In this case, the potential will be uniform throughout the electrode surface

and the uncompensated IR drop will be small.

With this consideration in mind, we modified the design of our first generation
flow cell as follows: instead of having the working and the counter electrode next to each
other on one slide, we placed them right across fi'om each other. This was achieved by
using the top ITO-coated glass slide as the working electrode, and the bottom slide (also
an ITO-coated glass slide) as the auxiliary electrode (Figure 3-7b). Ohmic potential drop

can also be decreased by using a high concentration of supporting or fully dissociated

58

electrolyte, thereby increasing the conductivity of the solution and decreasing the
resistance. Therefore, in all experiments a high concentration of supporting electrolyte

(lOOmM) was used.

The first generation system was subject to an additional problem of
uncompensated IR drop or solution resistance, because the reference electrode had been
placed far away from the working electrode. Placing the working and reference
electrodes in close proximity minimizes the voltage errors resulting from ohmic loss
(Mabbott 1983). To achieve this, we modified the set-up to incorporate the reference
electrode as close to the working electrode as our existing flow cell layout would permit.
A “no leak” reference electrode (EE009, ESA, Cypress Systems, Lawrence, KS) was
purchased and inserted into a flangeless ferrule (P342, Upchurch Scientific, Oak Harbor,
WA) fitted in a T-joint (P-712, Upchurch Scientific, Oak Harbor, WA). Teflon tubes
were connected to the remaining two ends of the T-joint. One end of the teflon tube was
connected to the inlet syringe, while the other end was flanged to stay in the bottom slide
of the flow cell. To prevent formation of air gaps/bubbles in the tube, the flow cell was
filled with solution at the beginning of each experiment, prior to insertion of the reference

electrode into the T-joint (Figure 3-7).

Air gaps in the tubes caused a disruption/break in electrical contact between the
reference electrode and the solution in the flow cell. By conducting control experiments
in the flow cell and in the conventional three-electrode set-up, we determined that when

the reference electrode was not in contact with the solution in the flow cell (due to air

59

gaps or bubbles), it resulted in blackening of the ITO counter electrode (due to cathodic
polarization). This blackening damaged the ITO coating, resulting in either a decrease or
a loss of conductivity of the ITO layer. To ensure that the reference electrode always

stayed in contact with the solution, the opening in the T-j oint was expanded.

Initially, wires epoxyed to the electrode were used to provide electrical contacts.
However, when the flow cell was assembled and tightened with the wires in place, we
noticed air gaps and some leakage of solution from the flow cell. Leakage of solution
and air gaps increased solution resistance and, in some instances, resulted in blackening
of the ITO electrode (cathodic polarization). To minimize these problems, we replaced
the wires with copper foils. We noticed that wetting of the copper foil during
experiments always resulted in leaching of the copper into the solution. Aluminum foils
were, therefore, used instead to provide electrical contact. All these modifications

resulted in the second generation of our optical electrochemical set—up (Figure 3-7).

3.3.4.1 Problems encountered with the second generation experimental set-up

Using aluminum foils as electrical contacts worked to prevent major leaks of
solution out of our flow cell. However, at the end of a failed experiment (blackening of
the electrode), we noticed that the aluminum foils and the leads were slightly wet (even
though no major leaks were detected during the experiment). Wetting of the leads and
the foil resulted in increased contact resistance. To prevent this problem, we tried
isolating the foil by encasing it in parafihn or double-sided tape, to prevent the solution

from touching the foil. This did not work, because we were limited by how much

60

pressure we could apply to the flow cell, since tightening the flow cell too hard most
often resulted in breaking the bottom slide.

We were also limited by having to work with the existing flow cell used for
TIRFM experiments. Hence, we decided to epoxy the whole assembly, connecting the
top slide, bottom slide, aluminum foils and the spacer into one unit. Epoxying the flow
cell assembly prevented all leaks completely. The drawback is that, at the end of an
experiment, the whole set-up had to be immersed in acetone to pry the slides apart and, in
some cases, the slides were rendered unusable if the epoxy could not be completely
removed. We recognize that epoxying slides together is not the ideal solution, so we are
in the process of designing a flow cell that would be compatible for both optical and

electrochemical experiments.
3.3.4.2 Proof-of-concept of optical/electrochemical set-up

To test the simultaneous TIRFM/CV set-up, we needed to use a compound that
was both electrochemically active and could be optically tracked in either of the redox
states (oxidized or reduced). We had originally planned to study the redox reaction
catalyzed by the enzyme secondary alcohol dehydrogenase (sADH). The cofactor for
this reaction is nicotinamide adenine dinucleotide (NAD). NAD undergoes the following

redox reaction to convert to NADH:

sADH
CZHSOH + NAD <:> CH3CHO + NADH + H+ (3.8)

The reduced form of the cofactor NADH is fluorescent upon excitation at 340nm,

61

with an emission maximum at 450nm. Therefore, the conversion of NAD to NADH
during the enzymatic reaction can be monitored both optically and electrochemically.
Upon production of NADH, we would expect to see fluorescence emission at 450nm,
coincident with the presence of a cathodic peak in the cyclic voltammogram. We
characterized the reversibility of adsorption of sADH onto ITO surfaces, and NAD
attachment to ITO using a boronic acid linker molecule. However, we were not able to
do simultaneous optical and electrochemical characterization of this system because of
the need for a UV light source, which we do not have in our laboratory at this time. This

study remains as planned future work to be accomplished in our laboratory.
3.3.5 Introduction to resazurin

To assess the feasibility of our set-up, we selected a chemical dye, resazurin,
which undergoes a reduction reaction to produce a form (resorufin) that is fluorescent.
Resazurin (commercial name Alamar Blue) is a versatile chemical dye which is widely
used in resazurin reduction assays. For example, the resazurin reduction assay has wide
application in cell toxicity work to test the viability of cells (O'Brien et a1. 2000; Dayeh et
al. 2004; Riss and Moravec 2004). Resazurin has also been used to monitor bacterial and
yeast contamination of milk (Reinheimer and Demkow 1990), to test mitochondrial
function (Zhang et al. 2004), to test quality of sperm by colorimetry (Wang et al. 1998;
Zrimsek et al. 2004), to probe environmental redox reactions (Tratrryek et al. 2001), and
to differentiate between bacterial respiratory types (Karakashev et al. 2003). It serves as

an indicator in various enzyme assays (Chapman and Zhou 1999; Maeda et al. 2001) as

62

well as to monitor NADH turnover (Candeias et al. 1998; Batchelor and Zhou 2002).
Resorufin (produced in the resazurin reduction assay) is fluorescent and can be used as a

colorimetric indicator of the progress of the reaction under study.

Resazurin undergoes a two-step electrochemical reaction. The first step is an
irreversible reaction in which resazurin is converted to resorufin upon reduction (Figure
3-8a). The reduced product, resorufin, is fluorescent and thus serves as a measure of
monitoring the reaction optically. The substrate resazurin is purple in color, while the
product resorufin is pink in color. In the second step of the reaction, resorufin is reduced
reversibly to dihydroresorufin (Figure 3-8b). While the substrate in the second reaction,
resorufin, is pink and fluorescent, the dihydroresorufin is colorless and non-fluorescent.
The optical properties of the dyes have been tabulated in Table 3-4. The oxidation and
reduction products generated can be monitored electrochemically at the oxidation and

reduction potentials listed in Table 3-3.

Thus, this reaction can be monitored both optically and electrochemically. When
resorufin is generated, we should expect to see an increase in fluorescence emission
coincident with a cathodic peak at about ~160mV. The maximum excitation wavelength
of resorufin is 571nm. Since a 571nm light source was not available, we used a 532.5 nm
source for excitation. The excitation efficiency at this wavelength is about 56%. While
we recognize that this efficiency is less than ideal, this source still worked well as an

appropriate excitation light source.

63

3.3.5.1 Materials and methods: Resazurin Experiments

Resazurin (R12204) and resorufin (in kit V23110) were purchased from
Molecular Probes (Eugene, OR). A 0.02mM solution of resazurin and a 0.002mM
solution of resorufin were made in pH 7.4, 100mM phosphate buffer (PBS) prepared
according to the protocol outlined in Chapter 2 (under Materials and methods). All
solutions contained 100mM ultra pure sodium chloride (NaCl, SX0420-1, EMD

Chemicals Inc., Gibbstown, NJ).

The instrumentation used for the optical experiments was similar to that in
Chapter 2, except for a few minor modifications. A 32x microscope objective (440851,
Achroplan, Zeiss, Thomwood, NY) was used for fluorescence emission measurements.
A green laser module (DPSS 60, Beam of Light Technologies, Inc., Clackamas, Oregon)
at 532nm was used as the excitation source. The filter set (45 13 66, Zeiss, Thomwood,
NY) used for these experiments comprised three components:

- an excitation bandwidth filter that allows light only between 510-560nm to pass
through

- a beam splitter that directs light below 580nm up into the microscope objective,
and

- an emission filter that allows light above 590 nm to pass through to the

photomultiplier tube.

3.3.6 Experimental details: Resazurin Experiments

64

A flow cell identical to the epoxyed second generation flow cell described earlier
was used for these experiments. A program written in Labview was used as the data
acquisition sofiware (see Figure 5-2 in Appendix). The parameters used for the runs with

resazurin as the working solution are given in Table 3-2.

3.3.7 Results and Discussion: Resazurin Experiments

A 0.02mM resazurin solution in 0.1M PBS solution was introduced into the flow
cell at a flow rate of 0.36 mI/min. The reference electrode was inserted into the T-joint
after the solution flowed past it, and all three electrodes were connected to the
voltammetric analyzer. Once the flow cell was filled with the resazurin solution, the
labview data acquisition program for collection of fluorescence emission intensities
(Figure 5—2 in the appendix) and the CV-50W program for the collection of

electrochemical data were started simultaneously.

The cyclic voltammogram is shown in Figure 3-9. The potential was cycled three
times between 400mV and -400mV, at a scan rate of 30mV/sec. During scans towards
increasing negative voltages, resazurin was reduced to resorufin at about the reduction
potential for this irreversible couple (-160mV) (shown as reaction “a” in Figure 3-8).
Further along the same scans, the resorufin produced was reduced to dihydroresorufin at
about the reduction potential for this reversible couple (-277mV) (reaction “b” in Figure
3-8). In the reverse scan towards increasing positive voltages, the dihydroresorufin was

oxidized back to resorufin at about -23 8mV.

65

The fluorescence emission data are shown in Figure 3-10. The observed increase
in fluorescence emission was due to the production of resorufin by either reduction of
resazurin or oxidation of dihydroresorufin. A decrease in fluorescence was a result of the

reduction of resorufin to dihydroresorufin.

To illustrate how the simultaneous optical-electrochemical system functions to
provide corroborative data, the fluorescence versus time and potential versus time data
have been combined in one plot (Figure 3-11). The potential scan was started at 400mV
towards increasing negative voltages (400mV to -400mV) at a rate of 30mV/sec. During
the first scan, we observed an increase in fluorescence emission at approximately
-160mV. This increase is due to the reduction of resazurin to resorufin (which is
fluorescent). When the potential was scanned further towards increasing negative
voltages (400mV to ~400mV), the fluorescence emission dropped at about —277mV as a
result of the reduction of the fluorescent product resorufin to non-fluorescent
dihydroresorufin. This reaction occurred only in scans following the formation of
resorufin during the first scan. When the direction was reversed to scan towards more
positive voltages (-400mV to 400mV), an increase in fluorescence emission was
observed at about -—238mV, a result of the oxidation of non-fluorescent dihydroresorufin
to fluorescent resorufin. Thus, after the first scan, observed increases in fluorescence
emission could be the result of either reduction of resazurin to resorufin or oxidation of
dihydroresorufin to resorufin. The data appear to confirm that our TIRFM-CV set-up

works properly.

66

To further evaluate the system, we conducted a similar experiment with resorufin.
After the flow cell was filled with 0.002mM resorufin in 0.1M PBS solution, the
collection of fluorescence and CV data was started. The potential was scanned ten times
to obtain CV data (Figure 3-12). The reduction of resorufm to dihydroresorufin occurred
at about -277mV on scanning towards increasing negative voltages. The oxidation of
dihydroresorufin back to resorufin occurred at about ~238mV on scanning towards

increasing positive voltages. The CV peaks are not pronounced as explained earlier.

