£53 WW“W“WNW!Ill!l1\Ul|\Hl\lll|\\l\\\\\|\i‘a|

   

r:~4ESlS

Kin-2°

llllllllllllllllHIHHHIHIH"llIllllllllllllllllllll

31293 01687 0994

This is to certify that the

thesis entitled
A study of the diffusion of bovine serum albumin
at the oil-water interface using total internal
reflection fluorescence and fluorescence photobleaching
recovery

presented by

Parijat Jauhari

has been accepted towards fulfillment
of the requirements for

M.S. degree in Chemical Engineering

 

441) (/0214 /

Mgér professo

Date flag lg; /[2? 2

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

 

LIBRARY

Michigan State
Unlversity

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/90 WM“

A STUDY OF THE DIFFUSION OF BOVINE SERUM ALBUIVIIN AT THE OIL-
WATER INTERFACE USING TOTAL INTERNAL REFLECTION
FLUORESCENCE AND FLUORESCENCE PHOTOBLEACIIIN G RECOVERY

By

Parijat Jauhari

A THESIS
Submitted to
Michigan State University
in partial fiilfillment of the requirements
for the degree of
MASTER OF SCIENCE

Department of Chemical Engineering

1997

III lli’é? ‘1

ABSTRACT

A STUDY OF THE DIFFUSION OF BOVINE SERUM ALBUMIN AT THE OIL-
WATER INTERFACE USING TOTAL INTERNAL REFLECTION FLUORESCENCE
AND FLUORESCENCE PHOTOBLEACHIN G RECOVERY

By

Parijat Jauhan'

We have combined total internal reflection fluorescence (TIRF) with fluorescence
recovery after spot photobleaching (FPR) and fluorescence recovery after pattern
photobleaching (FRAPP) to study the difi‘usion of bovine serum albumin (BSA) labeled
with fluorescein-isothiocyanate (FIT C) at the oil-water interface. The primary objective
was to adapt the TIRF setup used in studies of protein adsorption at the solid liquid
interface to make it suitable for the liquid-liquid interface.

We assessed the integrity of our setup by measuring the equilibrium adsorption and
diffusion of BSA at the oil-water interface. Diflusion coefficients were obtained by fitting
the fluorescence emission data to the difl‘usion equation, using non-linear regression. We
obtained a lateral diffusion coefficient of 3.4 x 10" cmz/s, with a mobile fraction of 0.48;
and an apparent bulk difliision coefficient of 5.2 x 10" cm’ls, with a mobile fraction of
0.77. The difference of one order of magnitude in the two results indicates that bulk
difl’usion of FITC-BSA is much faster than lateral difliision, as would be expected. Signal
noise and errors in measuring the photobleaching spot size strongly affected the values of

the diffusion coefiicients.

Dedicated to my parents, to India my country, and to my friends.

iii

Acknowledgments

I would like to acknowledge the help given to me by Dr. Ofoli, who was more than my
research advisor. He guided me in my work and exposed me to the philosophy of research.
I would also like to thank An' and Abhijit who worked with me to solve the intricacies of
our respective projects. Special thanks must be given to Dr. Perry Ng and his group who
let me borrow their electronic balance, to Dr. Kris Berglund for the use of his centrifuge

for the spin coating procedure, and to Dr. Alec Scranton for allowing me to use his

centrifuge.

iv

Table of Contents

ABSTRACT .................................................................................................................... ii
Acknowledgments .......................................................................................................... iv
Table of Figures ............................................................................................................. vii
Table of Tables ............................................................................................................. viii
1. Introduction ................................................................................................................. l
1.1 Motivation for studying the liquid-liquid interface .................................................. 1
1.2 Objectives of this study .......................................................................................... 2

2. Theoretical Considerations ........................................................................................... 4
2.1 Total internal reflection .......................................................................................... 4
2.2 Fluorescence Recovery After Photobleaching ......................................................... 5
2.2.1 Spot Photobleaching ....................................................................................... 7
2.2.2 Pattern Photobleaching .................................................................................... 7

3. Experimental Materials and Techniques ....................................................................... 8
3.1 Experimental equipment ......................................................................................... 8
3.1.1 Setup for TIRF experiments ............................................................................ 8
3.1.2 FRAPP Setup ................................................................................................ 10
3.1.3 The experimental cell ..................................................................................... 10

3.2 Material preparation procedures ........................................................................... 13
3.2.1 Protein labeling ............................................................................................. 13
3.2.2 Coating of bottom slide in experimental cell ................................................... 14
3.2.3 Immobilization of fluorophore ....................................................................... 14

3.3 Experimental Techniques ..................................................................................... 14
3.3.1 Formation of the Oil-water interface .............................................................. 14
3.3.2 Introduction of proteins into the experimental cell ......................................... 15
3.3.3 FPR experimental protocol ............................................................................ 15

3.3.4 FRAPP experimental protocol ....................................................................... 16

3.3.5 Experimental precautions .............................................................................. 16

3 .4 Determination of interfacial concentrations of fluorophores on the basis of
fluorescence emission data ......................................................................................... 19
3.5 Data analysis procedures ...................................................................................... 23
3.6 Calculation of the amplitude of the fringe pattern ................................................. 27

4. Results and Discussion ............................................................................................... 29
4.1 Check of the TIRF setup using steady state adsorption of BSA-FIT C ................... 29
4.2 Results of spot photobleaching experiments ......................................................... 33
4.3 Results of pattern photobleaching experiments ..................................................... 36
4.4 Discussion ............................................................................................................ 40
4.5 Possible sources of errors ..................................................................................... 41

5. Summary and Conclusions ......................................................................................... 43
6. Suggestions for fixture work ....................................................................................... 45
6.1 Difliision of surface-bound species ....................................................................... 45
6.2 Diflirsion of bulk species ...................................................................................... 45
6.3 Accounting for concentration quenching .............................................................. 46

7. Bibliography .............................................................................................................. 48
8. Appendix: Mathcad program ..................................................................................... 51

Table of Figures

Figure 1. Total internal reflection induced by a laser beam incident on an interface made
up of materials of different refiactive indices. ........................................................... 6
Figure 2. Setup for total internal reflection fluorescence experiments, showing the optical

components used to direct the beam to the oil/water interface. ................................. 9
Figure 3. Optical configuration used for FRAPP experiments (after Robeson 1995). ..... 11
Figure 4. Components of the experimental cell ............................................................... 12
Figure 5. Protein adsorption experiment conducted to assess the effect of flow rate on the

integrity of the oil-water interface ........................................................................... 18
Figure 6. Graph used to calculate the amplitude of the fiinge pattern. ............................. 28
Figure 7. Fluorescence emission profile during adsorption of BSA at the oil-water

interface. ................................................................................................................ 30
Figure 8. Fluorescence emission profile of BSA ............................................................ 32

Figure 9. Actual, averaged and fitted fluorescence emission versus time after

photobleaching. ...................................................................................................... 34
Figure 10. Fluorescence recovery after photobleaching for a spot size of 3 1.25 pm ....... 34
Figure 11. Fluorescence recovery after photobleaching for a spot size of 31.25 um ....... 35
Figure 12. Fluorescence recovery after photobleaching for a spot size of 3 1.25 urn. ....... 35
Figure 13. Normalized fluorescence recovery afier pattern photobleaching with a fiinge

width of 6.25 pm. .................................................................................................. 37
Figure 14. Normalized fluorescence recovery afler pattern photobleaching with a fringe

width of 5.68 um. .................................................................................................. 37
Figure 15. Normalized fluorescence recovery after pattern photobleaching with a fringe

width of5.21 um ................................................................................................... 38
Figure 16. Characteristic time of recovery versus the square of the fiinge half period ..... 38

vii

Table of Tables

Table 1. Difiirsion coeflicients and mobile fi'actions obtained from three replicate spot
photobleaching experiments. .................................................................................. 33
Table 2. Characteristic times and mobile fractions fi'om pattern photobleaching
experiments ............................................................................................................ 39

viii

1. INTRODUCTION
1.1 Motivation for studying the liquid-liquid interface

The liquid-liquid interface is fundamentally important in many biological systems. For

example:

1. Liquid-liquid interfaces may have played a role in the origin of life. It has been
suggested that the first living cells probably utilized lipid-like hydrocarbon derivatives
for their membrane structures. The chemical and physical properties of these oil/water-
like interfaces could have contributed to the mechanisms by which energy captured
fi'om the environment was used to control photochemical reactions (Deamer and

Volkov 1996);

2. the behavior of drugs at biological membranes is an important factor in determining
pharmacological activity. Irnmiscible oil-water interfaces could serve as an artificial
model of biomembranes and can, therefore, be used in drug delivery experiments

(Arai er a1. 1996); and

3. the distribution and dynamics of proteins at the liquid-liquid interface is pertinent to
many technological processes. In particular, protein adsorption at the liquid-liquid
interface is an important component of many processes in the food and pharmaceutical
processing industries where proteins are extensively used as stabilizers in emulsions
and microemulsions. Information on amphiphilic behavior at the oil-water interface
could be used to model these systems. Secondly, studies of protein adsorption at the
liquid-liquid interface allow us to gain a better understanding of amphiphile dynamics
at this important interface, making it possible to develop better processing and
characterization techniques.

