University of Southern California Dissertations and Theses

The interaction of photo-responsive surfactants with biological macromolecules

(USC Thesis Other)

The interaction of photo-responsive surfactants with biological macromolecules

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
THE INTERACTION OF PHOTO-RESPONSIVE SURFACTANTS WITH
BIOLOGICAL MACROMOLECULES

by

Khiza L. Mazwi

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(MATERIALS SCIENCE)

December 2012

Copyright 2012 Khiza L. Mazwi
ii

Acknowledgements
I would like to express by deepest gratitude to my advisor, Dr. Ted Lee Jr.,
who was always a source of expert advice, technical guidance and encouragement.
His patience, thoughtful insights and pleasant nature created the ideal research
atmosphere.
I would like to thank the members of my guidance committee, Dr. Ralf
Langen, for his helpful suggestions and support, and Dr. Ed Goo, who has become
something of a mentor to me during my time at the Viterbi School.
I would also like to thank Dr. William Heller and Dr. Sai Pingali Venkatesh
for their technical assistance collecting neutron scattering data, as well as my
undergraduate research assistants and fellow research group members.
Finally, I recognize that I am indebted to all the other excellent teachers and
educators who have passed through my life and instilled in me a sense of wonderment
at the nature of materials.

iii

Table of Contents
Acknowledgements ii
List of Figures iv
Abstract xiii

Chapter 1: Introduction 1
1.1 Protein Structure 1
1.2 Photo-responsive Surfactants 8
1.2.1 Surfactant-Protein Interactions 10
1.3 Protein-Protein Interactions and Amyloidosis 13
1.4 Proteins 16
1.4.1 Insulin 16
1.4.2 Catalase 17
1.4.3 Papain 19
1.5 Experimental Techniques and Methods 21
1.5.1 Dynamic Light Scattering 21
1.5.2 Small Angle Neutron Scattering 22
1.5.3 Atomic Force Microscopy 24
1.5.4 Circular Dichroism 25

Chapter 2: The Effect of Photo-responsive Surfactants on Insulin
Fibrillation 27
2.1 Abstract 27
2.2 Introduction 28
2.3 Experimental Techniques and Methods 32
2.3.1 Materials 32
2.3.2 Dynamic Light Scattering 34
2.3.3 Atomic Force Microscopy 35
2.3.4 Small Angle Neutron Scattering 35
2.3.5 Circular Dichroism 38
2.4 Results and Discussion 38
2.5 Conclusion 62

iv

Chapter 3: Photocontrol of Catalase Quatenary
Structure 64
3.1 Abstract 64
3.2 Introduction 65
3.3 Experimental Techniques and Methods 67
3.3.1 Materials 67
3.3.2 Circular Dichroism 68
3.3.3 Small Angle Neutron Scattering 69
3.3.4 Dynamic Light Scattering 71
3.3.5 Flourescent Spectroscopy 72
3.3.6 Activity Measurement 72
3.4 Results and Discussion 73
3.5 Conclusion 87

Chapter 4: Structural Changes Papain induced by Photo-responsive
Surfactants 88
4.1 Abstract 88
4.2 Introduction 89
4.3 Experimental Techniques and Methods 92
4.3.1 Small Angle Neutron Scattering 93
4.3.2 Circular Dichroism 95
4.4 Results and Discussion 96
4.5 Conclusion 103
4.5.1 Future work 104

Chapter 5: Future Studies 105
5.1 Introduction 105
5.2 β-Amyloid Peptide Fragment (1-42) 107
5.3 Protein Dynamics and Activity 111
5.3.1 Neutron Spin Echo Spectroscopy 111
5.3.2 Phosphoglycerate Kinase 112

References 115
Bibliography 127

v

List of Figures
Figure 1.1. Generalized structure of the amino acid. The nature of the R group
determines classification. 1

Figure 1.2. The four levels of protein structure

(adapted from RSCB PDBs) 6

Figure 1.3. Structure of the two isomers of the azoTAB surfactants. 9

Figure 1.4. UV/vis absorption profile for the trans and cis azoTAB isomers (1 mM, 1
mm pathlength) 9

Figure 1.5 Theorized mechanism of fibrilogenesis following the pathway:
Monomers → Oligomers → Protofibrils → Protofilaments → Fibrils 14

Figure 1.6 Structure of the native insulin hexamer (a), and the monomer (b),
showing the A chain (blue) and B chain (green) [PDB 4INS and 1AIY] 16

Figure 1.7 Structure of the native catalase tetramer from bovine liver. [7CAT] 18

vi

Figure 1.8. Structure of papain protein. Domain 1 is primarily helical, whereas
domain 2 is characterized by a strong β-structured content. [PDB 1PPP] 19

Figure 2.1. Structure and isomerization of the azoTAB surfactants 32

Figure 2.2. Time dependent DLS data (a), and UV/vis absorbance (b), for insulin,
and insulin in the presence of azoTAB surfactants DLS 39

Figure 2.3. Time dependent DLS data for pure insulin (10 mg/ml) incubated at 60
o
C
and pH 1.6. Inset diagrams illustrate potential oligomeric intermediates along with
their hydrodynamic radius calculated by the modified Stokes-Einstein equation for
the infinite dilution diffusion coefficient of a cylindrical species. 40

Figure 2.4. Atomic Force Microscope (AFM) images of 10mg/ml insulin incubated
at 60
o
C at pH 1.6 for (a) t = 0 minutes, monomer (b) t = 12 minutes, early
oligomers/hexamers visible (c) t = 35 minutes, evidence of lateral aggregation of
hexamers (d) t = 65 minutes, multiple oligomeric species present (e) t = 120 minutes,
Lateral aggregation of fibrils (f) t = 500 minutes, network of large fibrils. 43

vii

Figure 2.5. Height profiles of prefibrillar and fibrillar species in the amyloidosis
pathway for 10 mg/ml pure insulin after incubation at 60
o
C in pH 1.6 solution for (a)
12 min (rod-like oligomer height ~ 1.5 nm), (b) 35 min (oligomer height ~ 8 nm), and
(c) 500 min (fibril height ~ 8 - 10 nm) 44