The fluorescence emission data were collected for the same time period as the
CV scans and are shown in Figure 3-13. A decrease in fluorescence was observed when
resorufin was reduced to dihydroresorufin, while an increase in fluorescence emission

occurred during the reverse redox reaction.

To illustrate how the simultaneous optical-electrochemical system functions to
provide corroborative data, the plot of potential versus time and fluorescence emission
versus time have been combined in Figure 3-14. The fluorescence emission intensity
goes through cycles as the applied potential is cycled. In the first scan towards increasing
negative voltages (400mV to -400mV), the fluorescence emission intensity decreased at
about —277mV, as fluorescent resorufin was reduced to non-fluorescent dihydroresorufin.
The same trend was observed in all subsequent potential scans towards increasing

negative voltages.

As the scan direction was reversed towards increasing positive potentials, the

dihydroresorufin was oxidized back to resorufin at about —238mV and an increase in

67

fluorescence emission was observed, due to the production of fluorescent resorufin. The
same trend was observed in all subsequent scans towards increasing positive voltages.
Thus the CV and fluorescence data provide confirmation of the occurrence of the same
event: the production of non-fluorescent dihydroresorufm during the reduction cycle, and
resorufin during the oxidation cycle. These data confirm that our modified second
generation TIRFM-CV set-up works as expected, and is capable of simultaneously using

two different characterization techniques on a single sample.
3.4 CONCLUSION

We have modified our TIRFM flow cell to incorporate cyclic voltammetry (CV),
an important characterization technique for monitoring electron transfer reactions. We
have tested the TIRFM-CV flow cell to demonstrate its feasibility, using fluorescent
resorufin which, upon reduction, is converted reversibly to non-fluorescent
dihydroresorufin. The production of the reduced product (dihydroresorufin) upon
scanning towards increasing negative voltages was confirmed by an accompanying
decrease in fluorescence emission. Similarly, the production of fluorescent resorufin by
oxidation of dihydroresorufin upon scanning towards increased positive voltages

coincided with an increase in fluorescence, as expected.

This experimental set-up provides a very powerful tool for characterization of
complex biomimetic interfaces for use as biosensors. Incorporating the additional
characterization technique has two advantages: it can provide corroboration of the same

reaction in cases where a single species generates both an optical and an electrochemical

68

signal, and also enable characterization of one sample undergoing different interactions at

the biomimetic interface.

3.5 FUTURE WORK

With the current flow cell design, we have to dip the slides into acetone at the end
of each experiment to pry apart the slides connected by epoxy. It would be beneficial to
build a leak-proof flow cell that can be easily assembled before any experiment and
disassembled after every experiment, with sufficient space to put in the electrical
contacts. The versatility of the modified TIRFM/CV set-up can be improved by using
other transparent electrodes such as boron-doped diamond and transparent gold
electrodes. In addition to TIRFM, we can use other optical techniques for
characterization, such as fluorescence recovery after pattern photobleaching (FRAPP).
We can also use the same flow cell with other electrochemical techniques such as time-

based voltammetry and chronoamperometry.

69

Table 3-1. Input parameters used to obtain cyclic voltammetry measurements for the
fenicyanide experiments.

 

 

 

 

 

High (mV) 700
Low (mV) 0

Initial (mV) 700

Scan rate (mV/sec) 100
# of scans 6

 

 

 

 

70

Table 3-2. Input parameters used to obtain cyclic voltammetry measurements for the
resazurin and resorufin experiments.

 

 

 

 

 

 

 

 

High (mV) 400
Low (mV) -400
Initial (mV) 400
Scan rate (mV/sec) 30
# of scans 6 (resazurin experiments)
10 (resorufin experiments)

 

71

Table 3-3. Electrochemical properties of the dyes resazurin and resorufin (Maeda et al.
2000; Maeda et al. 2001).

 

Dye Electrochemical properties

 

Oxidation potential (mV) Reduction potential

(mV)

 

Resazurin -203mV*

(-160mV vs. Ag/AgCl)

 

Resorufin/dihydroresorufin -28lmV* -320mV*

(-238 mV vs. Ag/AgCl) (-277mV vs. Ag/AgCl)

 

 

 

 

 

* Glassy carbon working electrode, saturated calomel reference electrode and
platinum working electrode.

72

Table 3-4. Optical properties of the dyes resazurin and resorufin (Maeda et al. 2000;
Maeda et al. 2001).

 

 

 

 

 

Dye Color Optical Properties
Resazurin Purple Not fluorescent
Resorufin (in pH Pink Excitation Emission
7.4 buffer) maximum 563nm Maximum 587m
Dihydroresorufin Colorless Not fluorescent

 

 

 

 

 

73

 

 

 

 

 

 

 

 

 

 

x BF”
3. 1pc
'5 """"""""
2 """""""
a - ,-
o ............. “I. ...... I
t/fl— .
lpa
Decreasing potential
b Epa
Potentlal (mV)

Figure 3-1. A typical cyclic voltammogram.

A cyclic voltammogram is a plot of current versus potential. The applied potential is the
excitation signal and the current produced is the response signal. The parameters of
interest in a voltammogram are the anodic peak current (ipa), the cathodic peak current
(ipc), the anodic peak potential (Em) and the cathodic peak potential (Epc).

74

Prism

 

6>6c

 

Aluminum spacer TOP 81355 Slide

L

 

Bottom glass slide

Teflon tubing

To syringe pump

 

 

Q 0 O-ring

 

 

 

 

Figure 3-2. Diagram of flow cell used for TIRFM characterization experiments.

It is made up of two glass microscope slides separated by a 900nm thick aluminum
spacer. An oval channel cut into the aluminum spacer and lined with a rubber o-ring
forms the flow channel. Teflon tubes inserted into holes drilled into the bottom slide
were used to infuse and withdraw solutions into and out of the flow cell.

75

 

Laser Prism Potentiostat

ITO Slid\ / I
L/ V ...................
spacer /l T 1 Wire from CE

Bottom slide syringe
pmnp

 

 
 
 
 

 

 

 

 

 

    
 

 

 

 

 

Exit

 

 

Figure (a)

 

ITO WE -

nonconductive etched __, -

region / -

ITO AE

wires to
potentiostat

 

 

Figure (b)

Figure 3-3. Diagram of flow cell for optical and electrochemical experiments (first
generation design).

The top slide in figure (b) is an ITO-coated glass slide. The ITO electrode was etched
into three conductive bands insulated from each other. One of the bands served as the
working electrode (WE), while either of the other two bands served as the
auxiliary/counter electrode (AE). Wires epoxyed to the conductive areas established
electrical contact. The bottom slide was a glass microscope slide. An oval cut into a
rubber spacer served as the flow channel. The reference electrode (RE) was placed in the
exit beaker.

 

 

76

 

 

— Flow rate

3mllmin
— No Flow

 

 

 

 

 

 

 

Current (pA)

 

 

 

 

 

 

 

W I I I I I I i
700 600 500 400 300 200 100 0
Potential (mV)

Figure 3-4. Cyclic Voltammogram obtained in the first generation TIRFM-CV flow
cell.

The working solution was lOmM potassium ferricyanide in 0.1M potassium nitrate. The
solution was introduced into the flow cell at a rate of 3 mI/min. The scan rate for CV was
100 mV/sec. The curve in pink was obtained under stopped flow conditions.

77

 

 

 

 

 

 

 

 

a

.. 10 A
3.5
O
E
5

4’", —— F— - --" ’

r -5

b

w” I I I I U I

700 600 500 400 300 200 100 0
Potential (mV)

Figure 3-5. Cyclic voltammogram obtained in a conventional three-electrode set-up
for comparison.

Potassium ferricyanide (lOmM) in 0.1M potassium nitrate was the working solution for
these experiments. A platinum working electrode, a platinum counter electrode and an
Ag/AgCl reference electrode were used. All three electrodes were placed in the same
solution in close proximity to each other.

78

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

”"600 i i i .,_ I , , , -
fxx .
A / i ‘ — Flow rate
3 L—zee arm/min
; —— No Flow
E :
5 :9: , I
o M; =—-——»-—- / .
700 600 500 400 300 200 1 00 0
Potential (mV)
a) CV obtained in TIRFM-CV flow cell
' 15 ' I ' I
r 10
3 - .
i
:40”
__t/
I —s
700 600 500 400 300 200 100 0

Potewal (mV)

b) CV obtained in the traditional set—up

Figure 3-6. Comparison of cyclic voltammograms obtained in our modified first
generation TIRFM-CV set-up, and in a conventional three-electrode set-up.
Detailed descriptions of these figures have been provided in Figure 3-4 and Figure 3-5.

79

 

Laser

Potentiostat

 

 

 

ITO top slide

   
 

Wire from WE

 

 

 

 

Spacer coacoooooooccoooooooooo

_’.r
“flusmrfléé ...........
/ VWe from RE 3
ITO Bottom slide syringe RE
pump

 

‘-.----.--

 

 

 

 

 

 

 

C:
C:

 

 

 

 

 

 

 

 

 

 

 

 

 

O-ring Top and bottom slide with aluminum
foil for provided electrical contact

Figure 3-7. Diagram of second generation optical - electrochemical flow cell.

The first generation flow cell was modified to change the positions of the working
electrode (WE) and the auxiliary electrode (AB). The top and the bottom ITO-coated
glass slides were used as the WE and the AE respectively. Aluminum foils attached to
the ITO provided electrical contact. The RE was placed closer to the flow cell by
inserting it in a T-joint in the inlet tube.

80

.nmEOmocngfiu anemones—«-5: 9 @882 8.858 85 8:32
2£m5>2 m m_ :2: corona ”Eugene Eoomouosu 9 5353.8 3.8250 35 c0382 03655.5 5 m_ an: :ouomom .w-m oSwE

3v 3
ZEMOmmmomeEe ZEMOmmmV 955N<mmmv

:0 o 0: :0 o o :0 o o
\ \
z z z
_ +

m o

81

 

 

 

 

 

 

 

E

\_—/ t

/ 3

U

500 400 300 200 100 0 -100 -200 -300 400 -500
Potential (mV)

Figure 3-9. Cyclic Voltammogram of 0.02mM resazurin in PBS buffer.

The TIRFM-CV flow cell was filled with 0.02mM resazurin solution in 0.1M PBS buffer.
The curve above shows the cyclic voltammogram obtained at a scan rate of 30mV/sec
over three cycles. The oxidation and reduction potentials of resazurin and resorufin are
listed in Table 3—3. During scans towards increasing negative voltages, resazurin was
reduced to resorufin at about the reduction potential for this irreversible couple (~160mV)
(shown as reaction “a” in Figure 3—8). Further along the same scans, the resorufin
produced was reduced to dihydroresorufin at about the reduction potential for this
reversible couple (-277mV) (reaction “b” in Figure 3-8). In the reverse scan towards

increasing positive voltages, the dihydroresorufin was oxidized back to resorufin at about
—238mV.

82

 

30000 -. . . _ W .. .3. mm, - .. - W. .. .,..w,m,,

25000 <

20000 ~

    

15000 4

10000 J

Fluorescence (photonsl200msec)

5000

 

 

o I I I I I
0 1 2 3 4 5 0

Time (min)

 

.1

VJ
co

Figure 3-10. Fluorescence data obtained in the TIRFM-CV flow cell containing
0.02mM resazurin.

The working solution was 0.02mM resazurin in 0.1M PBS buffer. An increase in
fluorescence emission was due to the production of resorufin, either by reduction of
resazurin or by oxidation of dihydroresorufin. A decrease in fluorescence was a result of
the reduction of resorufin to dihydroresorufin. The potential was cycled three times. The
optical properties of the dyes are listed in Table 3-4.

83

 

 

 

 

 

25000 500
-— 400
§ 20000 -— 300
E
8 ~- 200
% 5*
E 15000 -- 100 g
0 a
fa _. ° '2
0
§ 10000 ~— -100 ‘6
D L
0
g -— -200
.2 5000 —- -300
ll.
-~ .400
o T . . . I -500
o 1 2 3 4 5 6
Time (min)

Figure 3-11. Combined plot of fluorescence vs. time and applied potential vs. time
for 0.02mM resazurin.