Various techniques have been used for studying amphiphiles at interfaces. These include
scintillation counting, used by Graham et a1. (1975) to obtain the concentration of proteins
at the oil-water interface; monitoring of changes in interfacial tension by Davies and
Mayers (1960); Raman scattering, used by Takenaka and Nakanaga (1976), Nakanaga and
Takenaka (1977) and Takenaka (197 8) to infer the orientation of molecules after
adsorption of dyes at the chloroform-water interface; and total internal reflection
fluorescence spectroscopy, used by Hirschfield (1965) to examine the fluorescence of dye

molecules near an interface.

We have adopted total internal reflection fluorescence (TIRF) as the primary tool in this
study, because it is the most suitable for molecular-level investigations of amphiphiles at
interfaces. TIRF allows one to excite fluorescent molecules in close proximity to the
interface, in preference to those in the bulk. One can then measure the concentration of the
fluorophore as a function of distance from the interface, and study the adsorption
equilibria and diffusion kinetics of species at the interface. However, most of the
published research so far utilizing TIRF has used the technique to study the adsorption
and interaction of protein molecules at the solid-liquid interface. Our goal was to modify
the standard TIRF technique to make it suitable for application at the liquid-liquid

interface.

1.2 Objectives of this study

The specific objectives of this project were to:

I. modify the total internal reflection fluorescence technique used for the solid-liquid
interface to make it suitable for application at the liquid-liquid interface, and to
incorporate capabilities for fluorescence photobleaching recovery (TIRF/FPR) and
fluorescence recovery after pattern photobleaching (TIRF/FRAPP);

2. establish qualitatively the relative importance of bulk and surface difi‘usion of a
representative protein at the liquid-liquid interface, using both pattern photobleaching
and spot photobleaching; and

3. determine the lateral and bulk diffusion coeflicients of bovine serum albumin (BSA) at
the oil-water interface, using both spot (FPR) and pattern photobleaching (FRAPP).

2. THEORETICAL CONSIDERATIONS

Total internal reflection fluorescence (TIRF) was first used by Tweet et al. (1964) to
measure the emission spectrum and quenching behavior of chlorophyll a monolayers at
the air-water interface. TIRF was used to exclude the signal fi'om direct and scattered
mercury arc excitation, so only the weak fluorescence generated by the dilute monolayer
would be detectable. TIRF has also been used at the oil-water interface by Morrison and
Weber (1987) to study the adsorption of 4,4-bis-l-phenylamino-8-napthalenesulfonate
(bis-AN S) at the decalin-water interface.

The theory of total internal reflection fluorescence has been described in detail in many
references, such as Axelrod et al.(1981), Axelrod (1990), and Axelrod et al. ( 1992). The
theory of spot photobleaching has been described by Axelrod et al. (1976), Barisas and
Leuther (1979), Elson (1985), Lanni and Ware (1984), and Webb and Thomas (1990).
Fluorescence recovery afier pattern photobleaching has been described by Abney et al.
(1992), Robeson (1995), Robeson and Tilton (1995), and Tilton et a1. (1990 a, b).
Therefore, only the details relevant to this study would be presented here.

2.1 Total internal reflection

When a beam of light is incident upon an interface made up of materials of different
refi'active indices at an angle 9 greater than a given critical angle, it is totally internally

reflected in the optically denser medium (refi'active index of 12,). The critical angle, 9 c , is

given by

a, = sin-{fl} (n, >n,) [11
"x
As a result of this total internal reflection, an evanescent field is formed in the optically

4

As a result of this total internal reflection, an evanescent field is formed in the optically
rarer medium with an exponentially decaying intensity, as shown in Figure 1. The intensity

of the evanescent wave is given by

 

{ z
I=Ioexp -d— [2]
K

where I is the scattering intensity at 2, la is the scattering intensity at the interface and d

is the depth of penetration of the evanescent wave, which is given by
d = 1°
47m1 ‘lsin2 G—nn2
Here, 34. is the wavelength of the incident radiation in vacuum and n21 is the relative

refractive indices of the two media:

[3]

"21 = "i [41
"1

The exponentially decaying evanescent field is a powerful tool to selectively excite species

in the vicinity of an interface, and to study their interactions.

2.2 Fluorescence Recovery After Photobleaching

In this technique, a brief pulse of strong laser light is flashed at the interface,
photobleaching all of the fluorescent species in the field of the beam. The subsequent
recovery of the signal towards the prebleach level by exchange of bleached non-
fluorescent species with unbleached fluorescent species is observed using a monitoring

beam (a laser beam of much less intensity).

The photobleaching recovery characterizes the rate at which the unbleached molecules
diffuse, hence it gives a measure of the transport characteristics of the species. In this

study, two modes of photobleaching (spot photobleaching and pattern photobleaching)

Laser beam

 

- n1
011 nl > n2
Water

‘4‘
v

Evanescent Wave

   

Figure 1. Total internal reflection induced by a laser beam incident on an interface made
up of materials of different refiactive indices.

were used to obtain information about the lateral and bulk diffusion coefficients of BSA
proteins tagged with the FITC fluorophore.

2. 2. 1 Spot Photobleaching

In spot photobleaching, a small spot is bleached at the interface. The photobleached spot
has a characteristic length I and a recovery time constant of I. These two variables are
related to the difi‘usion coeficient D by

12

D [5]

7

Thus, if we vary the length of the spot and measure the recovery time, then we can use the
plot of 1: versus I2 to obtain the surface difi‘usion coefficient, Dsurfacc- Conversely, if we
vary the penetration depth of the evanescent wave (1, while keeping I constant, then a plot
of 1 versus a“ would enable us go get the bulk diffusion coeflicient, D”. This procedure
gives both qualitative and quantitative information about the relative importance of the

two modes of diffusion.

2. 2.2 Pattern Photobleaching

In pattern photobleaching, two beams of equal intensity are intersected and totally
internally reflected at the interface. The interference pattern formed by the evanescent
waves fiom the two beams is used to photobleach molecules adsorbed at the interface.
The experimental setup for pattern photobleaching is more complicated than the one for
spot photobleaching. However, it has two advantages over spot photobleaching. It allows
one to more accurately control the characteristic length of the illuminated region by
varying the angle of incidence. It also allows the measurement of smaller length
dimensions, compared to spot photobleaching (Abney et al.1992).

3. EXPERINIENTAL MATERIALS AND TECHNIQUES

3.1 Experimental equipment

3.1.1 Setup for TIRF experiments

The TIRF setup consists of a 5-W Lexel 95 continuous wave argon ion laser (Lexel
Lasers, Inc., Fremont, CA), an inverted microscope (Axiovert 135 M, Carl Zeiss Inc.,
Thomwood, NY) and a photomultiplier tube (PMT) system using a modular automation
controller (MAC 2000, Ludl Electronic Products Ltd., Hawthorne, NY) with a
Harnamatsu R928 standard tube. Using an arrangement of flats and mirrors, as shown in
Figure 2, the laser beam was split into a monitoring beam and a photobleaching beam and
directed at the interface where the fluorescence emission was collected by a PMT.

The S-W argon ion laser allowed us to photobleach and monitor the interface with the
same light source. Using a microscope allowed us to incorporate capabilities for bright-
field, dark-field, phase contrast and epi-illumination techniques into the setup, if needed.
Using one flat to split the laser beam into a monitoring and photobleaching beam and then
recombining the beams using another flat ensures a perfect alignment of the monitoring
beam and the photobleaching beam (Thomas and Webb1990).

A shutter (D122, Vincent Associates, Rochester, NY) allowed us to control which light
beam would reach the interface. When the shutter was closed, only the monitoring beam
was directed to the interface; when it was open, both the photobleaching and monitoring
beams are focused on the interface. To make this possible, we cut a hole in the shutter to
enable the monitoring beam to be always incident on the interface while making
measurements. To increase the signal to noise ratio, the PMT was jacketed inside a
water-cooled housing system (TE177TSRF, Products for Research, Inc., Danvers, MA)
which maintained a constant temperature around the tube and reduced the noise level.

8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0
I ._........I %fi
F in
E
*8.
Oil a g
9s .
i i
1:: a 5
if
a i
3
g/ g E
/ / a
g // é :
. a
E‘ a 53 .. :1
° 3% “I § §
"‘ ,8 a
g 5
g unfit-«IL E r?
«E "”3 ' i
8 E 1L .5
E O
5. 3.
8 a,
U

 

 

 

Figure 2. Setup for total internal reflection fluorescence experiments, showing the optical
components used to direct the beam to the oil/water interface.

10

We used a computerized data acquisition system with the Viewdac software (Keithley
Instruments, Inc., Rochester, NY) for data collection. This allowed us to gather a large
quantity of data over short periods of time. A CCD camera (MTI, VB 1000, Carl Zeiss
Inc., Thomwood, NY ) was used to focus the laser beam into the experimental cell
without having to look through the eyepiece. We also have the capability to take pictures
of the interface using a 35-mm camera (Contax 167MT, Yashica Inc., Somerset, NJ).