Figure 2.6. AFM images of insulin + azoTAB surfactants (a) trans isomer, t = 30
minutes, (b) trans isomer, t = 60 minutes. (c) cis isomer, 12 minutes, (d) cis isomer,
65 minutes In the presence of the trans form of the surfactant, short rod-like
intermediates are observed although they show a tendency to lateral aggregation.. In
the presence of the cis form, fibrils are replaced by amorphous/globular
aggregates. 46

Figure 2.7. Circular Dichroism spectra for (a) Pure insulin at 1h, 4h, 7h, 9h, 12h and
22h, (b) Insulin + 0.5mM azoTAB cis isomer at 1h, 4h, 7h, 9h, 12h and 22h, (c) Pure
insulin, Insulin + 0.5mM azoTAB cis isomer, and Insulin + 0.5mM azoTAB trans
isomer at t = 1h, (d) Pure insulin at t = 22h, Insulin + 0.5mM azoTAB cis isomer at t
= 22h, and Insulin + 0.5mM azoTAB trans isomer at t = 4h (after extensive
aggregation). 48

Figure 2.8. SANS scattering curves for 1 mg/ml insulin, incubated 1.5h, 5h, 10h and
20h in a pH 1.6 solution with 0.1 M NaCl, in the absence and presence of 0.5 mM cis
and trans azoTAB surfactants. 50
viii

Figure 2.9. PDDF for pure insulin (10 mg/ml) at t = 0.5 hrs illustrating the how R
g

and D
max
can be obtained from the plot. For a spherical scatterer, R is the diameter.
For a dimer, R describes the inter-molecular distance (illustrated in inset). Adjacent
table lists R
g
and D
max
for insulin-azoTAB surfactant systems. 52

Figure 2.10. PDDFs from (a) Insulin + 0.5 mM azoTAB (trans) after 20 hrs of
incubation, (b) Insulin + 0.5 mM azoTAB (cis) after 20 hrs of incubation 52

Figure 2.11. PDDFs of (a) Pure insulin assay after 20 hrs of incubation, (b) Insulin +
0.5 mM azoTAB (trans) after 20 hrs of incubation, assuming an extended rod-like
scatterer. The D
max
in these graphs corresponds to the diameter of the scattering rod,
and the most probable dimesion, R, to the cross-sectional radius. 53

Figure 2.12. Structure of the low pH insulin dimer. Shown from three orthogonal
views are the ribbon diagram and CPK space filling models from X-ray
crystallographic studies, along with the shape reconstructed consensus envelope in
blue, and best fit model (red insets). The consensus envelope is based on 10
independent, ab initio runs. 56

ix

Figure 2.13. Shape reconstruction of the early insulin rod-like oligomer.
Dimensions of the structure are consistent with a slightly compressed, extended
hexamer. Hypothetical ribbon diagram and spacefilling model illustrate potential
monomeric arrangement in the pre-amyloid assembly. 58

Figure 2.14. Structure of a number of molecules known to interact with amyloid
fibrils with high affinity and relative specificity. Presence of the molecules during the
aggregation process leads to structural disruption of amyloid fibrils, in some cases
blocking neurotoxicity. 61

Figure 3.1. Nile red fluorescence as a function of azoTAB concentration, for the
trans and cis isomer. The trans conformation has a marked effect on the protein
structure as seen by the sharp increase in nile red intensity (hence, increase in
hydrophobicity). 74

Figure 3.2. Hydrodynamic diameter of solution structures in catalase-azoTAB
system. Inset diagrams illustrate possible protein assembies associated with each
distribution profile 75

Figure 3.3. Catalase (4 U/ml ) activity in the presence of azoTAB surfactants. The
increase in activity suggests the presence of a super-active dimer. 77

x

Figure 3.4. SANS scattering data for the catalase – azoTAB systems at various
concentrations and after various incubation times. 78

Figure 3.5. Pair distance distribution functions for tetrameric and possibly dimeric
catalase 80

Figure 3.6. Structure of bovine liver catalase native tetramer. X-ray crystallography
based ribbon diagram and CPK space-filling model [PDB 7CAT], along with the
shape reconstructed model from SANS data. The consensus envelope, in blue, is
based on 10 independent runs. The best fit individual model is shown
in the red inset. 81

Figure 3.7. Shape reconstruction of catalase in the presence of azoTAB surfactants.
The dimer is formed at a surfactant concentration of 0.2 mM. At the higher
concentration (1.0 mM), the lopsided dimeric structure may indicate the existence of
monomeric catalase. X-ray crystallography based ribbon diagram and CPK space-
filling model [PDB 7CAT] illustrate how separation along different planes in the
tetramer results in different dimer structures. The consensus envelopes in these
diagrams are based on 10 independent runs. 83

xi

Figure 3.8. Circular Dichroism spectra of catalase in the presence of various
azoTAB concentrations and illumination states. An unfolding of α-helical secondary
structures is observed. 85

Figure 4.1. Structure of papain protein. Domain 1 is primarily helical, whereas
domain 2 is characterized by a strong β-structured content. [adapted from PDB:
1PPP] 90

Figure 4.2 CD spectra of (1mg/ml) papain secondary structural changes induced by
photo-surfactants 96

Figure 4.3. Papain SANS scattering curves of 1 mg/ml papain solution 98

Figure 4.4. Model independent PDDFs calculated for pure papain, and papain in
the presence of 0.5 mM and 1.0 mM azoTAB. An increase in the most probable
distance, r, is accompanied by a shift of D
max
to higher values. 99

Figure 4.5. Solution structure of native papain. The ribbon diagram and CPK space
filling models from X-ray crystallographic studies [PDB 9PAP], along with the shape
reconstructed consensus envelope in blue, and best fit model (red insets) are shown
from three orthogonal views. The consensus envelope is based on 10 independent, ab
initio runs 100
xii

Figure 4.6. azoTAB surfactant induced unfolding of papain tracked by SANS shape
reconstruction. The shape reconstructed consensus envelope (in blue) and best fit
model (red insets) are shown from three orthogonal views for three different
surfactant conditions. 102

Figure 5.1. SANS scattering curves for Pure Aβ(1-42) (red circles), as well as Aβ(1-
42) in the presence of cis azoTAB (green squares), and trans azoTAB (blue
triangles). 109