The working solution was 0.02mM resazurin in 0.1M PBS buffer. After the flow cell was
filled, fluorescence and CV data collection was started simultaneously. When the
potential was scanned towards negative potentials, resazurin was reduced to resorufin at
about —160mV, with an increase in fluorescence. Resorufm was further reduced to
dihydroresorufin at about ~277mV with a decrease in fluorescence. In the reverse scan,
when the potential was scanned towards more positive potentials, the dihydroresorufin
was oxidized to resorufin at about —238mV, with an increase in fluorescence. The
potential was cycled three times. The electrochemical properties of the dyes are
summarized in Table 3-3, while the optical properties are listed in Table 3-4.

84

 

Current (pA)

 

 

 

:b

I I I I

SIG 4%: 300 200 100 I 10 -100 -200 -300 400 -5 DO

*—

 

 

 

 

 

Potontlal (mV)

Figure 3-12. Cyclic Voltammogram of 0.002mM resorufin in PBS buffer.

The TIRFM-CV flow cell was filled with 0.002mM resorufin solution in 0.1M PBS
buffer, then fluorescence and CV data collection was started simultaneously. The curve
above shows the cyclic voltammogram obtained at a scan rate of 30 mV/sec. The
reduction of resorufm to dihydroresorufin occurred at about -277mV on scanning towards
increasing negative voltages. The oxidation of dihydroresorufin back to resorufin at -
238mV occurred on scanning towards increasing positive voltages. The CV peaks are
not pronounced, as explained earlier. The potential was cycled five times. The oxidation
and reduction potentials of resorufin and dihydroresorufin are listed in Table 3-3.

85

6000

 

sooo - i '
4ooo -
3000 ~
2000 -

1000 —

Fluorescence (photonsIZOOmsec)

 

 

 

Time (min)

 

+ fluorescence

 

Figure 3-13. Fluorescence data obtained in the TIRFM-CV flow cell for 0.002mM
resorufin.

The working solution was 0.002mM resorufin in 0.1M PBS buffer. A decrease in
fluorescence was observed when resorufin was reduced to dihydroresorufin, while an
increase in fluorescence resulted from the reverse redox reaction. The potential was
cycled five times. The optical properties of the dyes are listed in Table 3-4.

86

 

P tentlal

~ -200
P -300
-400
0 r l v r r -500

Fluorescence (photons/200msec)

 

 

 

 

Tlme (mln)

 

b-fluorescence . potentlal]

 

Figure 3-14. Combined plot of fluorescence versus time and applied potential versus
time for 0.002mM resorufin.

The working solution was 0.002mM resorufin in 0.1M PBS buffer. After the flow cell
was filled, the fluorescence and CV data collection was started simultaneously. When
the potential was scanned towards negative potentials, resorufin was reduced to
dihydroresorufin at about —277mV, with a decrease in fluorescence emission. In the
reverse scan, the dihydroresorufin was oxidized to resorufin at about —238mV, with an
increase in fluorescence. The potential was cycled five times. The electrochemical
properties of the dyes are summarized in Table 3-3, while the optical properties are listed
in Table 3-4.

87

4 DEVELOPMENT AND OPT [CAL-ELECTROCHEMICAL
CHARACTERIZATION OF A BIOSENSOR BASED ON D-FRUCTOSE
DEHYDROGENASE (FDH)

4.1 ABSTRACT

We have developed and characterized a fructose dehydrogenase (FDH) biosensor,
based on indirect bioelectrocatalysis to ensure electrical communication with the
electrode, using coenzyme Q6 as the mediator. FDH, a naturally membrane-bound
enzyme, was incorporated into a bilayer lipid membrane (BLM) deposited on indium tin
oxide (ITO), along with Q6. The sensor was evaluated by using cyclic voltammetry (CV)
to monitor the change in anodic current upon addition of D-fi'uctose to the sample cell.
However, this FDH biosensor exhibited poor sensitivity, because electron tunneling

through the BLM was very low.

To improve electron tunneling, the carbon chain length characteristic of lipid
bilayers was shortened. Octyltrimethoxysilane was used to form an 8-carbon monolayer
on ITO, and liposomes containing FDH and Q6 were deposited to form the outer leaflet of
a biomimetic layer. This sensor demonstrated a clear dependence of the anodic current
on D-fi'uctose concentration, with a linear dependence over D-fi'uctose concentrations of
0.056mM to 1.98mM and a plateau at higher D-fructose concentrations, as expected for
an enzymatic sensor. Since the biomimetic interface was formed on an (IT O)-coated

glass slide, we were also able to characterize the bilayer optically using fluorescence

88

recovery afler pattern photobleaching (FRAPP). The diffirsion coefficient of the upper
leaflet of the biomimetic layer was 4.4 x 10'10 cmz/sec 3; 0.7 x 10'10 cmz/sec, with a

membrane fluidity of 0.59 i 0.099.
4.2 INTRODUCTION

In the previous chapter of this thesis, we demonstrated a protocol for
simultaneous optical and electrochemical characterization, using the chemical dye
resorufin as a test molecule. Both optical and electrochemical measurements provided
corroborative evidence of the reversible conversion of fluorescent resorufin to non-
fluorescent dihydroresorufin upon reduction. Besides providing confirmation of the same
event, the experimental set-up can also be used to monitor independent events occurring
within the same sample. In this chapter, we present details of work on developing a
biosensor based on D-fructose dehydrogenase (FDH), and its characterization by the

combined optical-electrochemical protocol.

Biosensors serve as reliable miniaturized analytical devices for the detection of
various analytes with high sensitivity. They have applications in numerous fields such as
environmental monitoring, food technology and medical diagnostics. In this study, FDH
was immobilized in two different membrane-mimicking environments: a bilayer lipid
membrane (BLM) and a biomimetic layer. Fluorescence recovery after pattern
photobleaching (FRAPP) was used to characterize the two systems, and cyclic

voltammetry (CV) was used to monitor the electrochemical redox reaction catalyzed by

89

FDH.

4.3 DEVELOPMENT OF AN FDH-BASED SENSOR

4.3.] Introduction to FDH

FDH (EC 1.1.99.11) is isolated from Gluoconobacter Industrius, and is a
membrane-bound quinoprotein that was first discovered by Yarnada and coworkers
(Yamada et a1. 1966). It is a 140kDa pyrroloquinone quinone (PQQ)-containing
oxidoreductase. PQQ is the cofactor associated with the enzyme FDH and remains
permanently bound to it. Amayema and coworkers first purified and characterized the

catalytic and molecular properties of FDH (Ameyama et al. 1981; Ameyama et al. 1981).

FDH catalyzes the conversion of D-fi'uctose to 5-keto-D-fructose. Upon
enzymatic oxidation of D-fructose, the prosthetic group in the enzyme (PQQ) is reduced

to PQQHZ , according to the following scheme:

D-Fructose + FDH-PQQ —> 5-keto-D-fi'uctose + FDH-PQQHZ (4.1)

Amayema and coworkers recognized that FDH exhibited substrate specificity, i.e.
it catalyzed the conversion of the monosaccharide D-fructose only, but no other forms of
sugar (Ameyama et a1. 1981; Ameyama et al. 1981). Thus, an analytical assay based on
FDH would be useful for the quantitative determination of D-fructose. Fructose, being
sweeter than glucose and sucrose, is frequently used as a diabetic sweetener. It is also

widely present in fi'uits, vegetables, honey and soft drinks. Fructose also serves as an

90

indicator of the ripening process in fruits (Paredes et al. 1997). It is also present in the
seminal fluid, where its concentration is an indicator of seminal vesicle size and secretory
capacity. Thus, the ability to precisely determine fructose concentrations is of great and

general importance.

Numerous FDH-based arnperometric biosensors have been developed to date.
These sensors exhibit a unique response, based on the different environments the FDH is
incorporated into. Examples of FDH-based sensors include FDH incorporated into
conducting polypyrrole matrices (Begum et al. 1993), immobilization in self-assembled
monolayers or membrane mimetic layers on gold (Kinnear and Monbouquette 1993;
Kinnear and Monbouquette 1993; Kinnear and Monbouquette 1997; Darder et al. 2000;
Campuzano et al. 2003; Campuzano et al. 2004), incorporation in carbon paste matrices
(Parellada et al. 1996; Paredes et al. 1997), immobilization in cellulose acetate
membranes (Tkac et al. 2001), and free FDH adsorbed on gold, platinum and glassy

carbon electrodes (Khan et al. 1991).

In this study, we present an FDH-based biosensor immobilized in a biomimetic
interface assembled on ITO. We exploited the transparency of the ITO film, along with
its conductive property, to conduct both optical and electrochemical characterization of

the sensor.

Quinoproteins are especially advantageous for use in biosensing elements,
because they do not undergo auto-oxidation. This is because the cofactor is tightly bound

to the enzyme, and oxygen does not compete for its reoxidation (Khan et al. 1991;

91

Paredes et al. 1997). Reports in the published literature suggest that FDH has a
hydrophobic and a membrane-lipid-interacting structure, and is thus an intrinsic
membrane protein (Ameyama et a1. 1981; Ameyama et al. 1981). A membrane bound
enzyme typically requires a hydrophobic environment for optimal activity and stability of
the enzyme. Studies also report that the enzymatic activity is higher when FDH is bound

to a membrane than when it is free in solution (Kato et al. 2003).

Besides mimicking its natural environment, the presence of a lipid bilayer
prevents easily oxidized electroactive interferents such as ascorbic acid (typically present
in fruits, etc.) from reaching the electrode surface. Erroneous signal generation due to
electroactive interferents is thus eliminated. The lipid bilayer also causes a reduction in
background noise, resulting in increased sensitivity. Studies have shown that
incorporation of enzymes into such interfaces creates systems that exhibit greater

reproducibility and robustness (Campuzano et al. 2003).

4.3.2 Theory

A common problem in the construction of enzyme-based amperometric
biosensors is the achievement of direct electron transfer between the enzyme and the
electrode. This can be a problem with enzyme-based sensors, because the prosthetic
group in most enzymes is located deep within the enzyme and, as a result, is not available
for efficient electron transfer. The prosthetic center, in most cases, also undergoes
irreversible electrochemical reaction when present by itself. High overpotentials are

required for the redox processes in these prosthetic groups (Begum et al. 1993). Ideally,

92

amperometric biosensors operating at less extreme potentials need to be developed, to
reduce the effect of interfering reactions caused by easily oxidizable components at high

potentials.

To address the problems listed above, “indirect bioelectrocatalysis” has become a
commonly used method to achieve efficient electron transfer. In indirect
bioelectrocatalysis, a mediator is added to act as the electron acceptor and to shuttle
electrons from the enzyme to the electrode. Thus, after the completion of the enzymatic
reaction given by Equation (4.1), the PQQHZ from the enzymatic reaction is reoxidized
electrochemically with the aid of the redox mediator, coenzyme Q6. The current
produced as a result is directly proportional to D-fructose concentration (Paredes et a1.
1997). Coenzyme Q6 is known to be a relatively stable mediator that can be regenerated
at an electrode (Kinnear and Monbouquette 1993; Kinnear and Monbouquette 1993).

The electrochemical reaction occurring in the presence of the coenzyme is given by

Q
FDH(PQQH2) 3 FDH(PQQ) + 2H+ + 2e' (4.2)

Electron transfer between the prosthetic group (PQQ) and the coenzyme is
catalyzed at less extreme potentials. The oxidation of the reduced form of the enzyme
(PQQHZ) takes place at about 400mV (versus Ag/AgCl). Coenzyme Q6, required for the
FDH biosensor, is typically incorporated into the liposomes used to form the lipid
bilayer. Incorporation results in immobilization of the coenzyme, preventing the

mediator from diffusing away from the electrode into the bulk solution (Begum et al.

93

1993). This is important, because diffusion of the mediator into the bulk solution causes

a loss in functionality of the biosensor (Kinnear and Monbouquette 1997).

4.3.3 Introduction to FRAPP

Fluorescence photobleaching recovery (FPR) and fluorescence recovery after
pattern photobleaching (FRAPP) are both effective techniques used to measure apparent
diffusion coefficients. However, diffusion of lipids and proteins in BLMs is
characterized by non-ideal behavior, and cannot be well—characterized as ideal self-
diffusion. In addition, the spatial profile of the excitation source in FPR can deviate from
the ideal Gaussian shape as the beam passes through a series of optical components
before becoming a focused spot. Since the form of the fluorescence recovery data
depends critically on the beam shape, FPR cannot be used to measure non-ideal diffusion

in lipid bilayers (Starr and Thompson 2002).