The whole setup was housed on a vibration isolation table (RS 4000, Newport
Corporation, Irvine, CA). The table floats on a cushion of air and prevents building
vibrations from destabilizing the oil-water interface.

3.1.2 FRAPP Setup

The setup shown in

Figure 3 was adapted fi'om the apparatus used by Robeson (1995) to measure surface
difi‘usion of bovine serum albumin on polymeric substrates. The laser beam was split into
two beams of equal intensity, using a 1-inch cubic non-polarizing beamsplitter (Newport
Corporation, Irvine, CA). The two beams were then directed to meet at the interface using
an arrangement of two mirrors, a cubic beamsplitter and a focusing lens. One of the
mirrors was mounted on a miniature piezo-flexure mirror mount (17 ASM 001,Melles
Griot, Inc., Boulder, CO) and was driven by a piezo electric controller (17 PCS 001,
Melles Griot). This arrangement allowed fine focusing of the two beams so they would
intercept at the interface. It was also used to calculate the amplitude of the interfering
beam pattern as described later in Chapter 4.

3.1.3 The experimental cell

The experimental cell shown in Figure 4a forms the heart of the setup. It consists of a top
microscope slide (Fisher Scientific, Pittsburgh, PA) optically coupled with an immersion
oil (16482, type A, RP. Cargille Laboratories, Inc., Cedar Grove, NJ) to a dovetail prism

11

 

 

 

 

 

 

 

 

 

 

 

352
M3 to
A
L \< BS1
PM
flowcell
/
\
r1 objective

BSl,BSI-Beamquitters,m,m.m -Minurs,L-Forusshgbns,HJ-mlrmrnnunled
ontlnmmtbxnnmrtarrlconnettedtolhepleroehrutdrlrer

Figure 3. Optical configuration used for FRAPP experiments (after Robeson 1995). The
setup produces two intersecting beams at the interface to create the fiinges required for
pattern photobleaching.

 
 
 

12

THE FLOWCELL

afrwviu

 

 

 

 

 

 

 

hmpviu

 

 

(.

 

 

 

 

O-ring

Figure 4. Components of the experimental cell. While all experiments were conducted
under stopped-flow conditions, the cell can also be used as a standard flow cell.

13

of angle 64°. The bottom of the top slide is coated with the same immersion oil and serves
as the oil part of the interface (the complete formation of the oil-water interface is given in
the next section). Below the top slide, we have an aluminum spacer of the same

dimension as the slide, and a bottom slide to complete the cell assembly.

The spacer has a channel cut into it (Figure 4b). When the channel is filled with water or
protein solution and the top microscope slide is placed over the cell, an oil-water interface
is formed with the oil coating on the bottom of the top slide. An O-ring (Parker Seals,
NJ) is inserted into the spacer channel, and used to seal the cell to prevent leakage of the
protein solution. The bottom slide has two holes drilled into it (Figure 4b) which allow us
to connect a syringe pump to the experimental cell for infusing and withdrawing the
protein solution. The whole assembly of the prism, top slide, aluminum spacer and the
bottom slide is tightly screwed on to an anodized aluminum shell, and sits on the top of the

microscope stage during experiments.

While all experiments were conducted under stopped-flow conditions, the experimental

cell can also be used as a standard flow cell.

3.2 Material preparation procedures
3.2.1 Protein labeling

For the purpose of studying interactions at the oil-water interface, bovine serum albumin
(BSA) (A-7511, Sigma Chemical Co., St. Louis, MO) was labeled with Fluorescein-S-
isothiocyanate (F -1907, Molecular Probes, Eugene, Oregon). Also albumin, bovine-
fluorescein isothiocyanate (A-9771, Sigma Chemical Co., St. Louis, MO) with a labeling
ratio of approximately 12:1 was mixed with unlabelled BSA to obtain the appropriate
labeling ratios needed to avoid concentration quenching of the signal. A bufl‘er of
concentration 0.1 M and pH of 9 was made from Borax (B-9876, Sigma Chemical Co.),

and was used as in all the labeling reactions and for making all the protein solutions.

14

Protein concentrations were kept below the critical micelle concentration (cmc); therefore,
only single particle interactions have been characterized in this study. The procedure used
to label the proteins and calculate their labeling ratios has been described by McKinney et
al. (1965).

3.2.2 Coating of bottom slide in experimental cell

A 0.5 wt% solution of polyethylene oxide (PEO) with a molecular weight of 8x10‘5
(37,283-8, Aldrich Chemical Company, Inc., Milwaukee, WI) was made with deionized
water. A clean microscope slide was soaked in the PEO solution for 2 hours, causing a
film to be formed on the quartz slide. The slide was then washed with water to remove any
excess PEO. The film coating on the bottom slide of the sample cell was applied to
prevent the adsorption of proteins at the bottom interface formed by the glass and water
(Cheng et al. 1987).

3. 2.3 Immobilization of fluorophore

A 15 wt % solution of Gelatin (Knox, Nabisco Inc., Hanover, NJ) was made in warm
Borax bufl‘er (0.1 M, pH ~ 9). Approximately 1.5x10'3 M FITC was reacted with the
boiling gelatin solution for about one hour. The moist solution was then spin-cast onto a
clean slide. The slide, with the immobilized fluorophores, was used in experiments to
calculate the amplitude of the fringes and the bleaching parameter k (Robeson, 1995), as
described later. The slide was also used in the calibration process to calculate the
instrument constant of the TIRF setup.

3.3 Experimental Techniques
3. 3.1 Formation of the Oil-water interface

The oil-water interface was formed with an immersion oil ( 16482, type A, RP. Cargille
Laboratories, Inc., Cedar Grove, NJ) in contact with distilled deionized water. The
following procedure was used in preparing the interface:

15

1. Each slide was cleaned using the procedure described by Cheng et al. (1987), and
heated for 15 minutes under a heat lamp.

2. A drop of oil was placed in the center of the slide, where it quickly spread over the
slide to form a thin film of oil.

3. The slide was then optically coupled to a prism with the same immersion oil used to

form the oil-water interface.

4. The slide and the prism were then placed on top of the sample cell filled with buffer or
protein solution to form the oil-water interface. We estimated the oil layer in contact

with the water to be approximately 100 um thick.

3. 3.2 Introduction of proteins into the experimental cell

The sample cell was first filled with a buffer solution to establish the oil-water interface.
Then the protein solution was pumped into the cell using a double syringe pump which
withdrew the buffer solution fi'om the cell at the same rate as the protein solution was
pumped. The solution was allowed to flow into the cell until the bufl‘er solution in the cell
has been completely displaced by protein solution. At that time, the syringe pump was
turned off, and a stable interface re-established. Our goal was to put the oil layer into
contact with a protein solution approximately 1 mm thick, which is thin enough to allow
us to simultaneously accomplish two objectives: 1) give an equilibrium adsorption time on
the order of minutes rather than hours; and 2) provide the ability to use hydrodynamic

shear to sweep the interface clean within a reasonable amount of time.

3. 3.3 FPR experimental protocol
The FPR experiments were conducted according to the procedure described by Axelrod

et. al (1976). The protein solution was introduced into the sample cell and adsorption was
allowed to proceed to steady state, as determined by a constant fluorescence signal. We

then photobleached a spot at the interface and measured the rate of fluorescence recovery.

16

3. 3.4 FRAPP evqrerimental protocol

All FRAPP experiments were conducted with 0.067mM FITC-BSA (labeling ratio of
1.06) adsorbed at equilibrium at the interface, with unlabelled BSA of the same
concentration in the bulk. First we passed Borax buffer through the cell. We then
introduced FITC-BSA at the same flow rate into the sample cell. Afier steady state had
been reached, we introduced unlabelled BSA at the same flow rate into the cell and
allowed the system to reach steady state, again as indicated by a constant fluorescence
signal (T ilton et. a1 1990). After steady state had been reached, we photobleached the

interface and observed the rate of fluorescence recovery.

The introduction of unlabeled BSA into the sample cell serves two purposes. It allows the
exchange of loosely bound FITC-B SA with unlabelled BSA, so only the mobility of
irreversibly surface bound proteins is measured. It also assures us that any signal obtained
is fi'om the irreversibly bound proteins at the interface and not from the bulk. Since we
allowed the system to reach steady state afler introducing unlabeled BSA, any drop in
fluorescence measured during photobleaching will only be due to the lateral exchange of
surface bound species. Essentially, we have assumed that any exchanges between
unlabeled BSA and FITC-BSA would have also reached a steady state.

The same experimental cell was used for all FRAPP experiments, simply by choosing
different spots, at the interface. Since the time constant for the exchange of species at the
interface by lateral diffusion is much larger than the time scale of our experiments, this

approach should not present a problem by introducing experimental artifacts.

3. 3. 5 Experimental precautions

The following precautions were observed during all FRAPP experiments:

1. The mirrors and optics were kept clean and fixed so that the Gaussian profile of the
laser beam is not disturbed and the fiinge pattern does not shifi during measurements.