Figure 5.2. PDDFs for Aβ(1-42) in the presence of (a) 0.5 mM trans isomer, (b) 0.5
mM cis isomer. 110

Figure 5.3. . Structure of phosphoglycerate kinase molecule showing binding of
substrate in active site 113

xiii

Abstract
The interaction of photo-responsive surfactants with proteins has been
considered as a means to exert reversible control over a number of aspects of protein
structure and function. The azobenzene trimethylammonium bromide (azoTAB)
family of cationic surfactants undergo a photo-reversible cis to trans isomerization
upon exposure to light of the appropriate wavelength. The trans form of the molecule
has a lower dipole moment across its azo linkage, and is more hydrophobic than the
cis isomer. This results in a higher binding affinity with proteins for the trans isomer,
inducing a greater degree of unfolding of tertiary and secondary structures. The
surfactant has been applied to the study of the amyloid fibrillation pathway in insulin,
in which the protein self-associates into long, insoluble, rod-like structures. The
fibrillation rate in insulin is enhanced in the presence of the trans- isomer while the
formation of fibrils is largely inhibited in the presence of the cis- isomer, where
amorphous aggregates are observed instead. Additionally early fibrillar species
formed in the trans-azoTAB assays exhibit a greater tendency to lateral aggregation
than do structures in the pure protein, resulting in a more truncated, bundled final
aggregate morphology. Use of the surfactants as a means to control protein
quaternary solution structure has also been explored in the subunit dissociation of
tetrameric catalase. In the presence of azoTAB surfactants, catalase dissociates first
into a super-active dimer, then at higher concentrations into an aggregation prone
monomer. Finally, the structural changes associated with azoTAB-induced unfolding
xiv

of the two domain protein papain are tracked. The denaturation pathway involves a
progressive loss in secondary structure with increasing azoTAB concentration, along
with a relaxation of the compact tertiary structure, and a spatial separation of the two
domains. A number of complementary experimental techniques are combined to
determine the solution structure of non-native protein conformations, including light
scattering, circular dichroism and small angle neutron scattering.

1

Chapter 1

Introduction
1.1 Protein Structure
Proteins are organic macromolecules that take part in virtually all cell
functions throughout living organisms. They may be enzymes responsible for
catalyzing biochemical reactions, or have structural and mechanical functions, such as
generating movement or providing a system of scaffolding that supports the cell
structure. Many proteins also play an important role in cell signaling and adhesion,
immune response and growth and differentiation.[83]
Proteins are polymers comprised of different combinations of 20 naturally
occurring amino acids, known as protein residues. Each amino acid consists of a
central alpha-carbon covalently bound to a carboxyl group, an amine group, a unique
side chain, and a hydrogen atom (figure 1).

Figure 1.1. Generalized structure of the amino acid. The nature of the
R group determines classification.

2

The structure of the side chain attached to the alpha carbon can vary in size
from a single hydrogen atom, as in glycine, to long chains or large heterocyclic
groups, as in tryptophan, and determines many properties of the amino acid. The
standard amino acid side chains may be generally classified as polar, non-polar,
acidic or basic. Peptide linkages are formed between the individual amino acids
through the translation process, in which messenger RNA produced by transcription
is decoded by the ribosome, producing the main chain protein backbone. The
function of a protein is heavily dependent on its structure, and the complex three-
dimensional interactions necessary for correct protein function require diverse and
irregular protein structures. However, there are some aspects of protein structure
common to all proteins, and specific, regular structural motifs have been, and
continue to be identified.[129]
The structure of a protein is generally considered in four stages. The first
aspect is the one-dimensional sequential arrangement of amino acids along the
polymer main chain, defined as the protein’s primary structure. Due to the differing
degrees of polarity and solvent affinity of the amino acid side chains, some segments
of the polymer main chain are typically hydrophobic, while other segments are
hydrophilic. As a result, in a suitable solvent, the protein spontaneously folds,
adopting free energy-minimizing local structures in which the hydrophobic portions
of the molecule are protected from interactions with water. This folding is stabilized
by hydrogen-bonding interactions between main-chain NH and C=O groups on non-
adjacent amino acids, forming the protein’s secondary structure. The earliest
3

identified, and most prevalent secondary structural elements are α-helices and β-
Sheets, though a number of other structures have been described, including the β-
bulge, β-bend or chain reversal, as well as several less energetically favorable helical
structures. α-helices arise when hydrogen bonds running parallel to the helical axis
form between main-chain atoms on amino acids that are part of a consecutive set of
residues along the protein backbone. This type of helix is characterized by 3.6
residues per turn and the hydrogen bonds occur between the C=O of residue n and the
NH of residue n + 4. The residues at the ends of the helix do not take part in
hydrogen bonding, making them polar, and usually exposed to the solvent. There are
variations on the helix in which it is coiled more tightly (3
10
-helix, n + 3) or more
loosely ( -helix, n + 5). Both of these modified helices are less energetically
favorable than the true α-helix, and occur at lower rates, or at the ends of α-helices.
Helices have lengths ranging from as low as 5 residues to over 40, and may
theoretically have a right or left handed twist, though right handed helices are more
common in proteins as a result of the L amino acid side chains. Many α-helices
exhibit amphipathic character. These are helices with opposing polar and non-polar
faces oriented along the long axis of the helix. In globular proteins that exist in
solution, this amphipathic nature results in most helices being half buried on the
surface of the protein, with one side facing the solution and the other side facing the
hydrophobic interior of the molecule.[130] The hydrogen bonds in α-helices run
parallel to the helical axis. Since each peptide unit has a dipole moment due to the
difference in polarity between the NH and C=O, the sum effect in the helix leads to
4

an overall positive charge at the amino end of the helix and a partial negative charge
at the carboxy end. These charges attract oppositely charged ligands, though binding
sites are more frequently observed at the N-terminus. Though the stabilizing
hydrogen bonds in helices are formed between main chain atoms, different side
chains show a weak but clear preference for being incorporated in α-helical
structures. [129,83]
The second major secondary structural element found in protein is the β-sheet.
Unlike α-helices, which are formed from one contiguous segment of peptide
backbone, β-sheets are formed by a number of individual sets of consecutive amino
acid residues known as β-strands. These strands may be approximated as nearly fully
extended peptide sections which are laterally aligned, and stabilized by hydrogen
bonding between the C=O and NH main chain groups. Most β-sheets are between 5
and 10 residues long, and when a number of strands are involved in forming the
sheet, it is referred to as a pleated β-sheet. In these structures successive side chains
generally alternate positions above and below the plane of the sheet. The alignment of
the β-strands’ amino and carboxy termini which respect to one another may be
parallel or anti-parallel, arranged in a head-to-tail fashion. While the two forms of β-
sheet association lead to distinctive differences in the hydrogen bonding pattern
between strands, in both moieties, all possible main chain hydrogen bonds are
formed. Though a single pleated β-sheet may include both parallel and anti-parallel
strands, there is a clear bias against mixed sheets. All β-sheets observed in proteins
have a right-handed twist in their structure. [83]
5