In the case of FRAPP, developed by Smith and McConnell in 1978 (Smith and
McConnell 1978), the spatial profile of the excitation intensity is known. Thus FRAPP is
commonly used to measure the lateral diffusion of fluorescent phospholipid probes in
either natural lipid bilayers or in biomimetic layers. The redistribution of fluorescence
emission alter photobleaching is monitored in a pattern of well-characterized periodic

stripes, which yields information about the motion of the fluorophores.

A mathematical model that takes diffusion non-idealities into account was

developed by Starr and Thompson (Starr and Thompson 2002). In this model, the system

94

is characterized as being composed of a number of discrete fluorophore populations, with
each population exhibiting ideal diffusion with specific diffusion coefficients. The time-
dependent fluorescence emission of a two-dimensional sample containing fluorophores

that undergo ideal diffusion is given by

F(t)=Q I [1(x.y)C(x,y,0dxdy (4.3)

-oo-oo

where F (t) is the time-dependent fluorescence, Q is a proportionality constant, 1(x, y) is
the monitoring beam intensity, and C(x, y, t) is the density of unbleached fluorophores at
position (x,y) at time (t). For the specific case of the FRAPP set-up where a Gaussian
laser beam is intersected by a ronchi ruling in the back image plane of a microscope, the

intensity is given by

2 2
1( x, y) = [gexp [flax—3:2] x I] + 2 C" cos(nkx)I (4.4)
2 S nodd

where 10 is the beam intensity at the origin, 3 is the 1/e2-radius of the expanded beam, k is
the spatial frequency of the striped pattern, and “n odd” signifies summatiOn over all odd

numbers.

The parameters k and C,. are given by

 

_ (n-1)/2
kg C" A 1)

a "71'

(4.5)

95

where a is the spatial period.

For ideal diffusion, C(x, y, t) is given by

 

(1) oo ' ' - '2 - '2 ' '
C(x,y,t)=-4—”-1—EIIC(x,y,0)xexp[-(x “4;? Y) :Idxdy (4.6)

-00 -00

where the initial density of unbleached molecules can be approximated as
C(x, y, 0) = Ce'x’W) (4.7)

where C is the total fluorophore density (bleached and unbleached) and K is a constant of
proportionality that depends on the duration of the bleach pulse, absorptivity of the

fluorophores, and the quantum efficiency for bleaching.

In the situation where the striped periodicity is small compared to the illuminated
area, the normalized fluorescence photobleaching recovery curve for a sample containing
one diffusive population of fluorophores is given by (Smith and McConnell 1978; Wright

et al. 1988)

¢(t) = ¢(O) + (gjfl - ¢(0)]{1—(;[§2—)Iexp[- 4:22” I — [Ii-j exP[- 36::Dt III , t 2 0

 

 

(4.8)

96

where ¢(t) is the ratio of the postbleach fluorescence [F (t>0)] to the constant prebleach

fluorescence [F(t<0)], and m is the fraction of fluorophores that are mobile.

Thus (1 -m) is the fraction of immobile fluorophores; and a is the periodicity of the
striped pattern which, in our experiments, was 12.5p.m. A similar model was developed
by Wright and coworkers for a sample containing two mobile species (Wright et al.
1988). However, all data obtained in this work were well-described by the “single-
mobile species” model; therefore, the “two-mobile species” model will not be described

here.
4.4 EXPERIMENTAL SECTION

4.4.1 Materials

D-Fructose dehydrogenase (FDH, Catalog # F-5152, 400-l,200 units/mg protein)
isolated from gluconobacter industrius, was purchased from the Sigma Chemical
Company (St. Louis, MO) and used without firrther purification. FDH is stable between
pH 3 and 5.5, with 4.5 being optimum. Thus, all solutions were used at a pH of 4.5.
Sodium phosphate monobasic (Catalog # 3818-01, J .T. Baker, Phillipsburg, NJ) and
sodium chloride (Catalog # S1240, Spectrum Chemical Corporation, Gardena, CA), used
in the preparation of the pH 4.5 buffer, were purchased from the Sigma Chemical
Company (St. Louis, MO). Egg phosphatidylcholine (Egg-PC, Catalog # 850375C) and
fluorescently labeled lipid (NBD-PC, Catalog # 810130), used in the preparation of
liposomes, were purchased from Avanti Polar Lipids (Alabaster, AL). Triton X-100

(Catalog # 21123), used for the solubilization of the protein, was purchased from Supelco

97

(Bellefonte, PA).

Indium tin oxide (ITO)-coated glass slides (Catalog # CG-41IN-S107) were
purchased from Delta Technologies Ltd. (Stillwater, MN). Co-enzyme Q6 from
Saccharomyces cerevisiae (Catalog # C9504), used as the mediator in the electrochemical
assays, was purchased from the Sigma Chemical Company (St. Louis, MO).
Octyltrimethoxysilane (Catalog # 376221), used for the preparation of hydrophobic
monolayers on ITO, was purchased from the Sigma Chemical Company (St. Louis, MO).
Dialysis was carried out with a regenerated cellulose porous membrane (Spectra/Per 1,
Catalog # 132655, molecular weight cutoff - 6000 to 8000) purchased fiom Spectrum
Laboratories (Rancho Dominguez, CA). Bio-beads (Catalog # 152-8920), used for the
removal of triton, were purchased from Bio-rad Laboratories (Hercules, CA). D-Fructose
(catalog # F2543) was purchased from the Sigma Chemical Company (St.Louis, MO).
Potassium ferricyanide (catalog # 6912), used for the colorimetric assay, was purchased
from Mallinckrodt, Inc. (Paris, KY). Ferric sulfate (Catalog # I-142), used for the
colorimetric assay, was purchased from Fisher Scientific (Springfield, NJ). RBS-SO
detergent concentrate (catalog # 27952), used for cleaning the glass slides and the ITO-

coated glass slides, were purchased from the Pierce Chemical Co. (Rockford, IL).
4.4.2 Preparation of the proteoliposome solution

FDH and the coenzyme were incorporated into liposomes using standard
protocols (Meuller et al. 1997; Graneli et al. 2003; Girard et al. 2004). One ml of 20

mg/ml Egg-PC in chloroform, 1 m1 of 2 mg/ml NBD-PC in chloroform and 16.67 ml of

98

co-enzyme Q, in ethanol were added to a glass test tube, and the solvent was evaporated
under a nitrogen blanket. The lipid-coenzyme mixture was freeze-dried overnight under
vacuum at -47°C, in a rotovac freeze drier. Sodium phosphate buffer (0.01M) containing
0.1M sodium chloride was prepared, and the pH was adjusted to 4.5 using a strong acid

or base as required.

One ml of pH 4.5 buffer was added to the lipid-coenzyme mixture and the
solution was vortexed and sonicated to form small unilarnellar vesicles. Solubilization of
the protein was achieved by adding 50 ug/ml of Triton. FDH (0.5 mg) was added to the
solution, followed by mixing for 45 minutes. The detergent was removed from the
solution using two different methods. One method involved three different subsequent
additions of 0.5 mg/ml of biobeads to the solution, followed by removal of the biobeads
by decantation after the mixing was complete. The other method dialyzed the solution
(in a regenerated cellulose membrane) against phosphate buffer for about 3-4 days at 4°C,

using about 6 buffer changes.
4.4.3 Adsorption of proteoliposome solution onto ITO-coated glass slides

An ITO-coated glass slide was cut into smaller pieces and sonicated for 30
minutes in RES-50 detergent solution, followed by rinsing under distilled water and
drying using nitrogen. The slide was oxygen plasma cleaned for a period of about 20
seconds immediately before each experiment. After plasma cleaning, the slide was

placed in a cell holder to ensure that the working area would be reproducible. The

electrode area exposed to the solution was about 0.17 cmz. About 100 pl of the enzyme

99

solution was added to the exposed circular ITO, and the solution was allowed to adsorb
for a period of about 45 minutes to one hour to ensure formation of a lipid bilayer.
Following this, the cell holder was carefully immersed in a beaker containing pH 4.5

buffer solution, ensuring that the working area always stayed in contact with the solution.
4.4.4 Preparation of hydrophobic silane monolayers on ITO-coated glass slides

The silylated surfaces were prepared by deposition from aqueous alcohol,
following the protocol outlined in a technical documentation obtained from Gelest, Inc.1
A 95% ethanol and 5% water solution was prepared, and the pH was adjusted to between
4.5 and 5.5 with acetic acid. Octyltrimethoxysilane was added with stirring to yield a 2%

v/v solution. Five minutes was allowed for hydrolysis and silanol formation (Figure 4-1).

The ITO-coated glass slide was plasma cleaned under oxygen for 20 seconds,
followed by piranha cleaning for 30 seconds. The slide was then washed with deionized
water and dried under nitrogen. It was then immersed in the silane solution and allowed
to react for 30 minutes. The slide was then rinsed in ethanol and baked in an oven at

160°C for 12 hours to cure the silane layer.
4.4.5 Adsorption of proteoliposomes onto silanized ITO slides

The hydrophobicity of the silanized ITO slide was checked visually by placing a
drop of water on the surface. The contact angles were observed to be in excess of 90°.

The silanized slide was placed in the cell holder to ensure a reproducible circular working

2

area of 0.17 cm . About 100 pl of the enzyme solution was added to the exposed area,

 

' “Applying a silane coupling agent”, technical documentation, Gelest Inc., Morrisville, PA

100

and the solution was allowed to adsorb for a period of about 45 minutes to one hour to

ensure formation of a monolayer of liposomes on the C3 hydrophobic monolayer.
4.4.6 Cyclic Voltammetry Experiments

The modified electrode (either the lipid bilayer on ITO or a monolayer of
liposomes formed on the C3 silanized monolayer) was placed in the cell holder, and was
immersed in deoxygenated pH 4.5 buffer. The working electrode from the potentiostat
was connected to the ITO, and the platinum counter electrode and the Ag/AgCl reference
electrode were placed in the buffer solution. D-fructose in varying amounts was added to
the buffer solution, and cyclic voltammetry measurements were canied out at l mV/sec
to obtain steady state data. Films were scanned at very low rates (1 mV/sec), to allow for
near redox equilibrium to be established between the electrode and the redox protein.
Slow scan rates also allow us to obtain midpoint potentials (Haas et al. 2001). The

parameters used for the cyclic voltammetry scans are given in Table 4-1.

4.4.7 Set-up for FRAPP Experiments

4.4.7.1 Optical set-up for FRAPP experiments

A diagram of the optical set-up used for the FRAPP experiments is shown in
Figure 4-2. The laser beam from the argon ion laser was expanded using a 5x beam
expander (NT 55-577, Edmund Industrial Optics, Barrington, NJ). The expanded beam
was then passed through a pair of optical flats. The first flat was used to split the beam

into two lines of different intensities — a high intensity beam used for photobleaching

101

(200-500 mW) and a low intensity beam used for monitoring the interfacial dynamics (3-
4 uW). The second optical flat was placed after the shutter, and exactly recombined the
two beams. The combined beam was then directed through a set of mirrors into the epi-
port at the back of an inverted microscope (Axiovert, 135M, Carl Zeiss Inc., Thomwood,
NY). The beam passed through a 50 line per inch ronchi ruling (NT 56-592, Edmund
Optics, Barrington, NJ) placed in the back image plane of the microscope, and through a
filter cube which directed the light into a microscope objective (32x, 440851, LD
Acrophlan, Carl Zeiss Inc., Thomwood, NY) placed right below the stage housing the

flow cell.

When a beam passes through a ronchi ruling placed in the back image plane of the
microscope, a real image of the ronchi ruling is projected onto the sample plane. The
image appears as alternating bright and dark fringes (Figure 4-3) when viewed through
the eyepiece of the objective or by a CCD camera (VB-1000, DAGE-MTI, Michigan
City, IN). Prior to the start of an experiment, bands were marked out on an ITO-coated
slide using a fluorescent marker. The position of the laser beam (adjusted using the
mirrors) and the position of the microscope objective were adjusted such that the fringes
(created using the fluorescent bands) came into sharp focus and filled the viewing area.

This approximately located the point of focus of the objective.

Beam alignment was further ensured by using the CCD camera to confirm that the
bright fiinges formed by the photobleaching beam overlapped the faint fringes formed by

the monitoring beam. After the experiment was started, the position of the objective was

102

readjusted minutely to relocate the fringes, because the position could have changed as a
result of minor variations in slides or a change in position induced by the presence of a

monolayer/bilayer adsorbed on the slide.