2. The microscope was focused exactly at the interface so that the focal plane and the

plane where the fiinges are formed are the same.

17

There was always a concern about unintended photobleaching of adsorbates by the
monitoring beam, particularly when adsorption is very slow to reach equilibrium and the
monitoring beam is left on for an extended period of time (Axelrod 1990). A beam
chopper was used during long adsorption experiments to reduce the total amount of
radiation reaching the interface, and to prevent unintended photobleaching of the

fluorophores.

Flow cells such as our sample cell are routinely used in studies at the solid-liquid interface
without the risk of destroying the interface, even during flow. With the liquid-liquid
interface, however, there is a strong possibility of sweeping away adsorbed molecules as a
result of shear imposed on the interface by the flowing fluid. A double syringe pump which
simultaneously infirses and withdraws very minute quantities of liquid was used to prevent
shear fi'om clearing the interface. As a further precaution, nearly all the experiments were

conducted in stopped-flow situations.

To obtain the maximum allowable flow rate to be used for pumping protein solutions
during flow situations, we conducted a number of replicate experiments in which we:

1. Introduced the protein solution into the sample cell and measured the fluorescence

2. Then we introduced a bufl‘er solution and measured the fluorescence emission again.

3. We then reintroduced the same concentration of protein solution used in step 1, and

measured the fluorescence emission again.

At a flow rate of 0.17 ml/min, the fluorescence emissions fi'om introduction of protein in
steps 1 and 3 were statistically the same (fluorescence versus time, in arbitrary but
consistent units) as shown in Figure 5. Also no oil was found either in the syringe or the
tubing connected to the experimental cell. This assured us that we were not destroying the

oil-water interface during periods of flow.

 

‘—'niirl “reduction of protein
sotrtion He the flow col
A tapioca-m of proteh w ih
buffer
a rephcemerl of buffer in In
prote'n souion

 

 

 

 

 

tlrna(s)

Figure 5. Protein adsorption experiment conducted to assess the effect of flow rate on the
integrity of the oil-water interface. Fluorescence emissions from identical protein
solutions (introduced before and after the cell was filled with a buffer) were statistically
the same at flow rates of 0. 17 ml/min and lower.

19

3.4 Determination of interfacial concentrations of fluorophores on the basis of
fluorescence emission data

Devising a reliable calibration technique for the quantification of the surface concentration
of amphiphiles based on a given fluorescence signal is a difiicult problem at the liquid-
liquid interface, in comparison to the solid-liquid interface. A number of approaches have
been used to determine the interfacial concentrations at the solid-liquid interface which are
made possible by the fact that a) protein adsorption is irreversible at this interface, and b)
the interface is non-mobile and stable. These techniques include depletion of adsorbates in
the bulk phase (Burghardt and Axelrod 1981), use of non-adsorbing species (such as
polystyrene spheres) as a standard, use of scintillation counting coupled with TIRF (Lok
et al. 1983a,b) and combination of reflectometry with TIRF (Robeson 1995).

Unfortunately, none of these techniques will work at the liquid-liquid interface, due to two

limitations:

1. There is no reliable technique that can be used to drain the bulk solution from the
sample cell without incorporating adsorbed species as well as the oil layer in the

effluent, even if we assume that adsorption is irreversible at this interface.

2. Quantitation methods such as combining reflectometry or scintillation counting with
TIRF rely on the immobilization of proteins due to irreversibility of adsorption at the
solid-liquid interface. As a result, the solid can be removed from the sample cell and
the quantity of adsorbed species determined accurately by any one of several methods.

This approach is not feasible at the liquid-liquid interface.

To circumvent these difficulties, we propose a calibration technique here that uses the
experimental cell along with photobleaching recovery to determine the interfacial
concentration of adsorbed species in situ, without destabilizing the oil-water interface. The

theoretical basis for this technique is presented below.

The intensity of fluorescence emission, F, from proteins at the interface and in the bulk is

given by

20

F= II<z>C<z>g<z>dz [61

where I (z) is the intensity of the incident beam and varies with depth, C(z) is the
concentration of the fluorescent species, and g(z) is a function which depends on the
relative quantum yield of the fluorescent species and on instnrment characteristics. The
integral quantum yield of molecules adsorbed at the interface and those in the bulk can be
different. However, Zimmerman et al.(1990) and Robeson (1995) have shown that the
quantum yields of proteins adsorbed at the solid-liquid interface and in the bulk do not
differ appreciably. Also, at labeling ratios below the concentration at which quenching
occurs (as is the situation in this case), the relative quantum yield is constant. Therefore

we can rewrite equation [6] as

F: g[1(z)C(z)dz [7]

Substituting equation [2] into [6] we get

F =13]: e""'C(z)dz [8]

For immobilized fluorophores, we can make the following substitutions:
1. The evanescent wave extends to A, the thickness of the immobilized fluorophore film.
Therefore, we can change the upper limit of the above integral to this variable.

2. Because of the uniformity of the matrix, C (2), the concentration of the immobilized
fluorophores (BSA-FITC immobilized in gelatin) is uniform and can be set equal to C
at both the interface and in the bulk.

Incorporating these, we can rewrite equation [8] as

A -3. -3. A A
F=KCIe dCdz=KC[-de 4] =KC[-de " +d) [9]
0

0

21

where K is a proportionality constant which incorporates the relative quantum yield and
the intensity of the incident laser beam. K is referred to as the instrument constant, and is
unique to each setup.

Since A>>d, equation [9] reduces to

F=KdC [10]
The result is that, since F, d and C are all known, we can evaluate K.

For a protein adsorbing from a bulk solution to the oil-water interface, we can assume that
a simple monolayer is formed at the interface in equilibrium with the bulk concentration of
proteins (Zirnmennann et al. 1990). We can also assume that this layer of adsorbed
molecules has a thickness I and a surface concentration C, (units of mass or moles per
area). For a monolayer of proteins at the liquid-liquid interface, the thickness 1
corresponds approximately to the diameter of the hydrophilic head of the protein,
assuming that the hydrophobic segment of the protein penetrates into the oil layer and is,

therefore, not observable by the evanescent wave.

Incorporating this, we can rewrite equation [8] as a sum of the fluorescence emission fiom

the interface (0 s z s I) and from the bulk solution (I s z s 6) :

 

6 (s—l)
F=K[C,+[e 6 Quiz) [11]
l

where C, is the concentration of the fluorescent species in the bulk (units of mass or mole

per volume), and 6 is the depth of the sample cell.

We know that lis approximately 3nm (Zimmermann et. al 1990). Also, dis 138 nm for

our setup, and 6 is 1 mm. Therefore,

6 _(:—l)
[e d C,~dc, [121

i

 

22

Substituting [12] into [11] we get

F=K(C,+dC,,) [13]
If a bleaching pulse is flashed at the interface, the recovery of bulk emission should be
almost instantaneous, in comparison to the emission from the surface because of the
difl‘erent transport time scales for bulk and surface diffusion. This means that difi‘usion in
the bulk will lead to a rapid replacement of the bleached protein molecules by unbleached
ones, and will be observed as a partial recovery of fluorescence. Since this signal is

entirely fi'om the bulk, equation [13] becomes

 

FM = K d C, [14]
Combining equations [13] and [14], we can calculate the interfacial concentration of
amphiphiles:

C, = F ' FM [15]

K

Using Eq. [15], a plot of C, versus F will yield the surface concentration at any

fluorescence intensity.

Three assumptions are implicit in this analysis:
1. exchange of proteins adsorbed at the interface with the bulk species is very slow

compared to the time scales of the experiment (for both lateral and bulk difl‘usion);

2. K, which is the product of the instrument constant and the quantum yield of the
species, is constant in the specified concentration range and over the time scales of the
experiment; and

3. The glass-gelatin and oil-water interfaces have identical critical angles for total internal

reflection. This assumption is supported by Robeson (1995) who reported that, due to

its high water content, gelatin has essentially the same refractive index as water.

23

3.5 Data analysis procedures

The procedure developed by Axelrod et al. (1976) was used for analyzing the spot
photobleaching data, and the Robeson (1995) procedure was used for FRAPP analysis.
The theoretical basis of the two procedures are briefly outlined here.

When two coherent plane polarized beams intersect at an interface, the intensity of the
resulting interference pattern varies spatially as
I(x) == I,(1+Asin(7rx/w)) [16]

where w is the hinge width and A is the amplitude of the interfering fiinges (Robeson
1995) . For spot photo bleaching, the spatial intensity of a Gaussian beam varies as

212J exp[- 2?) [17]

 

[03:11}? F

where P is the power of the laser beam and R is the half width of the laser beam at a depth
of He2 (Axelrod et al. 1976).

The diflirsion of proteins in the absence of convection is governed by Fick’s second law

of 31C
7037 “8]

where C is the concentration of the diffusing species, D is the diffusion coemcient, t is the

time of diffusion and r is the position vector of the species.