These secondary structural elements are then arranged into a three
dimensional conformation, stabilized by a number of effects, including ionic
interactions, hydrophobic interactions, hydrogen bonding, disulfide bonds and
interactions with the solvent. A few specific geometric arrangements of secondary
structural features have been identified as occurring regularly, and are referred to as
supersecondary structures. Although some of these motifs have been shown to be
important in forming ligand binding sites, or the active site, they function as part of
the larger molecule. The folding and energy-minimization of the secondary structures
and functional motifs leads to the solution conformation of the protein, known as the
tertiary structure. The structure of a globular protein is usually characterized by a
stable hydrophobic core, with the loop and irregular structures on the surface. The
C=O and NH main chain groups in the loop segments are thought to form hydrogen
bonds with the solvent, further stabilizing the structure. These residues generally
contain charged and polar, hydrophilic side groups.[131] Flexibility in loop
structures may affect protein dynamics, and are frequently involved in controlling
protein function, or accessibility of the active site.
Some proteins form oligomeric assemblies of individual protein subunits held
together by non-covalent associations. The spatial arrangement of these subunits
defines the quatenary structure of the protein.(figure 1.2) The association of
monomeric subunits may allow them to work co-operatively, modulate their activity,
or enhance the peptide’s stability against degradation.
6

Figure 1.2. The four levels of protein structure [adapted from RSCB
PDBs]
7

Correct translation and subsequent structural organization and folding of the
molecule leads to the formation of the so-called native state, or biologically active
form of the protein. This native protein structure is not rigid, but a dynamic collection
of equilibrium structures representing local minima in the potential energy of the
system.[1] Depending on the flexibility of the peptide structure, these motions may
include hinge-bending, molecular breathing or conformational changes in addition to
the atomic vibrations and backbone and side chain motions which take place on a
sub-picosecond to nanosecond time scale. Though we only beginning to understand
the relationship between these motions and the biochemical activity of the proteins, it
is clear that they are functionally important.[2] Such dynamics may be influenced by
a number of factors such as pH, temperature, and ligand binding.[3] While the
modification of protein dynamics and domain motions may affect the function and
activity of the protein, further structural alterations such as protein misfolding and
protein malfunction arising from non-native conformations has been implicated in the
progression of a number of debilitating and fatal diseases such as Parkinson’s disease,
Alzheimer’s disease and Huntingdon’s disease. In these disorders, the misfolded
proteins are observed to aggregate into large fibrillar species, as will be discussed in
the following sections of this report.
In order to better understand the protein folding/unfolding processes as well as
the effects of structural fluctuations and the presence of metastable intermediates, a
number of methods to induce protein unfolding have been examined, including
chemical denaturants, pressure, temperature and pH. Surfactants, amphiphilic
8

molecules with a hydrophobic tail and polar or charged hydrophilic head group,
represent a class of small molecules that have proven quite useful in the study of
these systems. When surfactants bind to proteins, the interaction of the surfactant tail
with non-polar amino acids protects the hydrophobic core from interactions with the
solvent, allowing the protein to unfold. In this report, we employ azobenzene based
photoresponsive surfactants to control protein folding and conformation in the study
of amyloidosis and the protein form-function relationship.

1.2 Photo-responsive Surfactants
The surfactants used in these studies are part of the Azobenzene
trimethylammonium bromide (azoTAB) family of molecules. These cationic
surfactants have a hydrophilic trimethylammonium head group, a hydrophobic alkyl
tail, and between them a photo-responsive azobenzene group. An alkyl spacer group
may be added between the head group and photo-responsive motif, increasing the
molecule’s hydrophobic character. azoTAB generally exists as the planar trans
isomer in the dark or under visible light. Under ultraviolet (UV) illumination
(350nm), the molecules undergo a photoisomerization resulting in the cis form of the
molecule (figure 1.3). This change in structure is reversible, and the more
energetically stable trans form of the molecule is recovered upon exposure to visible
light (434nm). Absorption of photons of the appropriate wavelength (350nm for the
cis, 434nm for the trans) allows for rotation about nitrogen double-bond (-N=N-) in
the azobenzene group, switching between the two isomers.[4]
9

Figure 1.3. Structure of the two isomers of the azoTAB surfactants.

UV-vis spectroscopy has shown that under UV light, more than 90% of the
surfactant is in the cis form, while under visible light, around 25% exists in the cis
state.[6]

Figure 1.4. UV/vis absorption profile for the trans and cis azoTAB isomers
(1 mM, 1 mm pathlength)
10

There is a lower dipole moment across the planar trans isomer of the azoTAB
molecule than the bent cis isomer (0.5 Da for trans compared to 3.1Da for the cis
structure). This difference may be exploited to allow reversible photo-control of
various physical properties, such as electrical conductivity, surface tension and
hydrophobicity.[6]