After allowing 45 minutes for the formation of the lipid bilayer (steps outlined in
section 4.1.8), the illuminated area was subjected to a brief 400ms flash of the high
intensity photobleaching beam. The photomultiplier tube (R4632, Hamamatsu,
Bridgewater, NJ) was blocked with a shutter during the bleaching period to protect it.
Fluorescence recovery data at the end of the bleaching process were collected from a
much smaller area than had been bleached (Figure 4-3). This was done to ensure that the
fluorescence recovery is only from unbleached fluorophores in the fiinges, and not from
unbleached fluorophores in areas outside the fringes. The photon counting electronics and

the data acquisition software are the same as described earlier in Chapter 2.
4.4.7.2 Experimental cell used for FRAPP experiments to study FDH system

A new cell had to be designed to characterize the properties of the biomimetic
layer, because of the small working volume of FDH-liposomc solution (only 1 ml of
working solution). The small volumes did not allow for flow experiments to be

conducted in the modified second generation flow cell described earlier (Chapter 3).

The new cell consisted of a box 2 inch high, 1 inch wide, and 1 inch deep,
constructed with a silanized ITO slide as the bottom surface. The walls of the box/cell

were made of microscope glass slides. The box was assembled using epoxy. In

103

electrochemical experiments, the bottom ITO slide served as the working electrode, and

the counter and the reference electrodes were placed directly in the cell.

4.4.7.3 Preparation of bionrimetic layer for FRAPP experiments

After preparation of the octylsilane monolayer on the ITO bottom slide, the cell
was assembled. Then, 60-100 pl of the FDH-liposome solution was placed onto the
bottom silanized ITO surface of the cell, and the solution was allowed to adsorb for about
45 minutes. Loosely adsorbed FDH-liposomes were removed by rinsing with about 10
ml of buffer solution. The cell was then placed on a stage on top of the objective of our

inverted microscope (Axiovert 135M, Call Zeiss Inc., Thomwood, NY).

4.5 RESULTS AND DISCUSSION

The goal of this work was to construct a bilayer lipid membrane (BLM)-based
FDH sensor, using indirect bioelectrocatalysis. Indirect bioelectrocatalysis requires close
contact between the enzyme, the mediator and the electrode for efficient electrical
communication (Ciucu and Ciucu 2002). To achieve this close contact, the correct ratio
of the enzyme, lipid and mediator must be used. Numerous trials were conducted to
determine the ratio of lipid:enzyme:coenzyme that would make a functional
amperometric biosensor. BLMs consisting of varying coenzymezlipidzenzyme ratios
were formed on ITO-coated slides, and the catalytic response of the sensor upon addition
of D-fructose was measured electrochemically (amperometrically). In cases where the

electrochemical assay exhibited no catalytic activity in the presence of D-fructose, a

104

colorimetric assay was conducted to determine if the FDH was active.

Activity of FDH was determined using a colorimetric assay in the presence of

potassium ferricyanide. The reaction catalyzed in the assay is given by

FDH
D-Fructose + 2K3Fe(CN)6 —> 5-keto-D-Fructose + 2K4Fe(CN)6 (49)

2K4Fe(CN)6 + Fez ($04 )3 —> PrussianBlue + 3KZSO4

The production of Prussian Blue can be monitored and measured at 660nm by
spectrophotometry (Toyobo; Sigma 1994), allowing the activity of FDH to be calculated
in terms of units, where one “unit” is defined as the amount that causes the oxidation of
one micromole of D-fi'uctose per minute (Toyobo; Ameyama et al. 1981; Ameyama et al.

1981; Sigma 1994).

For all the different ratios of lipid:enzyme:coenzyme used, the colorimetric assay
outlined in Equation (4.9) always confirmed the production of Prussian blue, proving that
the enzyme was incorporated in an active conformation in the BLM and was able to
oxidize D-fructose to 5-keto-D-fructose. While the colorimetric assay always indicated
that the enzyme was active, the electrochemical assay did not exhibit any appreciable

catalytic activity upon addition of D-fi'uctose.

One of the major problems in these experiments was the need to use low
concentrations of the enzyme because of its high cost. Sensors containing lower

quantities of enzyme incorporated into liposomes did not exhibit any catalytic activity

105

upon addition of D-fructose in the CV experiments, even though a positive colorimetric
assay indicated that the enzyme was active. Hence, we concluded that electron transfer to
the electrode was being inhibited. We speculated that the probability of the enzyme not
being in close contact with the coenzyme is enhanced when the enzyme concentration is

low, resulting in lack of electrical contact.

Similarly, using the correct amount of mediator is also necessary. Studies have
shown that a scarcity or excess of the co-enzyme resulted in the absence of a fully
functional amperometric sensor. Having the mediator in low concentrations affected the
biosensor response, since it was limited by enzyme-mediator kinetics (Li et al. 1996;
Wang et a1. 1998; Miao et al. 2001; Yu et al. 2003; Wang et al. 2005). Similarly, having
the mediator at higher concentrations than required affected the biosensor response, since
it was limited by enzyme-substrate kinetics. Using a very high mediator concentration
could result in a high background signal, and thus the signal from the coenzyme could
overwhelm the signal due to the presence of D-fructose. This behavior is typical of

mediator-based sensors.

The concentration of liposome was also changed to determine the level high
enough to form a BLM while also allowing for efficient electrical contact between the
FDH, coenzyme and the electrode. Too high a liposome concentration may increase the
probability of the enzyme being far away from the cofactor. We decided to use 1 mg/ml
of liposomes, based on experiments conducted in our laboratory using FRAPP that

showed formation of lipid bilayers at this concentration on various substrates (V aidya

106

2005)

Proteoliposomes containing the coenzyme Q, in the ratio outlined in Section 4.4.2
were prepared. A lipid bilayer was formed on a freshly cleaned ITO slide, using the
procedure outlined in Section 4.4.3. The solution was purged with nitrogen, and a
nitrogen blanket was maintained during the experiment. The amperometric biosensor
was tested by monitoring the catalytic activity exhibited upon addition of D-fructose.
Addition of a substrate to the redox system promotes multiple turnovers of the enzyme,

and a steady state transfer of electrons through the enzyme to and from the electrode.

A background CV measurement was obtained using the parameters listed in Table
4-1. Redox waves of the coenzyme in buffer, similar to that observed by Kinnear and
Monbouquette (Kinnear and Monbouquette 1997), were visible in the background scan,
with a decrease in current in each successive cycle. At the end of the background scan,
varying amounts of D-fi'uctose were added to the solution, followed by mixing of the
solution by gentle pipetting. The response obtained for the BLM-based FDH sensor is
shown in Figure 4-4. An increase in catalytic current (compared to the background) was
observed upon addition of D-fi'uctose. For easier viewing, the region between 400mV

and 800mV has been expanded in Figure 4-5.

On scanning towards positive voltages beyond the redox potential of the
coenzyme, the anodic current increased with increasing D-fructose concentration in our
BLM-based FDH sensor, due to oxidation of D-fiuctose. However, the increase in anodic

current appears to be a relatively weak function of D-fructose concentration. To assess

107

the strength of this dependence, we added D-fi'uctose-free buffer to the solution, and
observed a corresponding decrease in current. Upon addition of lOmM D-fructose, a
19% net increase in current was observed (at 800mV) compared to the background.
These results appear to indicate that the BLM-based FDH biosensor is firnctional, albeit

one that functions at very low sensitivity.

While the results indicate that we have a functional biosensor, it does not display
the ideal behavior typical of enzyme-based sensors. This could be due to the fact that we
did not achieve steady state during our measurements. However, we believe it is more
likely that these results are an indication that electron transfer is not efficient and may,

perhaps, even be inhibited as explained below.

One of the causes of inefficient electron transfer in BLMs is that electron
tunneling through a lipid bilayer of high resistance may be very slow or negligibly small
(Bard and Faulkner 2000). There is an exponential decrease in the rate of electron
tunneling with distance, which may explain why electron tunneling through a BLM is
slow. There is also a possibility that some of the electron transfer that we observe occurs
because the BLMs are not highly insulating. A BLM is typically on the order of about 3-
5 nm in thickness. To decrease the electron tunneling distance, we could use one of two
approaches. We could form a self-assembled monolayer of a shorter carbon chain (6-10
carbons) on the ITO surface, following which the upper leaflet could be formed using
lipids. Or we could use custom-synthesized lipids with shorter carbon chains, which

would in effect decrease the electron tunneling distance upon formation of the bilayer.

108

We decided to form a self—assembled monolayer on ITO.

4.5.1 Incorporation of FDH in a biomimetic layer

Octyltrimethoxysilane was used as the linker molecule to form a self-assembled
monolayer on ITO. The electron tunneling distance was decreased by using an 8 carbon
chain in place of the l8-carbon lipid chain which forms the lower leaflet of the lipid
bilayer. The silanization protocol outlined in Section 4.4.4 was used to form the
hydrophobic C3 monolayer on the ITO-coated glass slide. The biomimetic layer
containing the FDH and the coenzyme was formed on the C3 monolayer, using the

procedure outlined in Section 4.4.5.
4.5.1.1 Optical characterization

FRAPP experiments were conducted to confirm bilayer formation and to
characterize the properties of the biomimetic bilayer. The upper leaflet was formed with
liposomes that contained about 2% fluorescent lipids (NBD-PC), so the fluidity of the
upper leaflet could be characterized using FRAPP. FRAPP data were collected at
multiple spots using the procedure outlined in Section 4.4.7.1. Figure 4-10 shows a
typical FRAPP curve with the solid line representing the fit to the data, using the “single
mobile species” model given in Equation (4.8). The results of the FRAPP experiments
have been summarized in Table 4-2. The average diffusion coefficient was 4.4 x 104°
cmz/sec i 0.7 x 10'10 cmz/sec, while the mobility of the membrane was determined to be
0.59 i 0.099. Similar diffusion coefficients have been observed for lipid bilayers formed

on poly(diallyldimethylammonium chloride)/1l-mercaptoundecanoic acid (PDAD/MUA)

109

(Zhang et al. 2000; Vaidya 2005)and on poly(dimethyldiallylammonium chloride)

(PDAC) layers (Zhang et al. 2000; Vaidya 2005).
4.5.1.2 Electrochemical characterization

After formation of the biomimetic layer, the solution was purged with nitrogen
and a nitrogen blanket was maintained during the experiment. A background CV scan
was collected with the parameters listed in Section 4.4.6. After the background scan was
completed, D-fructose in varying amounts was added to the solution. The CV

measurements were recorded after each addition, following gentle mixing of the solution.

Upon addition of D-fructose, a stable catalytic current (or current density) was
observed when scanning towards positive voltages past 400mV (Figure 4-6), due to
oxidation of D-fructose and turnover of the coenzyme. The curve has been expanded for
better visibility in Figure 4-7, to illustrate the dependence of the catalytic current on D-
fructose concentration. Clearly, the biosensor is sensitive to D-fructose concentration.
The dependence of current density on D-fi'uctose concentration at 800mV is shown in
Figure 4-8, and may be used as a calibration curve. The response of the sensor was
checked at 800mV, since the current attained steady state beyond this point. We obtained
a linear dependence of the catalytic current on D-fructose concentration at D-fi'uctose
concentrations between 0.056mM and 1.98mM, with a correlation coefficient of 0.9813
R2 (Figure 4-9). We observed a 351% increase in current at 800mV compared to the
background scan upon addition of IOmM of D-fructose. This increase is highly

significant, in comparison to the 19% increase in current observed in the case of the

110

BLM-based FDH sensor described earlier. As expected, the current approaches a plateau
at higher D-fructose concentrations, indicating saturation of the enzyme. This is a typical
behavior of enzyme-based biosensors. The sensitivity of this biosensor
(0.189lpA/cm2mM) is two orders of magnitude lower than others reported in the
literature. For example Kinnear and coworkers (Kinnear and Monbouquette 1993)

reported a sensitivity of 15pA/cm2mM, using 300p] of 1 mg/ml FDH. We speculate that

the difference in sensitivity is a result of using much lower amounts of the enzyme (100pl

of 0.5 mg/ml F DH) in the current experiments.
4.6 CONCLUSIONS

We have developed an FDH-based biosensor that catalyzes the conversion of D-
fructose to S-keto-D-fructose. Using ITO as the substrate gave us the ability to
characterize the biosensor both optically and electrochemically. FRAPP was used to
characterize the properties of the biomimetic layer, while CV was used to monitor the
catalytic response to increasing D-fructose concentration. The FDH and coenzyme Q6
were immobilized in a biomimetic layer on a silanized ITO slide. We observed a 351%
increase in current at 800mV upon addition of lOmM of D-fructose, compared to the
background. The catalytic current was proportional to the concentration of D-fructose
with the current approaching a plateau at higher fructose concentrations. The calibration
curve was linear over D-fructose concentrations between 0.056mM to 1.98mM. The
diffusion coefficient of the upper leaflet of the biomimetic layer was 0.044 x 10'8 cmz/sec

: 0.007 x 10'8 cmZ/sec. The mobility of the membrane was 0.59 i 0.099.