Due to the large aspect ratio of the fiinges, only diffusion in a direction perpendicular to
the fiinges contributes significantly to the fluorescence recovery, so equation [18] can be

rewritten in one dimension and in non-dimensional form as (Robeson 1995)

.52: 52C, [19]
or ax

 

where X=xlw and T=Dt/w2.

24

We can calculate w from (Axelrod 1990)

A: [201

w = 2n,sin9sin(p/2)

 

where B is the angle formed by the two beams in the plane of TIR Therefore, by directly
measuring 9 and B, the hinge width can be calculated. Another more accurate way to
calculate the fiinge width is to insert a calibrated reticule in the eyepiece of the microscope
and measure the fiinge width directly.

The concentration profile immediately after photobleaching is given by

C(X,O) = Ce'W) [21]

where C is the concentration before photobleaching, and k is the bleaching parameter
which contains the rate constant for photobleaching and the duration and intensity of the
photobleaching pulse (Axelrod et al. 1976).

The fluorescence recovery profile following photobleaching is given by

F0) = j1(r.€)C(r.9. aq, (CW [221

The task now is to determine the functional form of C(r,t) so that equation [22] can be
integrated. This problem has been solved by Robeson (1995) and Axelrod (1976) for

pattern and spot photobleaching, respectively.

The solution of the diffusion Dirichlet problem in the domain (-1 < X <1) and
(0 < t < co) for FRAPP is given by (Robeson 1995)

C(X, T) = 53°- + i exp(—n’7r2 T)[a, cos(mrX) +b, sin(mrX)] [23]

where a, and b, are Fourier coefficients given by

25

a. = jm—mmcosam [241

b, = Iexp(-kI(X))sin(mtX)dX [25]

Similarly, for spot photo bleaching, the concentration profile is given by (Axelrod et al.
1976)

 

_ - °° (-k)" R’ _ 2w”
C(r’t)‘C§ n! (R2+8nDt)exp[ 8nDt+R2] [26]

Equation [23] or [26] can be substituted, as needed, into equation [22] to obtain the
fluorescence recovery profiles, assuming constant quantum yield. After carrying out the

integration, Robeson (1995) obtained the following equation for the fluorescence recovery

profile after pattern photobleaching

F“) = (1;; - F)[l — fl(l - exp{—:-}):l + F [27]

Here, F; is the prebleach signal, F is the signal immediately after photobleaching, f is the
mobile fraction of molecules and a, and b1 are the Fourier coefficients given by equations

[24] and [25], and t is the characteristic difi’usion time for pattern photobleaching:

“’ [281

 

The same integration process, when carried out for spot photobleaching, gives the

following solution (Axelrod et. al 1976)

17(1): FE" { (I) 21)] [29]
"=°n. 1+ 1+—
7

where ‘t is the characteristic diffusion time for spot photobleaching with a Gaussian beam:

 

26

2
w
4D I ]

For small kvalues, equation [29] can be simplified to (Axelrod et. al 1976):

F(t)= ' 1- k [311
1+ —
r
The mobile fraction f is given by
/=——E"F(_°°’ [321

17,-}?

where F (co) is the asymptote of the fluorescence recovery curve.

Similarly if the amplitude of the interfering fiinges is close to unity, we can rewrite
equation [27] as

Fm—F “"9159“; [331

E, - F
and f can be calculated either by non linear regression of this equation or by (Tilton et. al
1990a)

/=3 ———F°’F(.°°) [341
143—1:

Experimental data were fit to equations [31], [32], [33] and [34] to obtain the difiirsion
coefficients and the mobile fraction of FITC-tagged BSA proteins at the oil-water

interface.

. 1% II" I

 

27

3.6 Calculation of the amplitude of the fringe pattern

Ideally, the spatial amplitude of the fiinge pattern should be unity. However, if the two
intersecting beams are not perfectly at critical angle, the amplitude will be less than unity.
We used the process described by (Robeson 1995) to calculate the amplitude of the
interfering fiinge pattern created by the intersecting beams.

We bleached the surface of the slide containing the immobilized fluorophores, and
observed the fluorescence recovery using the monitoring beam. When the monitoring
beam pattern was shifted so that it was completely in phase with the stationary bleached
pattern, we get high fluorescence (F 180), and when we moved the monitoring beam
completely out of phase with the bleached pattern we obtained a low fluorescence (Fa).

We used the piezo-electric controller to oscillate the monitoring beam fiinges.

To calculate A, the amplitude of the fiinge pattern, an iterative process was used, based on

the following algorithm:
1. We assumed a value forA.

2. Then, by integrating equation [22] for F0 and F130 as firnctions of k, we generated a
plot of E, /F and (E80 —E,)/F versus k, using the Mathcad program. A typical

resulting plot is shown in shown in Figure 6.

3. When the experimentally measured values of bothFol F and (15;,0 —Fo)/F at the
same I: value are correctly predicted by the graph, we conclude that the assumed A
value is correct.

4. Ifnot, we chose a different value of A and repeated the process.

From this exercise, the value of the amplitude was found to be unity. As a further check,
difi‘erent intensities of the photobleaching beam were used (that is different k values) to
see if the A value chosen was satisfactory for all It. A was found to be equal to unity in all

cases.

 

it}: c. .. v. ._,_

28

 

—O—(F180(k) - F0 (k))lF
—Fo (kw

 

 

 

 

 

 

Figure 6. Graph used to calculate the amplitude of the fringe pattern.

 

 

 

4. RESULTS AND DISCUSSION

4.1 Check of the TIRF setup using steady state adsorption of BSA-FITC

The first exercise was to access the feasibility of using TIRF to characterize
macromolecular interactions at the oil-water interface. To do this, we collected
fluorescence intensity versus time data during the adsorption of BSA at the oil-water
interface (Figure 7). Steady state adsorption was interpreted as a constant fluorescence
intensity versus time. This initial experiment provides three important pieces of

information:

1. It confirmed that the TIRF apparatus is firnctioning properly.

2. It helped establish the time required for adsorption to reach steady state after the
pumping of the protein solution has been stopped. From the data, we estimated the

time required for steady state to be approximately 15 minutes.

3. The profile in Figure 7 assures us that the intensity of the monitoring beam does not
induce slow photobleaching of the adsorbed layer of protein molecules. If the
monitoring beam intensity were high enough to cause photobleaching, a slow but
steady decrease in the fluorescence signal would result. By adjusting the angles of the
optical flats, we found that a monitoring beam intensity of ~15 uW is the optimum
intensity to prevent slow photobleaching of the protein molecules, and yet produce
fluorescence emissions that can be easily detected over the background noise level. For
FRAPP experiments, the optimum monitoring beam intensity was 34 uW. The ratio of

the monitoring beam to the photobleaching beam intensity was 1:5000.

29

 

O-‘

7-.....

3O

 

£33.»... . .

f cccccccc occurrrflrod

0 O I. .~\.. .
T-.---"e‘ I. .
4 s

. it. b p.
o..&§K.\$w
l'-..‘q .

.. l50ll
l'flidrl
0 ‘.br.40.l.91¢-00

 

                 

Vba.
Q

r. 4%.“ 0".

I.
..5 .. . '
. .‘D
a 4 I-

r
......

I \ ~r.«..1 .0

. vow:

rat-{o3 'I
O.'

       

.P....-~L. I

“1.110:
.. ..

.......

...........
.......

r
rrrrrr

.......

. 11‘:
21.333:

fiver a...

r.
v ..

t . .l-IP ‘IIISI
. 9..

1000

 

 

time(s)

Figure 7. Fluorescence emission profile during adsorption of BSA at the oil-water

interface. From this data, we estimated the time required for steady state adsorption to be

ICS.

rninu

15

31

After establishing that our setup was firnctioning proper of our setup, we ran an
equilibrium adsorption experiment as explained in Section 3.3.4. We first introduced borax
buffer into the cell. At time t = 0, we started injecting a labeled BSA solution into the cell.
The fluorescence intensity rose steadily from a value of II, as shown in Figure 8, to a
steady state value of 13. We stopped the pumping and waited for a period of 10 minutes.
We then replaced the labeled protein with an unlabeled protein solution of the same

concentration. This resulted in a steady decrease in fluorescence emission until the signal

reached a new steady state value of 12.

The large difference in fluorescence intensity between 13 and 12 after the labeled protein
solution in the bulk was replaced by unlabeled protein solution can be interpreted in one of
two ways. It is possible that, contrary to what has been reported in the technical
literature, protein adsorption at the oil-water interface is reversible, with the result that
surface-adsorbed labeled proteins are immediately replaced by bulk unlabeled proteins,
thus reducing the fluorescence emission intensity. The second possible explanation for this
observation is that the majority of the fluorescence emission comes from the bulk solution,
so that afier the bulk was depleted of labeled proteins, there was a large drop in the
fluorescence emission. To difl‘erentiate between these two possibilities, we will need to
devise a calibration technique which can be used to calculate the concentration of proteins

adsorbed at the interface solely in terms of the fluorescence emission intensity, and as a

firnction of the concentration of the bulk solution. Such a calibration technique was

described in chapter 3, although it was not implemented in this study due to time and

equipment constraints at the time of this work.