1.2.1 Surfactant-Protein Interactions
Surfactant molecules are amphiphilic, containing polar and hydrophobic
moieties. At high aqueous concentrations the molecules associate, with their
hydrophobic chains interacting to form micelles, which have a generally hydrophobic
interior and hydrophilic exterior. The concentration at which this occurs in solution is
referred to as the critical micelle concentration (CMC), and depends heavily on the
ionic strength of the solution.[123]
In their native form, proteins' hydrophobic domains are buried at the core their
structure, minimizing unfavorable interactions with water. The co-operative action of
surfactant molecules in micelles has long been known to denature proteins by
stabilizing the proteins hydrophobic segments in the micelle core. However, at
concentrations significantly lower than the CMC, the polar head group allows the
surfactant to exist in solution as a generally free molecule, in spite of its hydrophobic
tail. At very low concentrations (well below the solution CMC) charged surfactants
have greatly increased binding affinity with peptides, and have been shown to lead to
11

significant unfolding, and in some cases complete protein denaturation.[126] In fact,
at these low concentrations, some surfactants may be around 1000 times more
efficient than traditional chaotropic chemical denaturants at altering protein structure
due to their fundamentally different interactions with peptides. Ionic surfactants can
bind to proteins through polar or electrostatic interactions involving the head group
and hydrophobic interactions involving the alkyl tail.[127] Depending on the protein
structure, the surfactant may have well defined binding sites, and due to strong
affinity with the protein, it has been proposed that, even concentrations much lower
than the CMC can lead to the formation of local micelle-like structures on certain
protein structures.[128]
Surfactants differ from other ligands in that they bind not only to the native
protein conformation, but also to the partially unfolded and denatured states, and it is
a higher affinity for these denatured states that drives protein unfolding. [124,125]
The systems discussed in the following sections employ azoTAB surfactants at
concentrations much lower than the CMC for the surfactant (~10 mM for trans
isomer in H
2
O), and with protein:surfactant ratios ranging from as low as 1:5 for
small proteins such as insulin, to over 1:100 for larger proteins (eg. Catalase
tetramer). Cationic surfactants like azoTAB are thought to bind to complementary
anionic side chains (Glu, Asp), while the alkyl chain binds to nearby hydrophobic
patches, or are involved in other energy lowering interactions.[128] Minor changes in
tertiary structure can have profound effects on wide variety of protein physical and
biochemical properties.[6]
12

Given the range of potential protein-surfactant interactions, it is not surprising
that surfactants have effects on proteins beyond simple denaturation. We explore
some of these effects in the following chapters using the photo-responsive surfactant
azoTAB. Of the two photo-isomers of this molecule, there is a lower dipole moment
across the planar trans form than the bent cis form (0.5 Da for trans compared to 3.1
Da for the cis structure). The relatively hydrophobic trans isomer has a greater
tendency to bind to proteins than the cis form, interacting more with hydrophobic
regions of the protein molecules. Binding eventually leads to the exposure of
hydrophobic residues to the solvent, and conformational changes often involving
extensive unfolding of the protein’s solution structure.[7]
It is the unique photo-responsive properties of these surfactants that make
them an important tool in studying protein structure. The ability to switch easily
between the more hydrophobic or hydrophilic forms of the molecule in solution gives
us a means to reversibly control the extent of protein binding, and induced unfolding,
essentially providing reversible control of protein conformation, function and
dynamics.[8,9] The degree of binding/unfolding can be further controlled by
adjusting the length of the hydrocarbon tail on the surfactant.[10] The longer the
hydrophobic hydrocarbon tail, the stronger the hydrophobic interactions with
proteins, and extent of unfolding. The structure of the protein also influences
surfactant binding. It has been shown that the azoTAB surfactants preferentially bind
to α-helical segments within the protein, which are relatively more hydrophobic than
β-structures, leading to greater relaxation in those sections of the protein’s secondary
13

structure. The unique properties of these azoTAB surfactants have been applied in our
group to provide photocontrol of protein secondary, tertiary and quatenary structure,
as well as enzyme activity and protein aggregation.[11,8,12,13] The surfactants are
synthesized by an azo-coupling reaction of phenol with 4-ethlyalaniline, then
reactions with 1,2-dibromoethane and trimethylamine, according to published
methods [14].

1.3 Protein-Protein Interactions and Amyloid
Fibrilogenesis
Protein aggregation is a physical process involving the association of several
protein molecules, mediated by non-covalent interactions. Fibrillation refers to a
mode of protein aggregation which results in the formation of long, highly ordered
rod-like structures called fibrils. One interesting feature of the fibrillation process is
that the observed fibrils share many striking morphological similarities with each
other, in spite of being derived from a number of disparate proteins with no obvious
similarities in residue sequence. This suggests a shared aggregation mechanism on the
molecular level. In addition to having similar dimensions (~ 4-14 nm diameter, 0.1-
10 um length), the fibrils all have intermolecular β-sheets running perpendicular to
the axis of elongation, and as such can bind dyes such as congo red, which are
thought to intercalate themselves into the newly formed β-structures. [15]
The ability of a protein to perform its biological function depends heavily on
the correct folding of the peptide chain making up the backbone of the molecule. It is
14

thought that incorrect folding may make the protein prone to aggregation. Amyloid
fibrillation represents one of the most dramatic consequences of protein misfolding.
However, given the large and growing number of proteins capable of forming such
structures under certain conditions it has been suggested that fibrillation may be a
generic property of structurally modified peptide chains, in which main chain
interactions outweigh the stable side-chain contacts that define the native, active state
of the protein.[16] The protein’s partially folded states, such as the molten globule in
which these tertiary contacts have been loosened, often have exposed hydrophobic
regions, which preferentially interact with each other rather than be exposed to
energetically unfavorable interactions with the solvent. This leads to a cascading
aggregation via a nucleated growth mechanism.[17]

Figure 1.5 Theorized mechanism of fibrilogenesis following the pathway:
Monomers → Oligomers → Protofibrils → Protofilaments → Fibrils

15

In the above scheme, unfolding of the monomer is followed by the formation
of nuclei (lag phase), then the nuclei associate into protofibrils, the first species to
display intermolecular β-sheets.
The specific molecular mechanisms underlying this phenomenon are a topic
of great interest, as fibrillation has been implicated in the progression of a number of
neurodegenerative diseases, including Alzheimer’s disease, Parkinson’s disease,
Huntington’s disease and Creutzfeldt-Jakob. The protofibrils associate into
protofilaments and eventually into fully formed fibrils, which accumulate in diseased
tissue forming plaques. However, it is the early oligomers that are increasingly
viewed as the primary pathogenic species. [17,18] It has been proposed that the
formation of large, highly ordered fibrils and inclusion bodies may be a protective
mechanism, reducing the concentration of pathogenic precursors along the fibrillation
pathway. [19] Thus, when targeting the fibrillation process as a potential therapeutic
strategy it is important that we understand the mechanism of inhibition, and can
ensure that the process is interrupted prior to the formation of the pathogenic species,
lest we exacerbate the problem by simply inhibiting the production of large fibrils,
shifting the equilibrium towards the small, potentially toxic oligomers.