111

4.7 FUTURE WORK

We believe that using higher concentrations of FDH would help us develop a
FDH-based biosensor with higher sensitivity. We speculate that controlling the
orientation of the prosthetic group, such that the reaction center is close to the electrode
would also increase the efficiency of electron transfer. This could be achieved by
controlling the orientation of the enzyme using short chain sulfides or promoter
molecules that bind and/or align to orient the enzyme, to facilitate electron transfer using

electrostatic interactions (Darder et al. 2000).

Further tests such as response to ascorbic acid, stability of the sensor, and
response time upon addition of D-fructose could be conducted to further characterize the
biosensor. The biosensor could in addition be tested in a real sensing application, for

example to detect D-fructose in fruit juice.

The versatility of this optical and electrochemical setup allows us to expand its
application to study many other enzyme-based biosensors and other biological systems.
This experimental ability to conduct simultaneous optical and electrochemical
measurements is being exploited, to explore the ability to tailor the properties of lipid
bilayers. Specifically, ongoing work to study the effect of electric fields on BLMs is
being investigated. This would be of enormous importance in any application where a
BLM is used, because it would allow us to improve membrane quality (by tailoring its

mobility).

112

Table 4-1. Input parameters used to conduct cyclic voltammetry measurements for the
D-Fructose dehydrogenase system.

 

 

 

 

 

 

 

 

High (mV) 800

Low (mV) -400

Initial (mV) -400
Scan rate (mV/sec) 1
# of cycles 1

 

113

Table 4-2. Summary of regression values obtained by fitting experimental FRAPP data
to the single species model [Equation (4.8) (Wright et al. 1988; Starr and Thompson

 

 

 

 

 

2002).
Run # Diffusion coefficient Mobile fraction f0
(D *1010) cmzlsec m
l 4.35 0.487 0.241
2 3.72 0.684 0.350
3 5.09 0.594 0.266
average 4.39 0.588

 

 

 

 

 

114

 

iii

\O/Zi \o/EKO/Zi \0/

l l I I l I trimethoxyoctylsilane i i l

I
| l =

 

 

 

Figure 4-1. Surface chemistry used to form an 8-carbon monolayer on ITO-coated glass
surface using trimethoxyoctylsilane as the linker molecule

115

5x beam expander

    
  
 
  

Ardonlaser

Optical flats

sampw

objective Filter
block detector

Figure 4-2. Experimental set-up for fluorescence recovery after pattern photobleaching
(FRAPP) experiments in the epi-illumination mode.

116

   

(a) Photobleached area (b) Interrogated area

Figure 4-3. Fringe pattern in illuminated region obtained by using 100 lines per inch
ronchi ruling. b) The observation area is restricted by placing an aperture in the
camera/PMT image plane, to ensure that fluorescence recovery is restricted to
fluorophores in the fringes.

117

     

 

 

Current Denslty ( "Alcnfl
c

 

 

 

 

 

-0.5 r _ OmM
— 0.00025MM
'1 1 — 0.20501"
— OAQCMM
.1 .5 _ — 0.967mM
-— 1.85701"
-- 10M"
.2 a 1 , , r , j
400 -200 0 200 400 500 800

Potentlal (mV)

Figure 4-4. CV curves obtained for a BLM-based FDH sensor.
A lipid bilayer containing FDH and the coenzyme was formed on IT O-coated glass and
the slide was then placed in a pH 4.5 sodium phosphate buffer solution. The working

area was 0.17 cmz. Response of the biosensor was monitored after addition of D-fi'uctose
in varying amounts, using a scan rate of 1 mV/sec.

ll8

1 .9 - — OmM
— 0.00625mllll
— 0.205mM
— 0.496rnM
— 0.967mllll
— 1.857mM

1 Omllll

 

 
  
  
  
  
  
   

r‘
a
f

 

 

 

 

 

Current Density orA/cmz)
O O
3: co

 

0. 1 I I l I I I l i
400 450 500 550 600 650 700 750 800
Potential (mV)

 

Figure 4-5. Expanded view of relevant portion of data.
Figure 4-4 has been expanded between 400mV and 800mV for better visibility. There
appears to be a dependence of the current density on D-fructose concentration.

119

 

 
  
 
 
 
 
   
  

 

 

 

 

115 _ —OmM
0.0187mM
A1195 " —-0.0561mM
"g 075 ——0.205mM
3 ° I —0.496mM
a 055 _ —0.967mM
'7. —1.857mM
£0.35 _ —3.818mM
E —1o.455mM
g 0.15
3

 

 

400 -200 0 200 400 600 800
Potential (mV)

Figure 4-6. Catalytic response of an immobilized FDH biosensor.

A biomimetic layer containing FDH and the coenzyme was formed on the ITO-coated
glass and the slide was placed in a pH 4.5 sodium phosphate buffer solution. The
working area was 0.17 cmz. Response of the biosensor was monitored after addition of
D-fructose in varying amounts, using a scan rate of l mV/sec.

120

 

 
 
 
 
 
 
 
 
   

 

 

 

 

 

—0mM
1'15 “ 0.0187mM
——-0.0561mM
A 0.95 - —0.205mM
NE —0.496mM
3:075 _ —0.967mM
' —1.857mM
E —3.818mM
E 0.55 — —-10.455mM
3
E
g 0.35
3
0.15
-o.os . . r . 1
300 400 500 600 700 800

Potential (mV)

Figure 4-7. Expanded view of the relevant portion of data.
Figure 4-6 has been expanded between 300mV and 800mV to more clearly show the
dependence of catalytic current on D-fructose concentration.

121

 

0.9 4

Current (pAlcm 2)
9 .° 9 .° P
15 0| 0) N a
l l l l 1

P
U
4

.°
N
1

O

0.1 a

 

 

 

o I I I
0 2 4 6 8 10 12

Concentration (mM)

—1
d

Figure 4-8. Calibration curve of current density versus D-fructose concentration.
From the catalytic response of the immobilized FDH biosensor, a plot of anodic current
versus D-fi'uctose concentration was generated at 800mV, which was chosen to plot the
calibration curve because the current reached steady state at this potential.

122

P
‘1

 

 

 

 

y = 0.1891x + 0.2074
2 -

0.5 . R - 0.9813
“5 0.5 ~
i
.._. 0.4 -
E
W
C
3 0.3 -
a I
5
g 0.2 F I
U

0.1 ~

0 I I I
o 0.5 1 1.5 2 2.5

Conc. (mM)

Figure 4-9. Linear part of calibration curve at 800mV.

The biosensor showed a linear relationship between the catalytic current and D-fructose
concentration from 0.056mM to 1.98mM.

123

 

 

 

 

 

 

0.45-
(a)
0.40~
J
8
§ 0.35-
U)
9
8
L: 0.30-
0.25-
0.20 , . fl I ' ' ' '
0 50 100 15° 200
Time (seconds)
0.08-
0.06- .
0.04'I ' I ' . a I .
uI ~:.;‘ ffzg. .~.I :..‘" .'.'. a?
- ' l s ' ." ' ' ' 4' I
0.02 .,._'. -, "."'-J-ri fii- ill" (b)
. .3}, ""1“ 1:11.51.
0.. ._ .. are ,1.
' I T ' "L 3711'” l 1'a I“
‘ u , L .3 I . " I ' l
. U. u “A i '. l‘ ' .'
.002- .“0. -.E:: fi? ::" I ‘ .' .‘ -..
u C -I I .. ' O
. " :' ,- " r'. ' .' I. - '
-0.04- . . -
I I T ' T I I

Figure 4-10. (3) Typical FRAPP curve obtained with FDH and the coenzyme
immobilized in the biomimetic layer. The solid line represents fit to the data, using a
model that describes a population containing a single mobile and immobile fraction
(Smith and McConnell 1978; Wright et a1. 1988; Starr and Thompson 2002). (b) Plot of
residual versus time showing excellent goodness of fit.

124

5 APPENDICES

5.1 APPENDIX A:

5.1.1 MATLAB program used for filtering analog measurements to remove dark
current measurements obtained during the phases when the chopper blocked
the laser beam.

function p = test(data)

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

min_val = data(l ,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, 1 );
end

end

threshold = (0.72) * (max_val-min_val) + min_val;
counter=1;

finaldata(l, 1) = 0;

finaldata(1, 2) = 0;

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

p=finaldata;
%mean and std. dev.%
m=mean(p, 1 )

average=m( l)
s=std(p, l )

125

stdev=s( 1 )

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

for i=1:length(p);
diff=p(i,1)-average;
absolute=abs(diff);
if (absolute<stdev)
endgame(counter, 1) = p(i,1);
endgame(counter, 2) = p(i,2);
counter = countertl;
end

end

save 'r:\co1loid-surface-science\output.txt' endgame -ASCII;

p=endgame;

5.1.2 MATLAB program used for averaging data

c=l;

b=1;

g=size(A, I )/ 8;
P=fl00r(g)+1

while b<p

sum=0;

d=c+7;

for j=c:d

sum=sum+A(j);

end

av g( 1 ,b)=sum/ 8;

c=c+8;

b=b+1;

end;

n=avg';

delete 'r:\colloid-surface-science\my_data.txt'
save my_data.txt n —ASCII

126

5.1.3 Labview Program for collecting fluorescence data in the analog mode.

 

 

    

 

V | >scan rate>
an rate member of chans< -_
scans/ set) E ‘ °

 

 

 

 

 

of ts

 

 

umber of has to were It; eadt data nolnt

. b -

l
i
g >errcir cluster>
1

ts bi ...... {I
+ @581: I-m -

Q

 

 

 

 

 

    
 

{my}

Figure 5-1. Labview flowsheet for collecting fluorescence data in the analog mode (left
hand side) (Figure 5.1 continued on next page).

127

Labview Program for collecting fluorescence data in the analog mode.

   
 
   

continuous

  
 

> scan rate >

 

<error cluster!
< taskID >

Figure 5-2contd. Labview flowsheet for collecting fluorescence data in the analog mode

(Right hand side of Figure 5.1)

128

5.1.4 Labview Program for collecting fluorescence data in the digital (photon
counting) mode.

 

 

 

 

 

  

J 0 [0 2] L- - - -
CONFIGUZE VI. ASSIGN INPUT 1 T0 COJNTER 'A',

10MHZTOCOUNT E "

 

 

Figure 5-3. Labview flowsheet for collecting fluorescence data in the digital (photon
counting) mode (Figure 5-3 continued on next two pages).

129

Labview Program for collecting fluorescence data in the digital (photon counting)
mode.

 

 

 

   
 

    

'A' GATE VI. SET MODE TO "CW", TSEI TO
EQUIVALENT TIME IN SECONDS.

 

Figure 5-4contd. Labview flowsheet for collecting fluorescence data in the digital
(photon counting) mode (Figure 5-3 continued on next page).

130

Labview Program for collecting fluorescence data in the digital (photon counting)
mode.

 

 

 

 

 

 

Figure 5-5contd. Labview flowsheet for collecting fluorescence data in the digital
(photon counting) mode.

131

REFERENCES

Abel, A. P., et al. (1996). "Fiber-optic evanescent wave biosensor for the detection of
oligonucleotides." A_r_r_alvtical Chemistry 68(1 7): 2905-2912.

Ameyama, M., et al. (1981). "S-Keto-D-Fructose Reductase of Gluconobacter-Industrius
- Purification, Crystallization and Properties." Agricultural and Biological
Chemistg 45(4): 863-869.