 

 

 

32

 

rooo -

900-

800'

 

700‘

600‘

400"E

FIuorescence(A U)

 

300-E

 

 

200 .i.‘ ’.

 

 

100

 

 

Figure 8. Fluorescence emission profile of BSA: at the initiation of the profile, unlabelled
proteins in the sample cell were replaced by labeled proteins; after approximately 2500
seconds, the bulk labeled proteins were replaced by unlabeled proteins.

 

33

4.2 Results of spot photobleaching experiments

A typical profile of fluorescence recovery afier photobleaching is shown in Figure 9. The
original data represented by the filled circles, were noisy over the entire range of
recovery. To get a sense of the trend in the data, we averaged each set of eight
consecutive data points sequentially (1-8, then 2-9, then 3-10, etc.) to obtain the points
represented by the open squares. This process was done to make it easier to locate the
prebleach and postbleach values with greater accuracy, since these are required in
subsequent calculations. It must be pointed out that the regression line shown in Figure 9
is the same for both sets of data; the sequential averaging was done only to improve

clarity.

The characteristic time, t, was obtained by non-linear regression of the fluorescence
recovery data using equation [30]. A typical regression analysis result is plotted in Figure
10 for the data shown in Figure 9. Results for the other two replicate experiments are
shown in Figure 11 and Figure 12. The mobile fi'actions were calculated directly from

equation [32]. The results of the three experiments are summarized in Table 1.

The average diffusion coefl'rcient fiom the three experiments is 52 :I: 1x 10"cm2/ s, which

is of the same order of magnitude (i.e., ~10“ cmzls) as the value reported by Zimmermann

et. al (1990). The average mobile fraction was f = 0.77 :l: 0.05 .

Table 1. Difi‘usion coemcients and mobile fractions obtained fi'om three replicate spot
photobleaching experiments.

 

 

 

 

Ir w (pm) 1 (sec) D (c?/s)x10‘ f
0.176 31.25 56.7 i 0.7 4.3 0.79
0.282 31.25 50.1: 0.1 4.9 0.71
0.323 31.25 39.1: 0.2 6.2 0.81

 

 

 

 

 

 

 

 

 

 

34

      

0 20 40 so 80 100 120 140 160 180
thus)

Figure 9. Actual (Series 1), averaged (Series 2) and fitted (Series 3) fluorescence emission
versus time after photobleaching.

 

‘—Fluorieaeenee (regression)

................ 8 PH. Mov, Avg.
(Fluorescence (m0)

 

 

 

 

 

Figure 10. Fluorescence recovery afier photobleaching for a spot size of 3 1.25 pm

35

 

 

 

 

""""thxaeoence «cannnku»

................ lipanlflovuAMl.
(thxIscence(achnin

 

 

 

1400+ - 1 ' I
“‘0 I .' II i‘ ="_‘. gird-.111
:,;. .1 .. .. . ,.
§1md Flu F: Tut-Win; .
.1 $1119 *1 ' ll
’ - E; :‘U i *1 I 1 ~
1310» f > 'I
if: 1 IN 1 I
1260 “3' I I
1’
IL 1210 ‘
1100
1110 1 1 1 1
0 50 100 150

thnlms)

Figure 11. Fluorescence recovery after photobleaching for a spot size of 31.25 um

 

 

 

1460- ! I I it; III
11! s'r'I! I
14104 L.-_ I J, i“i:' .QQLL 9!
g 130I :Ii IF 7;»; {:1 I ~ '1:
I . ”$in _' ' I‘I , 1‘ it]
1310 .. 1,. 233-1: nI ' ‘
:fsJI : E. ii l- I
II. 1‘" ' l
1’} I I
n. 1210 '
1160
1110 . +7 1 .
O 50 100 150

fluids)

 

“"""'thxeeeence enuresskxu

................ 8 per. Mov. AVG.
(thxeacence(hchl'»

 

 

 

Figure 12. Fluorescence recovery after photobleaching for a spot size of 3 l .25 um.

36

An average mobile fraction of ~0.4 has generally been reported in the technical literature
for BSA adsorption at the glass-water interface (Robeson 1995, Tilton et al. 1990). A
low mobile fraction indicates a high degree of irreversibility in protein adsorption. A high
mobile fraction could also indicate that most of the recovery is due to exchange of the
bleached near surface protein molecules with proteins from outside the field of the
evanescent wave. We attempted to study this process firrther by coating the top slide with
PEO and observing protein difl‘usion near this non-adsorbing interface to obtain an
estimate for the near surface bulk difi‘usion. However, experimental results were
inconclusive. So, at this stage, we are unable to distinguish between near surface difl'usion
and difl‘usion to the interface fi'om the bulk. This is why we refer to the bulk difi‘usion
coemcient as an apparent difl‘usion coefiicient, since is most likely a combination of bulk

and some surface diffusion.

4.3 Results of pattern photobleaching experiments

The results of three replicate FRAPP experiments are plotted in Figure 13, Figure 14, and
Figure 15. In these figures, we have plotted the normalized rather than actual
fluorescence data. The original fluorescence emission data were converted to normalized

fluorescence emissions by the equation

rim-i:
133—1?

Normalized Fluorescence = [3 5]
where F (t) is the fluorescence emission at any time, F is the post-bleach emission

intensity, and F; is the prebleach intensity.

The characteristic time for diflirsion, ‘t, and the mobile fraction, 1: were obtained by non-
linear regression of the normalized fluorescence recovery data, using equation [33].
Equation [3 4] was used to further check the value of f obtained fiom the non-linear

regression. It was found that the two methods gave the same result. The results are
summarized in Table 2.

37

 

 

 

 

 

2
1.8
1.0
14 ————— Normallzedexperlmentel
1 Fluorescence
'2 i I 1 'I 1 —Normelizedlluoreecence
‘ ‘ I“ . ' 1I It 1 (regression)
1; "H 'I‘ ll I ‘1 .
0.8 ' ‘ I ' ‘ l 1 I III" I11!
2 0.0 I . ' ' ' I I i" III
0.4 1
i 42. 1 r 1 a
-7 3 13 23 33
tinels)

Figure 13. Normalized fluorescence recovery afier pattern photobleaching with a fiinge
width of 6.25 um.

 

 

 

 

 

 

2 -.
1.8 --
1.6 «-
————— Nonmlhedexpemrental
1'4 ' Fluorescence
1.2 -~ I I I ‘ “—Nonnallzedtluoreecence
1 I I 1 (newton)
._v I I I II I 1" I 1. l].
0.81 I II | 1 ‘I I I 1 1 1 1 .1]
0.6 I I I l , II
0.4 4
0.2 r a r 9
0 10 20 30 40

time (s)

Figure 14. Normalized fluorescence recovery after pattern photobleaching with a fiinge
width of 5.68 pm.

38

 

 

— Normalized experimental
1-2 ‘ Fluorescence

" 1 I I I —Normellzed1horeecenoe
g .1, I. I IHIIII' . I II I (mansion)

 

 

 

 

 

0 10 20 30 40
time (s)

Figure 15. Normalized fluorescence recovery after pattern photobleaching with a fringe
width of 5.21 urn

 

 

 

 

 

 

14--

12.. “SEW-11.727

R2-0.9921

10 -»
.. a .. ‘13st 0 Serlee1
a R soeooe _.m(s.m1)
‘5 3.. —Llnet(8erlee1)

4 -I

2 ..

0 4— l 4. 1

0 0.0000001 0.0000002 0.0000003 0.0000004
w’ (cm‘)

Figure 16. Characteristic time of recovery versus the square of the fringe half period The
thick line is for regression of the data through the origin, as required on physical grounds;
the thin line is a best fit regression of the data.

39

Table 2. Characteristic times and mobile fractions from pattern photobleaching
experiments

 

 

 

 

 

w (M) I (see) 1'
6.25 13.5 0.48
5.68 8.6 0.53
5.21 5.9 0.45

 

 

 

 

A plot of r “the characteristic time,” vs. w2 “the square of the characteristic length,” is a
straight line (Figure 16), which implies that the fluorescence recovery is purely by
diffusion. Ifthe fluorescence recovery were by flow, then the characteristic time would be
proportional to w instead of w’. Since all the experiments were run in a stopped flow
format, this result is to be expected. The best fit regression of the data gave a regression
coeficient of 0.99, but does not pass through the origin. However, on physical grounds,
the plot of 1: versus w’ should pass through the origin. Forcing the regression line through
the origin gives a low regression coeflicient of 0.7, but satisfies the physics of the

situation.

The slope of this plot gives a diffusion coefficient of D = 3.38 x 10" cmzls, and an average
mobile fraction of f = 0.48 i005 , both of which are slightly higher than the value

reported by Tilton et. a1 (1990) (D = 1.1 x 10'9 cmZ/s and f = 0.42) for EITC-BSA (Eosin
isothiocyanate labeled BSA) adsorbing from a solution of SOug/ml to the glass-water
interface. Since EITC is a structural derivative of FITC, we do not expect it to afl‘ect the
adsorption any differently than F ITC, and certainly not to the extent of accounting for the

observed difl’erences. Therefore, it is necessary to look elsewhere for explanations.