16

1.4 Proteins
1.4.1 Insulin
The Insulin monomer has a molecualr weight of ~5800 kDa, and is comprised
of two peptide chains linked by disulfide bridges. Both peptide chains contain large
α-helical sections, and the longer B chain also contains a short β-structured section.
In its native state, at neutral pH insulin exists primarily as a toroid-shaped hexamer.
This is the form of the protein stored in the B-cells in the islets of Langerhans. It is
thought that the hexameric assembly facilitates efficient storage of high
concentrations of the hormone, as well as protecting the protein from degradation and
aggregation.[20, 106] The active form of the molecule is the monomer, but
destabilization of the monomer leads to self-association, and initiation of the
fibrillation pathway.[21]

(a) (b)
Figure 1.6 Structure of the native insulin hexamer (a), and the monomer
(b), showing the A chain (blue) and B chain (green) [PDB 4INS and 1AIY]
17

While many proteins may take part in the fibrillation process, Insulin is an
ideal model protein for studying the fibrilogenesis mechanism. While the large
fibrillar structures formed in this system are morphologically very similar to those
derived from disease proteins such as (Aβ) peptide fragments, insulin is readily
available, relatively cheap, and well-studied.
The primarily α-helical secondary structure of the protein makes it a target for
binding of the azoTAB surfactants. In the presence of the trans isomer, the
aggregation rate is increased with respect to the pure protein. Conversely, the cis
isomer extends the lag phase and apparently disrupts the formation of fibrils. The
effects of azoTAB surfactants on the fibrillation process will be discussed in more
depth in chapter 2 of this report.

1.4.2 Catalase
Catalase is a ubiquitous protein in aerobically respiring organisms. This
enzyme protects the cell from the harmful effects of hydrogen peroxide, by catalyzing
its decomposition to oxygen and water.[22] The catalase molecule is a 500 residue
long, 60 kDa, multi-domain protein that exists as a hydrophobically bound tetramer.
The first domain contains the active site, buried in a hydrophobic pocket near the
center of the structure (~20 Å below the molecular surface), consisting of a β-barrel
associated with a heme moiety. Access to this conformationally rigid catalytic motif
is controlled by a largely helical domain that is involved in forming a hydrophobic
18

channel to this portion of the molecule. The third and fourth domains appear to be
involved in maintaining the enzyme’s quaternary structure. A globular domain with
two α-helical arms, and a ‘wrapping’ domain which extends around the outside of the
molecule and is capable of forming β-structured contacts with other monomer units,
stabilize the native tetrameric structure of the protein.[22, 23, 24] Changes in
structure resulting in dissociation of the native oligomeric protein have been shown to
yield dimers (both inactive and super-active forms have been observed) and inactive,
aggregation-prone monomers. [25, 26]

Figure 1.7 Ribbon diagram structure of the native catalase tetramer from
bovine liver. [PDB code 7CAT]

In the presence of azoTAB surfactants, the catalase tetramer dissociates into
smaller species. At low surfactant concentrations, a dimer with enhanced activity is
favored, and at higher surfactant concentrations the equilibrium shifts towards the
monomers, which go on to form aggregates. This effect is more pronounced in the
19

presence of the trans isomer, due to its increased protein-binding tendency relative to
the cis isomer. The photo-responsive nature of the surfactant provides the potential
for reversible photo-control of the enzyme’s structure and activity. This system will
be discussed in greater detail in chapter 3 of this report.

1.4.3 Papain
Papain is a thiol protease from the Carica papaya plant. The ~23 kDa, 212
residue protein is a single polypeptide chain folded into two domains of a similar size,
but which are structurally very distinct. The first (domain 1) is primarily α-helical,
consisting of residues 10 to 111 and 208 to 212, while the second (domain 2) is
largely β-structured, spanning residues 1 to 9 and 112 to 207.

Figure 1.8. Structure of papain protein. Domain 1 is primarily helical,
whereas domain 2 is characterized by a strong β-structured content. [PDB
1PPP, 9PAP]

20

The active site of this protein rests in the cleft created between the two
domains, and involves residues from both. Domain 1 includes 3 helices, and has a
hydrophobic core. The longest of these helices, which is partially buried, contains the
catalytically important Cys25. The thiol group on this residue forms a catalytic dyad
with a histadine residue (His159) from domain 2, which deprotonates it allowing
nucleophilic attack of the substrate carbonyl group by the cysteine’s anionic sulfur.
The molecule’s remaining six cysteine residues are involved in forming disulfide
bridges. Two of the three helical segments in domain 1 are imperfect α-helices,
showing hydrogen bond disruption and 3
10
-character. Domain 2 is characterized by
its antiparallel β-sheet structure, though it does contain two short helices on the outer
shell on the domain. The residues involved in these β-sheets form a barrel-like cavity
near the center of the domain. The active site, near the molecular center, is thought to
be the most conformationally rigid portion of the protein, with residue mobility and
hydrogen bonding distance increasing towards the surface of the molecule.[86]
This protein represents an interesting target for the azoTAB surfactants. The
denaturation process in papain is thought to be biphasic, suggesting the domains may
be capable of folding or adopting non-native conformations independently of one
another.[83] The tendency of azoTAB surfactants to preferentially bind to α-helical
structures means that it may be possible to structurally alter one domain, while
preserving tertiary contacts in the other domain. In this study, we consider the effect
of azoTAB surfactants on the secondary and tertiary solution structure of papain.
This is discussed in greater detail in chapter 4 of this report.
21

1.5 Experimental Techniques
1.5.1 Dynamic light scattering (DLS)
In this visible light scattering technique, a laser beam is passed through a
solution containing protein molecules, which scatter the light. The scattered light
intensity is determined by the shape, size and molecular interactions of the particles in
solution. The time variation of the intensity of scattered light at some angle (usually
90 degrees) is measured, and related to the diffusion coefficient of scattering species
by an auto-correlation which describes how the measurement relates to itself over
time. The decay of the autocorrelation diminishes over time, and can be modeled by
an exponential decay function,