Ameyama, M., et al. (1981). "D-Fructose Dehydrogenase of Gluconobacter-Industrius -
Purification, Characterization, and Application to Enzymatic Micro-
Deterrnination of D-Fructose." Jouran meteriology 145(2): 814-823.

Andrade, J. D. (1985). Surface and interfacial aspects of biomedical polymers. New
York, Plenum Press.

Andrade, J. D., et al. (1996). "Poly(ethylene oxide) and Protein Resistance." ADV
CHEM SER 248: 51-59.

Arai, K., et al. (1993). "Drug Ion Transfer Across On Oil-Wwater Interface And
Pharmacological Activity." Bioelectrochemistrv And Bioenergetics 31: 65-76.

Armistead, P. M. and H. H. Thorp (2000). "Modification of Indium Tin Oxide Electrodes
with Nucleic Acids: Detection of Attomole Quantities of Immobilized DNA by
Electrocatalysis." Anal. Chem. 72: 3764-3770.

Asanov, A. N., et al. (1998). "Regenerable Biosensor Platform: A Total Internal
Reflection Fluorescence Cell with Electrochemical Control." Anal. Chem. 70:
1156-1163.

Axelrod, D. (1989). "Total Intemal-Reflection Fluorescence Microscopy." Methods in
Cell Biology 30: 245-270.

Axelrod, D., et al. (1983). "Total internal reflection fluorescence micrscopy." Journal of

132

Microscopy 129(1): 19-28.

Axelrod, D. H., E.H.; Fulbright, RM. (1992). Total internal reflection fluorescence.
Tmics in fluorescence spectroscopy Volume 3 : Biochemical Applications
Lakowicz, Joseph.R.: 289-335.

Bani-Jaber, A., et al. (2000). "Efficacy of the antimicrobial peptide nisin in emulsifying
oil in water." Journal of Food Science 65(3): 502-506.

Barak, L. S., et al. (1997). "A beta-arrestin green fluorescent protein biosensor for
detecting G protein-coupled receptor activation." J ourpal of Biological Chemistry
272(44): 27497-27500.

Bard, A. J. and L. R. Faulkner (2000). Electrochemical Methods - Fun_d_amentals and
Applications. John Wiley & Sons.

Baselt, D. R., et al. (1998). "A biosensor based on magnetoresistance technology."
Biosen_sors & Bioelectronics 13(7-8): 731-739.

Batchelor, R. H. and M. Zhou (2002). "A Resorufin-Based Fluorescent Assay for
Quantifying NADH." mlytical Biochemistry 305: 118—1 19.

Begum, A., et al. (1993). "New Electrocatalytic Biomolecular Interface for Fabricating a
Fructose Dehydrogenase-Based Sensing System." Analjaica Chimica Acta
280(1): 31-36.

Bioanalytical Systems (2005).
http://www.bioanalytical.com/products/ec/faqele.html#Three.

Bower, C. K., et al. (2001). "Lantibiotics as surface active agents for biomedical
applications." Colloids and Surfaces B- Biointerfam 22(4): 259-265.

Bower, C. K. M., J .; Daeschel, M. A. (1995). "Suppression of Listeria-Monocytogenes
Colonization Following Adsorption of Nisin onto Silica Surfaces." Applied and
Environmental Microbiology 61(3): 992-997.

Breukink, E. and B. de Kruijff (I999). "The lantibiotic nisin, a special case or not?"
133

Biochimica Et Biophvsfi Acta-Biomembrafls 1462(1-2): 223-234.

Breukink, E., et al. (1998). "The orientation of nisin in membranes." Biochemistry
37(22): 8153-8162.

 

Breukink, E., et al. (1999). "Use of the cell wall precursor lipid II by a pore-forming
peptide antibiotic." Science 286(5448): 2361-2364.

 

Brinkley, M. (1992). "A Brief survey of methods for preparing protein conjugates with
dyes, haptens, and cross-linking reagents." Bioconiugate Chem. 3: 2-13.

Campuzano, S., et al. (2003). "An integrated electrochemical fructose biosensor based on
tetrathiafulvalene-modified self-assembled monolayers on gold electrodes."
Apalytical and Bioanalytical Chemistry 377(4): 600-607.

Campuzano, S., et al. (2004). "An integrated bienzyme glucose oxidase-fructose
dehydrogenase-tetratl'riafulvalene-3-mercaptopropionic acid-gold electrode for the
simultaneous determination of glucose and fi'uctose." Bioelectrochemistry 63(1-
2): 199-206.

Candeias, L. P., et al. (1998). "The catalysed NADH reduction of resazurin to resorufin."
Journal of the Chemical Society-Perkin Transactions 2(11): 2333-2334.

Carnes, E. and E. Wilkins (2005). "The Development of A New, Rapid, Amperometric
Irnmunosensor for the Detection of Low Concentrations of Bacteria Part I: Design

of the Detection System and Applications." American Journal of Applied
Sciences 2(3): 597-606.

Castillo, J ., et al. (2004). "Biosensors for life quality - Design, development and
applications." Sensors and Actuators B-Chemica_l 102(2): 179-194.

Chapman, J. and M. Zhou (1999). "Microplate-based fluorometric methods for the
enzymatic determination of l-glutamate: application in measuring l-glutamate in
food samples." Analytica Chimica Acta 402: 47—52.

Chaubey, A., et al. ( 2000). "Immobilization of lactate dehydrogenase on
electrochemically prepared polypyrrole-polyvinylsulphonate composite films for

134

application to lactate biosensors." Electrochimica Acta 46(5): 723-729.

 

Chaubey, A. and B. D. Malhotra (2002). "Mediated biosensors." Biosensors &
Bioelectronics 17(6-7): 441-456.

Ciucu, A. and C. Ciucu (2002). "Organic Phase Amperometric Biosensor for Detection of
Pesticides." Roum. Biotechnol. Lett. 7(2): 667-676.

Cush, R., et a1. (1993). "The Resonant Mirror - a Novel Optical Biosensor for Direct
Sensing of Biomolecular Interactions .1. Principle of Operation and Associated
Instrumentation." Biosepaors & Bioelectronics 8(7-8): 347-353.

Daly, S. M., et al. (2003). "Coverage-dependent orientation of lysozyme adsorbed on
silica." Langmuir 19(9): 3848-3857.

Darder, M., et al. (2000). "Biosensors based on membrane-bound enzymes immobilized
in a 5-(octyldithio)-2-nitrobenzoic acid layer on gold electrodes." Analjaical
Chemist_ry 72(16): 3784-3792.

Davis, K. A. and T. R. Leary (1989). "Continuous Liquid-Phase Piezoelectric Biosensor
for Kinetic Irrrmunoassays." _A_n_alytical Chemim 61(11): 1227-1230.

Dayeh, V. R., et a1. (2004). "Evaluating the toxicity of Triton X-100 to protozoan, fish,
and mammalian cells using fluorescent dyes as indicators of cell viability."
Ecotoxicology_and Environmental Safeg 57: 375—3 82.

DiazGarcia, M. E. and M. J. ValenciaGonzalez (1995). "Enzyme catalysis in organic
solvents: A promising field for optical biosensing." Talanta 42(11): 1763-177 3.

Dickinson, E. (1977). Proteins at Liquid Interfaces. Introduction to Food Colloids.
Oxford University Press. Chapter 6: 140-173.

Dickinson, E. (1994). "Protein-Stabilized Emulsions." Journal of Food Engineering 22(1-
4): 59-74.

Dickinson, E. (1998). "Proteins at interfaces and in emulsions - Stability, rheology and
interactions." Journal of the Chemical Society-Faraday Transactions 94(12):
135

1657-1669.

Dickinson, E., et a1. (1993). "A neutron reflectivity study of the adsorption of .beta.-
casein at fluid interfaces." Langmuir 9: 242-248.

Dzyadevych, S. V., et al. (2002). "Assessment of the toxicity of methyl parathion and its
photodegradation products in water samples using conductometric enzyme
biosensors." Analjaica Chimica Acta 459(1): 33-41.

Enserink, M. (1999). "Microbiology - Promising antibiotic candidate identified." Science
286(5448): 2245-+. -

Fang, F. and I. Szleifer (2001). "kinetics and thermodyanmics of Protein Adsorption: A
Generalized Molecular Theoretical Approach." Biophysical Jouma_l 80: 2568-
2589.

Faulkner, L. R. (1983). "Understanding Electrochemistry: Some Distinctive Concepts."
Journal of Chemical Education 60: 262-264.

Ferri, T., et al. (1998). "Direct electrochemistry of membrane-entrapped horseradish
peroxidase. Part II: Amperometric detection of hydrogen peroxide."
Bioelectrochemistry An_d Bioenergetics 45(2): 221-226.

Gajraj, A. (1999). A TIRF technique for investigating macromolecular adsorption and
interactions at the liquid-liquid interface. Chemical Engineering. East Lansing,
Michigan State University: 1-153.

Gajraj, A. and R. Y. Ofoli (2000a). "Quantitative technique for investigating
macromolecular adsorption and interactions at the liquid-liquid interface."
Langmuir 16(9): 4279-4285.

Gajraj, A. and R. Y. Ofoli (2000b). "Effect of extrinsic fluorescent labels on diffirsion
and adsorption kinetics of proteins at the liquid-liquid interface." Langr_n_uir
16(21): 8085-8094.

Gerard, M., et a1. (2002). "Application of conducting polymers to biosensors." Biosensors

136

 

& Bioelectronics 17(5): 345-359.

Giffard, C. J ., et al. (1996). "Interaction of nisin with planar lipid bilayers monitored by
fluorescence recovery after photobleaching." J. Membrane Biol. 151: 293-300.

Girard, P., et al. (2004). "A new method for the reconstitution of membrane proteins into
giant unilamellar vesicles." Bipphvsical Journal 2004: 419-429.

Graham, D. E. and M. C. Phillips (1979). "Proteins at Liquid Interfaces. 3. Molecular-
structures of adsorbed films." Journal of Colloid and Interface Science 70(3): 427-
439.

Graneli, A., et al. (2003). "Formation of supported lipid bilayer membranes on SiOz from
proteoliposomes containing transmembrane proteins." Langmuir 19: 842-850.

Haas, A. S., et al. (2001). "Cytochrome c and cytochrome c oxidase: Monolayer
assemblies and catalysis." JoumaL of Physical Chemistry B 105(45): 11351-
11362.

Hunter, J. R., et al. (1991). "Beta-casein adsorption at the air-water interface." Journal of
Colloid and Interface Science 142(2): 429-446.

Hurst, A. (1981). Nisin. Advances in Applied Microbiology, Academic Press. 27: 85-
123.

Hynes, R. O. (1992). "Integrins - Versatility, Modulation, and Signaling in Cell-
Adhesion." Q1 69(1): 11-25.

Ivnitski, D., et al. (1999). "Biosensors for detection of pathogenic bacteria." Biosensors &
Bioelectronics 14(7): 599-624.

Jeon, S. l. and J. D. Andrade (1991). "Protein-Surface Interactions in the Presence of
Polyethylene Oxide." Journal of ColloiL and Interface Science 142(1): 159-166.

Karakashev, D., et a1. (2003). "A simple and rapid test for differentiation of aerobic from
anaerobic bacteria." World Journal of Microbiology & Biotechnology 19: 233-

137

238.

Karlsson, R. and A. Falt (1997). "Experimental design for kinetic analysis of protein-
protein interactions with surface plasmon resonance biosensors." Journal of
Immunological Methods 200(1-2): 121-133.

Kato, K., et al. (2003). "Enzymatic activity and stability of D-fructose dehydrogenase and
sarcosine dehydrogenase immobilized onto giant vesicles." Biotechnology and
Bioengineering 84(4): 415-423.

Kelley, S. O., et al. (1999). "Long-Range Electron Transfer through DNA Films."
Angew. Chem. Int. Ed. 38(7): 941-945.

Khan, G. F., et al. (1991). "Electrochemical-Behavior of Monolayer Quinoprotein
Adsorbed on the Electrode Surface." JournA of Electroanalytical Chemistry
315(1-2): 263-273.

Kinnear, K. T. and H. G. Monbouquette (1993). "Direct Electron-Transfer to Escherichia-
Coli Furnarate Reductase in Self-Assembled Alkanethiol Monolayers on Gold
Electrodes." Langmuir 9(9): 2255-2257.