The surface diffusion coemcient is a function of the surface concentration. Since we use a
higher concentration of BSA than Tilton et. a1 (1990), we expect a lower diffusion

coefficient for our experiment, since the interface is more “crowded.” However, the

difi
the

difl

3C5

4.4

ll?!

7..

In.

cm

is:

del

ms

aPl

Prc

nee
flm

bei

Wa

4o

diffusion coefficient is also afl‘ected by hydrophobic interactions between the protein and
the interface which, at this point, we do not fully understand. We suspect that the
difference in hydrophobicities between our interface and that of Tilton et al. (1990) may

account for the difl‘erence. However, this is currently under further study.

4.4 Discussion

In general, fluorescence recovery after photobleaching is possible due to four modes of
transport:
1. exchange of unbleached bulk molecules with bleached molecules in the evanescent

wave;

2. essentially one dimensional bulk diffusion-controlled exchange of bleached molecules
with unbleached molecules near the interface, yet outside the field of the

photobleaching pulse;
3. exchange of surface bleached molecules with unbleached molecules in the bulk; and

4. lateral diffusion of surface bound molecules.

In general, protein molecules are mobile in the bulk with a difi‘usion coefficient of ~10‘8
cm’ls (Zimmermann et al. 1990). Therefore, the time for recovery from the first process
is approximately given by equation [5]. In this study, I, the characteristic length is the
depth of the evanescent wave (i.e. I= 138 nm), which implies a difiirsion time of about 20
ms. As the time scale of our experiments is longer than 20ms, process 1 appears as an

apparently unbleachable fraction in our experiments.

Process 2 is roughly of the same duration as the bleaching pulse. Consider a small area
near the interface. The bleaching beam photobleaches this region during the time ‘t’ it is
flashed on the interface. The protein molecules in this region are however constantly
being replaced by unbleached molecules fi'om outside the field of view of the evanescent

wave by approximately one dimensional diflirsion. The region bleached by the beam is

41

about (Dom. Depending on the time of photobleaching, it will take the unbleached
molecules about the same time to replace the bleached molecules near the interface.

By using a bleaching beam diameter three times larger than the monitoring beam diameter,
we can eliminate lateral diffusion as a probable cause of recovery because the approximate
time of this diffusion, as calculated fi'om equation [5] is on the order of 900 sec. Hence,
the characteristic time measured by spot photobleaching is due to replacement of the
surface bleached molecules by bulk molecules, and the replacement of near surface

bleached protein molecules with molecules not in the field of view of the evanescent wave.

In pattern photobleaching, the bulk is free of labeled proteins, so the only recovery we
obtain is due to lateral diffusion of labeled BSA along the interface. The value of the
diflirsion coeficient in spot photobleaching is an order of magnitude larger than the one
obtained in pattern photobleaching, which tells us that we are indeed observing bulk
difl‘usion in the spot photobleaching experiments.

In principle, rotational difiirsion of BSA as it unfolds at the interface could also cause
fluorescence recovery. However, since we have no indication that this efl‘ect plays a

significant role in the time scale of our experiments, we did not attempt to quantify it here.

4.5 Possible sources of errors

The calculation of diflirsion coefficients using the equation described by the model
assumes that we can accurately locate the pre-bleach fluorescence intensity, the
fluorescence intensity immediately after photobleaching, and the asymptote of the
fluorescence recovery curve. The characteristic diffusion time, t , is calculated based on a
theoretical fit to the data, using the prebleach and post bleach beam intensities as end
points. Due to the high noise level of our data over the whole range of recovery, we were

not always able to obtain these points precisely.

42

In pattern photobleaching, the observable recovery is very small, as compared to spot
photobleaching. So the error in determining the exact location of the prebleach and post
bleach beam intensities is much higher. This error explains why the plot of the
characteristic time versus the square of difi‘usion length in pattern photobleaching does not
pass through origin. We suggest using a more sensitive photomultiplier tube, and reducing
the gain and the high voltage settings of the PMT in future experiments.

The diffusion coefiicient is proportional to the square of the beam size. As a result,
discrepancy in measuring the spot size could create large errors in the difl‘usion
coefiicient. We assumed a perfectly Gaussian laser beam profile to calculate the beam
width. However, imperfect optics and imperfect focusing of the beam can cause this
assumption to break down as these effects make our beam profile non-Gaussian
(Schneider and Webb 1981). Also, the mobile fi'action measured in all spot
photobleaching experiments was essentially the same. Since the mobile fraction does not
depend on the beam size, we conclude that any errors in the calculation of the difiirsion
coeflicient are due to errors in measuring the beam width. To overcome this problem, we
suggest that the more elaborate procedure described by Schneider and Webb (1981) be

used in subsequent studies.

 

5. SUMMARY AND CONCLUSIONS

The work done in this thesis has shown conclusively that total internal reflection
fluorescence can be used to study macromolecular phenomena at the liquid-liquid interface
and that our setup is functional. We have incorporated the capability to use both spot
photobleaching (FPR) and pattern photobleaching (FRAPP) to determine both the bulk
and lateral diflirsion coefiicients at the oil-water interface. We have estimated the time for
steady state protein adsorption at the oil-water interface to be approximately 15 minutes
fiom the initiation of the flow of proteins into the sample cell, based on a sample cell of
approximately lOOOum in depth. The optimum monitoring beam intensity was found to be
15 uW, making the ratio of the monitoring to photobleaching beam intensity about
125000. The time required for photobleaching varied from 200 to 800 ms.

From the experiments conducted using TIRF/FRAPP, we have estimated that FITC-B SA
has a lateral difl‘usion coefficient of 3 .4 x 10" cmzls at the oil-water interface, with an
average mobile fi'action of 0.48. The apparent bulk difiirsion coeflicient obtained from
TIRF/FPR experiments was 5.1 x 10" cmz/s, with a mobile fi'action of 0.77. The
difference of one order of magnitude in the two results appears to indicate that the bulk
diffusion of FIT C-BSA is faster than lateral diffusion, as would be expected. We were
unable to distinguish between diffusion of proteins in the bulk and the exchange of protein
molecules between the bulk and the interface. This is because of the overlapping time
scales of the two difi‘usion processes. Methods to differentiate between the two modes of
recovery are currently under further study.

Errors in measuring the spot size, as well as signal noise strongly afi‘ected the values of the

difiirsion coefficients. We have analyzed how these errors might effect our readings and

43

44

might be used in future studies to calculate the surface concentration of proteins on the
basis of fluorescence emission data. We have been unable to implement and verify the
protocol in this study, because we have not yet developed a method for measuring surface
concentrations independent of TIRF emissions.

6. SUGGESTIONS FOR FUTURE WORK

This thesis is an attempt to quantify protein adsorption at the liquid-liquid interface. We
hope to provide a first approximation to far more complex studies dealing with ligand-
receptor interactions, drug delivery and food processing problems. BSA was chosen
because its characteristics are well documented in the literature. Also, various diffusion
coeficients are reported for BSA adsorption at the solid-liquid interface which give an
approximate idea of the value that should be Obtained from our experiments. After
reduction in the noise level of the signal and the protocol for estimating the accurate beam
size has been established, we suggest the following experiments be conducted in the
firture.

6.1 Diffusion of surface-bound species

A beam expander should be used to increase the incident beam size, to enable a wider
range of beam widths to work with. The beam can then be passed through pinholes Of
varying radii to change the spot size. Since the depth Of the evanescent wave would
remain constant in these experiments, the slope Of the characteristic difi'usion time versus

the square of difi‘usion length will give the surface difi‘usion coefficient.

6.2 Diffusion of bulk species

We could obtain D” by varying the depth Of the evanescent wave while keeping the spot

size constant. This is a very elegant technique because one could plot the recovery time

versus the square of the characteristic length (d in this case) and obtain the bulk diffusion

45

46

run because, to change the depth of penetration Of the evanescent wave, we need to either
vary the wavelength or the incident angle of the laser beam. Changing the wavelength
requires a tunable laser. We would also need fluorescence tags that will fluoresce at
difl‘erent wavelengths and different filters for our photomultiplier tube. Even if this was
possible, various fluorescent tags will affect the diffusion characteristics of BSA in
different ways. Also, varying angles of incidence will cause different spot sizes to form at
the Oil-water interface, making it impossible to maintain a given spot size while
simultaneously varying the depth of penetration of the evanescent wave (Axelrod 1990).

Instead of the above procedure, it might be possible to vary the depth of the protein layer
in the sample cell. This will change the depth Of view of our evanescent wave while
maintaining the spot size. We suggest three ways to do this:

1. The thickness Of the PEO coating on the bottom slide could be varied. The thickness
of the slide can then be measured by reflectometry to Obtain the efl‘ective depth of the
experimental cell.

2. We could use aluminum spacers Of varying thickness (this will require O-rings Of
varying thickness) but constant area. Then, by measuring the thickness of the spacer,
we could calculate the depth Of view Of the evanescent wave.