Thus, we can express the autocorrelation in terms of the diffusion coefficient, D, and
the measurement vector, q, where q = 4π/λ·sin(θ/2), λ is the incident wavelength
(632.8nm) and θ the scattering angle. In general, small particles diffuse faster than
large ones. After the diffusion coefficient has been determined, the hydrodynamic
diameter is estimated (assuming a spherical shape) using the Stokes-Einstein
equation,

where k
B
is Boltzmann’s constant, n is solvent viscosity, and T is temperature.[27]

22

1.5.2 Small Angle Neutron Scattering (SANS)
Small angle neutron scattering is a relatively novel technique for studying
protein solution structure. A monochromatic beam of cold neutrons is scattered by a
liquid sample solution, and the two dimensional scattering pattern, determined by the
size and structure of the scattering species, is collected at various distances. One of
the fundamental differences between SANS and light scattering is that neutrons,
which have no electric field do not interact with the electron cloud associated with
large protein molecules, and are instead scattered by the atomic cores, giving
structural information on atomic position. However, as in light scattering, the
incident wavelength and scattering angle determine the length scales probed by the
relationship d ≈ λ/θ, where λ is the wavelength and θ the scattering angle. For SANS
which uses cold, long wavelength neutrons (~6 Å) and small scattering angles, the
structure of species in the 60 Å to 1000 Å size range can be probed. [28]
A common analysis performed on SANS data utilizes the Guinier
approximation for scattering intensity,

where Q is the measurement vector in inverse space, related to the real-space length
scale by Q = 2π/L and having units of Å
-1
, I(Q) is the extrpolated intensity at Q = 0 ,
and R
g
is the radius of gyration. [44] Through this approach, the radius of gyration
can be determined from the slope of a plot of ln(I(Q)) vs. Q
2
. This analysis can be
23

applied to non-globular scatterers by using the modified guinier analysis. For
aggregating systems in which we believe rod-like intermediates are present along the
fibrillation pathway, the approximation becomes

where I
c
(0) is the extrapolated intensity at Q = 0 and R
c
is the cross-sectional radius of
gyration, which can be calculated as the slope of a plot of ln(Q*I(Q)) vs. Q
2
.
Another data analysis technique that can be applied to SANS is the calculation
of a pair distance distribution function (PDDF), which provides information on the
maximum dimension (D
max
) within the protein or protein assembly by calculating the
probability (P(r)) of finding two scattering centers separated by a distance r, through
the following equation.

In the determination of the solution structure of scattering species in SANS,
shape reconstruction algorithms prove a powerful technique. This approach
determines the location of each scattering center within a protein based on the
scattering of the whole structure by assuming an initial number, and spatial
distribution of scattering centers, then iteratively adjusting the position of each until
the calculated I(Q) matches the experimental data. [13]
In studying aggregating systems SANS has two distinct advantages over
traditional characterization techniques. First, SANS is an absolute technique,
meaning that the weight-average molecular weight (M
w
) of the sample can be
24

determined directly from I(0), the scattering extrapolated at zero angle, through the
relationship,

where δ
s
and

δ
p
are the scattering length densities of the solvent and the
protein respectively, v is the protein-specific volume, and c is the protein
concentration. Second, scattering is additive, so that the scattered intensity from a
solution of a non- interacting species A and B, I
mix
= I
A
+ I
B
. This allows
contributions from multiple species to be deconvoluted and shape reconstructed
separately or even contrasted out through the use of appropriate solvents.[28]

1.5.3 Atomic Force Microscopy (AFM)
This scanning probe imaging technique generates a topographical image of a
surface by measuring the strength of forces between the sample surface and a very
sharp probe tip (several nm tip diameter) which scans across the sample surface.
While a number of imaging modes are available, the so called non-contact AFM
mode is used for soft biological samples, which may be damaged by contact with the
silicon tip. In the non-contact imaging mode, the probe tip is brought very close to
the sample surface (<1 Å) and the cantilever on which the probe tip is mounted
oscillates with a set frequency corresponding the resonant frequency for the sample.
As the tip is scanned across the surface of the sample, changes in the separation
between the tip and the sample can be measured by changes in the strength of the Van
25

der Waals attractive interaction, which dampens the oscillation of the cantilever.[30]
In these experiments, the samples are prepared by depositing protein solution on
freshly cleaved mica, and imaging is performed at ambient temperatures and
pressures.

1.5.4 Circular Dichroism (CD)
Circular dichroism refers to the differential absorption of left and right
circularly polarized light by an optically active chiral molecule. [31] Circularly
polarized light is described as light in which the electric vector is rotating about the
direction of movement of the photon. When the chromophores of the amides of the
peptide backbone are aligned in arrays, their optical transitions are shifted as a result
of exciton interactions.[113] As a result, in the far ultraviolet (190-250nm), different
protein secondary structural motifs have very distinctive CD spectra, making CD an
effective technique for quickly quantifying the secondary structural elements of
polypeptide molecules. α-helical proteins have negative bands at 222 nm and 208 nm
and a positive band at 193 nm, whereas well-defined anti-parallel β-pleated sheets
have a negative band at 218 nm and a positive band at 195 nm. Disordered peptide
segments, or ‘random coils’ have very low ellipticity over 210 nm and negative bands
near 195 nm.[114] By decomposing the protein spectrum into these three individual
characteristic spectra, we can estimate the contribution of each to the protein’s
secondary structure. [32, 112] While Circular Dichroism cannot provide direct insight
into the secondary structure of specific residues or the protein’s tertiary structure and
26

conformation, as X-ray crystallography or NMR do, it has the advantage that data can
be collected in a matter of minutes across a variety of solution conditions, providing
information on non-native protein states. Knowledge of the secondary structure can
be a good measure of the extent to which a protein has unfolded, or which portions of
the molecule have undergone structural changes. We have used CD to analyze the
conformational changes associated with surfactant binding in a number of proteins, as
well as in the study of amyloid fibrillation, where it can track the unfolding of the
monomer and formation of characteristic intermolecular β-sheets present in the
fibrillar species. One of the difficulties faced in the use of CD to study azoTAB
related structural changes is the broad UV absorption profile of the surfactant.
Ideally, CD spectroscopy buffers contain no optically active material and are as
transparent as possible. For optimal transparency, the maximum surfactant
concentrations used in CD measurements is 0.5 mM, and the wavelength range is
decreased slightly to exclude surfactant absorption below 195 nm.