Kinnear, K. T. and H. G. Monbouquette (1993). "Micelle-Supported Electroenzymology
- Succinate Dehydrogenation by Escherichia-Coli F umarate Reductase in
Decylubiquinone and Octyl Glucoside Micelles." Biotechnology and
Bioengineering 42(1): 140-144.

Kinnear, K. T. and H. G. Monbouquette (1997). "An amperometric fructose biosensor
based on fructose dehydrogenase immobilized in a membrane mimetic layer on
gold." Analytical Chemistry 69(9): 1771-1775.

Kissinger, P. T. and W. R. Heineman (1983). "Cyclic Voltammetry." Journal of Chemical
Education 60(9): 702-706.

Kissinger, P. T. and W. R. Heineman ( 1984). _L_aboratory Techniques in Electroapalytical
Chemispy, Marcel Dekker Inc.

Lakarnraju, M., et al. (1996). "Nisin adsorption and exchange with selected milk proteins

138

 

at silanized silica surfaces." Journal of Colloid and Interface Science 178(2): 495-
504.

Lee, W. K., et a1. (2000). "Interfacial tension kinetics of nisin and beta-casein at an oil-
water interface." Korean Journal of Chemical Engineering 17(2): 179-183.

Li, J ., et al. (1996). "Silica sol-gel immobilized amperometric biosensor for hydrogen
peroxide." Analflica Chimica Acta 335(1-2): 137-145.

Li, J. H., et al. (1996). "Direct electron transfer to cytochrome c oxidase in self-
assembled monolayers on gold electrodes." Journal of Electroanalytical;
Chemistg 416(1-2): 97-104.

Lin, V. S. Y., et al. (1997). "A porous silicon-based optical interferometric biosensor."
Science 278(5339): 840-843.

Mabbott, G. A. (1983). "An Introduction to Cyclic Voltammetry." Journal of Chemical
Education 60(9): 697-701.

Maeda, H., et al. (2000). "Resorufin as an electron acceptor in glucose oxidase-catalyzed
oxidation of glucose." Chemical & Pharmaceuticaa Bulletin 48(7): 897-902.

Maeda, H., et al. (2001). "Resazurin as an electron acceptor in glucose oxidase-catalyzed
oxidation of glucose." Chemical & Pharmaceutical Bulletin 49(5): 622-625.

Mehrvar, M. and M. Abdi (2004). "Recent developments, characteristics, and potential

applications of electrochemical biosensors." Analytical Sciences 20(8): 1113-
1126.

Mehrvar, M., et al. (2000). "Fiber-optic biosensors - Trends and advances." Analjaical
Sciences 16(7): 677-692.

Meuller, J ., et al. (1997). "Properties of a cysteine-flee proton-pumping nicotinamide
nucleotide transhydrogenase." Biochemical Journal 324: 681-687.

Miao, Y., et a1. (2001). "Amperometric Glucose Biosensor Based on Immobilization of
Glucose Oxidase in Chitosan Matrix Cross-Linked with Glutaraldehyde."
139

Electroanalysis 13(4): 347-349.

Motta, N. and A. R. Guadalupe (1994). "Activated Carbon Paste Electrodes for
Biosensors." Anal. Chem. 66: 566-571.

Mullor, S. G., et a1. (1996). "Alcohol biosensor based on alcohol dehydrogenase and
Meldola Blue immobilized into a carbon paste electrode." Talanta 43(5): 779-784.

Munteanu, F., et al. (2003). "Electrochemical and catalytic investigation of carbon paste
modified with Toluidine Blue 0 covalently immobilised on silica gel." Analjaica
Chimica Acta 476: 43-54.

Muramatsu, H., et al. (1987). "Piezoelectric Crystal Biosensor Modified with Protein-a
for Determination of Irnmunoglobulins." Apalytical Chemistry 59(23): 2760-
2763.

Napier, M. E., et al. (1997). "Probing Biomolecule Recognition with Electron Transfer:
Electrochemical Sensors for DNA Hybridization." Bioconiugate Chem. 8: 906-
913.

O'Brien, J ., et al. (2000). "Investigation of the Alamar Blue (resazurin) fluorescent dye
for the assessment of mammalian cell cytotoxicity." Eur. J. Biochem. 267: 5421-
5426.

Palecek, E., et al. (1998). "Electrochemical biosensors for DNA hybridization and DNA
damage." Biosensors & Bioelectronics 13(6): 621-628.

Paredes, P. A., et al. (1997). "Amperometric mediated carbon paste biosensor based on
D-fructose dehydrogenase for the determination of fructose in food analysis."
Biosensors & Bioelectronics 12(12): 1233-1243.

Parellada, J ., et al. (1996). "Amperometric flow injection determination of fructose in
honey with a carbon paste sensor based on fructose dehydrogenase." Analytica
Chimica Acta 330(1): 71-77.

Prodromidis, M. 1., Karayannis, MI. (2002). "Enzyme Based Amperometric Biosensors

140

 

for Food Analysis." Electroapalysis 14(4): 241-261.

Raiteri, R., et al. (2001). "Micromechanical cantilever-based biosensors." Sensors and
Actuators B-Chemical 79(2-3): 115-126.

 

Ramanathan, K. and B. Danielsson (2001). "Principles and applications of thermal
biosensors." Biosensors & Bioelectronics 16(6): 417-423.

Rao, T. N., et al. (1999). "Electrochemical Oxidation of NADH at Highly Boron-Doped
Diamond Electrodes." Anal. Chem. 71: 2506-251 1.

Reinheimer, J. A. and M. R. Demkow (1990). "Comparison of rapid tests for assessing
UHT milk sterility." J Daig Res. 57(2): 239-243.

Riss, T. L. and R. A. Moravec (2004). "Use of Multiple Assay Endpoints to Investigate
the Effects of Incubation Time, Dose of Toxin, and Plating Density in Cell-Based
Cytotoxicity Assays." ASSAY and Drug Development Technologie_s 2(1): 51-62.

Robeson, J. L. and R. D. Tilton (1996). "Spontaneous reconfiguration of adsorbed
lysozyme layers observed by total internal reflection fluorescence with a pH-
sensitive fluorophore." Langipuir 12(25): 6104-6113.

Russev, S. C., et a1. (2000). "Beta-casein adsorption kinetics on air-water and oil-water
interfaces studied by ellipsometry." COLLOIDS AND SURFACES B-
BIOINTERFACES 19(1): 89-100.

Saini, S., et al. (1991). "Organic-Phase Enzyme Electrodes." Analflica Chimica Acta
249(1): 1-15.

Sanocki, V. A. (2004). Cell culture at the liquid-liquid interface : effects of fluorocarbon
layer thickness on cell growth. Chemical Engineering & Materials Science. East
Lansing, Michigan State University.

Sigma, A. (1994). Enzymatic assay of D-fi'uctose dehydrogenase, Sigma Aldrich.

Smith, B. A. and H. M. McConnell (1978). "Determination of Molecular-Motion in
Membranes Using Periodic Pattern Photobleaching." Proceedings of the National
141

 

 

Academy of Sciences of the United States of America 75(6): 2759-2763.

Starr, T. E. and N. L. Thompson (2002). "Fluorescence pattern photobleaching recovery
for samples with multi-component diffusion." Biophysical Chemistry 97(1): 29-
44.

Thompson, N. L., et al. (1981). "Measuring surface dynamics of biomolecules by total
internal reflection fluorescence with photobleaching recovery or correlation
spectroscopy." Biophysical Journal 33: 435-454.

Tkac, J., et al. (2001). "Fructose biosensor based on D-fructose dehydrogenase
immobilised on a ferrocene-embedded cellulose acetate membrane." Analytica
Chimica Acta 439(1): 39-46.

Toyobo, E. D-Fructose Dehydrogenase.

Tratnyek, P. G., et al. ( 2001). "Visualizing Redox Chemistry: Probing Environmental
Oxidation-Reduction Reactions with Indicator Dyes." Chem. Educat_or_' 6: 172-
179.

 

T upy, M. J ., et al. (1998). "Total internal reflection fluorescence spectrometer to study
dynamic adsorption phenomena at liquid/liquid interfaces." Industrial &

Engineering Chemist_ry Research 37(8): 3159-3168.

Vaidya, S. and R. Y. Ofoli (2005). "Adsorption and Interaction of Fibronectin and
Human Serum Albumin at the Liquid-Liquid Interface." Langr_nuir 21(13): 5852 -
5858.

Vaidya, S. S. (2005). Adsorption and Interactions of Biomacromolecules at F luid-Like
Interfaces. Chemical Engineering & Materials Science. East Lansing, Michigan
State University.

Van Benschoten, J. J., et al. (1983). "Cyclic Voltammetry Experiment." Journal of
Chemical Education 60(9): 772-776.

Van heusden, H. E., et al. (2002). "Lipid II Induces a Transmembrane Orientation of the

142

 

Pore-Forming Peptide Lantibiotic Nisin." Biochemistry 41: 12171-12178.

Van Kraaij, C., et al. (1998). "Pore formation by nisin involves translocation of its C-
terminal part across the membrane." Biochemistry 37(46): 16033-16040.

 

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

Wagner, M. L. and L. K. Tamm (2000). "Tethered polymer-supported planar lipid
bilayers for reconstitution of integral membrane proteins: Silane-
polyethyleneglycol-lipid as a cushion and covalent linker." Biophysical Journal
79(3): 1400-1414.

Wang, B., et al. (1998). "Amperometric Glucose Biosensor Based on Sol-Gel Organic-
Inorganic Hybrid Material." Anal. Chem. 70: 3170-3 174.

Wang, H., et al. (2005). "A Biosensor Based on Immobilization of Horseradish
Peroxidase in Chitosan Matrix Cross-linked with Glyoxal for Amperometric
Determination of Hydrogen Peroxide." Sensors 5: 266-276.

Wang, S., et al. (1998). "Resazurin reduction assay for ram sperm metabolic activity
measured by spectrophotometry." Proceeding of the Society for Experimental
Biology and Medicine 217: 197-202.

Wertz, C. F. and M. M. Santore (1999). "Adsorption and relaxation kinetics of albumin
and fibrinogen on hydrophobic surfaces: Single-species and competitive
behavior." Langmuir 15(26): 8884-8894.

Wertz, C. F. and M. M. Santore (2001). "Effect of surface hydrophobicity on adsorption
and relaxation kinetics of albumin and fibrinogen: Single-species and competitive
behavior." Langmuir 17(10): 3006-3016.

Wertz, C. F. and M. M. Santore (2002). "Adsorption and reorientation kinetics of
lysozyme on hydrophobic surfaces." Langmuir 18(4): 1190-1199.

Wertz, C. F. and M. M. Santore (2002). "Fibrinogen adsorption on hydrophilic and
hydrophobic surfaces: Geometrical and energetic aspects of interfacial

143

 

 

relaxations." Langmuir 18(3): 706-715.

Willner, I., et a1. (1997). "Assembly of functionalized monolayers of redox proteins on
electrode surfaces: Novel bioelectronic and optobioelectronic systems."
Biosensors & Bioelectronics 12(4): 337-356.

Wright, L. L., et al. (1988). "Inhomogeneous Translational Diffusion of Monoclonal-
Antibodies on Phospholipid Langmuir-Blodgett Films." Biophysical Joumal
54(3): 463-470.

Yarnada, Y., et al. (1966). "A New Enzyme D-Fructose Dehydrogenase." Ag’c. Biol.
Chem. 30: 95-96.

Yu, J ., et al. (2003). "Glucose sensor for flow injection analysis of serum glucose based
on immobilization of glucose oxidase in titania sol-gel membrane." Biosensors &
Bioelectronics 19: 401-409.

Yuan, Y., et a1. (2003). "Mobility of adsorbed proteins studied by fluorescence recovery
after photobleaching." Langrauir 19: 3705-371 1.

Zhang, H. X., et al. (2004). "Assay of mitochondrial functions by resazurin in vitro."
Acta Pharmacol Sin 25(3): 385-389.

Zhang, L. Q., et a1. (2000). "Phosphatidylserine/cholesteral bilayers supported on a
polycation/alkylthiol layer pair." Journal of Colloid and Intergce Science 228(1):
82-89.

Zrimsek, P., et al. (2004). "Spectrophotometric application of resazurin reduction assay to
evaluate boar semen quality." International journal of andrology 27: 57—62.

144

 

 

 

I11111111111:111