3. We could also insert quartz wafers of varying thickness into the sample cell to change
the depth of the cell.

With each of the above modifications, the slope of a plot of recovery time with

characteristic length will give DI...“L

6.3 Accounting for concentration quenching

We are aware that, due to concentration quenching, the relative quantum yield of
fluorophores may not be equal to unity as we have assumed in our analysis. Concentration
quenching is a manifestation of Forster non-radiative energy transfer. In standard FPR and

FRAPP analysis, we assumed the quantum yield of the fluorophores to be constant. We

47

know that, for our system, concentration quenching could occur at high labeling ratios as
Shown by Robeson (1995) for FITC-Rnase adsorbing to glass Slides coated with polymer
substrates. For this reason, labeling ratios were kept sufficiently lower than the
concentration at which quenching of the signal might occur.

To account for a varying quantum yield, q(r) must be brought inside the integral sign of
equation [22]. Then, if fluorescence can be related to the surface concentration, a
firnctional relationship between the relative quantum yield and the surface concentration
can be Obtained as explained by Robeson (1995). For this purpose, it is necessary to
devise a calibration technique that can differentiate between bulk and interfacial
fluorescence emissions, even at high labeling. This could be done by treating K, the
instrument constant, as a variable. We could then try fitting experimental data to equation
[10] to Obtain K as a function Of the concentration.

10.

ll.

12.

7. BIBLIOGRAPHY

Abney, JR, Scalettar, BA, and Thompson N.L. Evanescent interference patterns for
fluorescence microscopy. Biophys. J. 61:542-552 (1992).

Arai, K., Kusu, F ., Takamura, K. Electrochemical behavior of drugs at the oil/water
interfaces. h: Liquid-Liquid Interfaces: Ilreory and Methods 37 5-400 (1996).

Axelrod, D., Koppel, D.E., Schlessinger, J., Elson, E., Webb, W.W. Mobility
measurements by analysis of fluorescence photobleaching recovery kinetics.
Biophysical Journal 16: 1055-1069 (1976).

Axelrod, D. Cell-substrate contacts illuminated by total internal reflection
fluorescence. J. Cell Biology 89: 141-145 (1981).

Axelrod, D., Thompson N.L and Burghart, T.P. Total internal reflection fluorescence
microscopy. J. [Microscopy 129 (pt 1) 19-28 (1983)

Axelrod, D., Burghart, T.P., and Thompson, N.L. Total internal reflection
fluorescence. Annu. Rev. Biophy. Bioeng. 13: 247-268 (1984).

Axelrod, D. Total internal reflection at biological surfaces. I_n: Noninvasive
Techniques in Cell Biology, pages 93-127 (1990)

Axelrod, D. Hellen, EH. and Fulbright RM. Q: Topics in Fluorescence
Spectroscopy, Vol. 3 LBiomedical Applications. Lakowicz, JR (Editor), 289-243
(1992).

Barisas, GB, and Leuther, M.D. Fluorescence photobleaching recovery measurement
of protein absolute diffusion constants. Biophysical Chemistry 10: 221-229 (1976).

Burghardt, T.P. and Axelrod D. Total internal reflection/fluorescence photobleaching
recovery study Of serum albumin adsorption dynamics. Biophys. J. 33:455-468
(1981).

Cheng, Y.L., Darst, S. A., Robertson, C.R. Bovine Serum Albumin adsorption and
desorption rates on solid surfaces with varying surface properties. Journal of Colloid
and Interface Sci. 118: 212-223 (1987).

Elson, E.L. Fluorescence correlation spectroscopy and photobleaching recovery.

48

49

13. Graham, D.E., Chatergoon, L. and Phillips, M.C. A technique for measuring
interfacial concentrations at the oil-water interface. Journal of Physics E. :Scientific
Instruments 8: 696-699 (1975).

14. Hansen, RL., Harris, J.L., Lateral difi‘usion of molecules partitioned into C-18 ligands
on silica surfaces Analytical Chemistry 67: 492-499 (1995).

15. Hirschfield, T. Total Reflection Fluorescence (TRF). Can Spect. 10: 128.
16. Lakowicz, 1R, Principles of Fluorescence Spectroscopy :Plenum Press (1986)

17. Lok, B.K., Cheng, Y.-L., and Robertson, CR Total internal reflection fluorescence a
technique for examining interactions of macromolecules with solid surfaces. J. Colloid
Interface Sci. 91: 87-103 (1983a)

18. Lok, B.K., Cheng, Y.-L., and Robertson, C.R. Protein adsorption on crosslinked
polydimethyl siloxane using total internal reflection fluorescence. J. Colloid Interface
Sci. 91: 104-116 (1983b)

19. McKinney, RM., Spillane, J.T., Pearce, G.W. A simple method for determining the
labeling eficiency of fluorescein isothiocyanate products. Analytical Biochemistry 14:
421-428 (1965)

20. Morrison, LE. and Weber, G. Biological membrane modeling with a liquid-liquid
interface. Biophys. J. 52:367-379 (1987)

21. Nakanaga, T. and Takenaka, T. Resonance Raman spectra of monolayers of a
surface-active dye adsorbed at the Oil-water interface J. Phys. Chem. 81: 645-649

(1977)

22. Robeson, 1L. Application of HRF to Probe Surface Difiirsion and Orientation of
Adsorbed Proteins . Ph.D. Thesis, Carnegie Mellon University (1995)

23. Robeson, J.L. and Tilton, R.D. Efi‘ect of concentration quenching on fluorescence
recovery after photobleaching measurements. Biophysical J. 68: 2145-2155 (1995)

24. Schneider, M.B. Measurement of submicron laser beam radii. Applied optics
20:1382-1388 (Year).

25. Strauss, W.A Partial Drflerential Equations- an introduction : John Wiley & Sons,
Inc. (1992)

26. Takenaka, T. Efi‘ect of electrolyte on the molecular orientation in monolayers
adsorbed at the liquid-liquid interface: studies by resonance Raman spectra. Chem.
Phys. Lett. 55:515-518 (1978)

2'7.

28.

29.

30.

31.

32.

33.

50

Takenaka, T. and Nakanaga, T. Resonance Raman spectra of monolayers adsorbed at
the interface between carbon tetrachloride and an aqueous solution of a surfactant and
a dye. J. Phys. Chem 80: 475-480 (1976).

Thomas, J., and Webb, W. W. Fluorescence photobleaching recovery: A probe of
dembrane dynamics. In: Noninvasive Techniques in Cell Biology, 129-152 (1990)

Thompson, N.L. and Axelrod, D. Immunoglobin surface binding kinetics studied by
total internal reflection with fluorescence correlation spectroscopy. Biophys. J. 43:
103-114 (1983)

Thompson, N.L., Burghart, T.P. and Axelrod, D. Measuring surface dynamics of
biomolecules by total internal reflection fluorescence with photobleaching recovery or
correlation spectroscopy. Biophys. J. 33: 435-454 (1981).

Tilton, RD, Robertson, CR, and Gast, AP. Lateral difiirsion Of bovine serum
albumin adsorbed at the solid-liquid interface. J. Colloid Interface Sci. 137: 192-203
(1990a).

Tilton, RD, Robertson, CR, and Gast, AP. Surface diflirsion of interacting
proteins: effects of concentration on lateral mobility of adsorbed bovine serum
albumin. Biophys. J. 58: 1321-1326 (1990b).

Tweet, A. G., Gaines, G.L., and Bellamy, W.D. Fluorescence of chlorophyll "'
monolayers. J. Chem. Phys. 40:2596-2600 (1964).

34. Volkov, AG., Deamer, D.W. Oil/water interfaces and the origin of life. In: Liquid-

35.

Liquid interfaces: Iheory and Methods 363-375 (1996).

Zimmermann, R.M., Schmidt, C. F., and Gaub, H. E. Absolute quantities and
equilibrium kinetics of macromolecular adsorption by fluorescence photobleaching in
total internal reflection Journal of Colloid and Interface Science 139: 1, 268-280
(1990).

8. APPENDIX: MATHCAD PROGRAM

Mathcad program used to develop the fluorescence profile for the immobilized
fluorophore system
Method to calculate the amplitude of the fluorescence fringe pattern

AssumeA
A:=1

a := 3.141592654

k1: 0,0.02..1

unitless profile for the photobleaching beam and the monitoring
beam when it is not in phase with the bleached pattern

I(X) := l + A-sidl-X)

unitless profile of the monitoring beam shified by a phase

Iu(X) 2:: 1 + A-sir(1r-X + n)

Fluorescence observed from the immobilized fluorophore system when the monitoring beam
has not been shified so that it is alligned to the bleached region divided by fluorescence
before photobleaching

1

exp(-k-I(X))°I(X) dX

F0(k) := '1

 

2

Fluorescence observed from the immobilized fluorophore system when the monitoring beam
has been shifted so that it is alligned to the bleached region divided by fluorescence before

photobleaching
l

exp(-k-I(X))-IN(X) dX
-1

 

Fl80(k) ;=

y(k) := F180(k) - F0(k)

51

"‘IIIIIIIIIIIIII“