27

Chapter 2
The Effect of Photo-responsive surfactants on Insulin
Fibrillation
2.1 Abstract
The azoTAB (azobenzene trimethylammonium bromide) family of surfactants
undergo a trans- to cis- conformational change on exposure to ultraviolet light. The
hydrophobic trans isomer (favored under visible light) shows a greater tendency to
bind with proteins. The ability to reversibly bind the surfactant to the protein by
changing light conditions allows for photo-control of protein conformation. We
explore its applications in the study of the amyloid fibrillation pathway.
Amyloid fibrilogenesis, a process by which proteins self-aggregate into
insoluble, fibril-like structures, has been implicated in the progression of a number of
neurodegenerative diseases. Morphological similarities in the fibrils formed from
various non-homologous proteins suggest that the fibrillation occurs via a shared
molecular mechanism. Insulin is one such protein, showing a tendency to form fibrils
under certain conditions. Insulin exists as a hexamer in its native state. At low pH
(~2) while the protein is more stable as a dimer, some fraction exists as the monomer,
and at elevated temperatures, significant fibrillation is observed. In addition to
exploring a number of techniques aimed at the identification and characterization of
transient intermediates, we have developed a method to inhibit the aggregation using
28

the azoTAB surfactants. Through dynamic light scattering (DLS), circular dichroism
(CD) atomic force microscopy (AFM) and small angle neutron scattering (SANS)
experiments, we have shown that the fibrillation rate in insulin is enhanced in the
presence of the trans- isomer while the formation of fibrils is largely inhibited in the
presence of the cis- isomer, where amorphous aggregates are observed instead.
Additionally early fibrillar species formed in the trans-azoTAB assays exhibit a
greater tendency to lateral aggregation than do structures in the pure protein.

2.2 Introduction
Late onset neuro-degenerative disorders such as Alzheimer's disease,
Parkinson’s disease, Creutzfeldt-Jakob, Huntington's disease, amyotrophic lateral
sclerosis (ALS) and familial amyloid polyneuropathy appear to be increasing in
prevalence. Alzheimer’s is a fatal dementia involving progressive memory loss, and
diminished speech and recognition skills. It has been estimated that by 2050
Alzheimer’s Disease could affect as many as 1 in 85 people globally.[33] However,
we are yet to develop effective treatment strategies for this family of diseases. This is
largely because the pathogenesis of these disorders remains a mystery. The issue is
being approached from a number of different perspectives; it has been shown for
example, that for some of these disorders there is a clear genetic link, and some
researchers are studying the synaptic dysfunction and altered neuronal connections
associated with these illnesses. However, recent work has converged on stable,
insoluble, ordered protein aggregates, which eventually associate into exclusion
29

bodies such as senile plaques, and characterize almost all such diseases. The
correlation and colocalization of neurodegeneration in disease tissue and protein
fibrillation suggests that the process either directly contributes to cell death and
toxicity, or at least is an inseparable phenomenon.[19] The threshold for the disease
state in vivo is very similar to threshold for aggregation in vitro, supporting the
hypothesis that aggregates or a tendency towards aggregation are responsible for
toxicity. Genetic studies further support this theory, as mutations identified as being
responsible for a number of disorders are localized to genes encoding associating
proteins observed in disease tissue, as with amyloid-β peptide fragment 1-42 in
Alzheimer’s disease.[34] Inhibition or interruption of the aggregation process is
considered a possible therapeutic route in treating these diseases, but requires a
thorough understanding of the molecular mechanisms associated with fibrillation.
One of the challenges when studying dynamic aggregation processes, such as
amyloid fibrillation, is the identification and characterization of transient
intermediates that may be in equilibrium with other species along the fibrillation
pathway. We have explored various features of the fibrillation process using the
model protein insulin, employing a number of techniques and methods aimed at
characterizing the solution structure of the protein.
Insulin is a hormone produced in the pancreas, within the B-cells of the islets
of Langerhans. Insulin controls the uptake of glucose in the bloodstream by target
cells, which would otherwise become toxic and plays a vital role in the regulation of
the metabolism of fats and carbohydrates. [35] The protein acts by binding to a
30

transmembrane tyrosine kinase receptor in adipose cells and muscle tissue, which in
turn phosphorylates a substrate protein engaging effector molecules.[106,107] The
Insulin monomer is a 51 residue protein with molecular weight of ~5800kDa. It is
comprised of two peptide chains; chain A of 21 amino acids and chain B of 30 amino
acids. Two disulfide linkages connect the two chains, and a third links two cysteine
residues within chain A.[36] The A chain contains two antiparallel helices connected
by a loop. One is a true α-helix, while the other shows 310-helix character. The B
chain also contains an α-helix, as well as an extended section near the C-terminus,
analogous to a β-strand, which takes part in the formation of intermolecular β-sheets
with other insulin monomers in the formation of quaternary structures.[37]
In its native state, insulin exists as a zinc-bound, toroid-shaped hexamer. The
destabilization of this native state can induce rapid aggregation of the protein into
insoluble fibers. The process has been studied quite extensively as insulin is used as a
drug to treat diabetes, and the instability of the material during processing, storage
and delivery continues to be a major challenge.[38] Factors leading to disruption of
the protein’s native conformation include temperature, acidic pH (~2), high
concentration, agitation and contact with hydrophobic surfaces. At low pH, the
insulin hexamer is no longer favored, and the monomer and dimer become the
dominant species in solution. Due to hydrophobic interactions, 12 C-terminal
residues of the B chain form hydrogen bonds between monomers, making the dimer a
stable solution structure.[37] It is thought the partial unfolding of the monomer leads
to amyloid fibrillation. As with other fibrillation processes, the aggregation is
31

thought to be hydrophobically driven, and the unfolding of the monomer and
nucleation of early intermediate states is the rate-limiting step.[39] However, the
underlying molecular mechanisms mediating fibril formation in insulin and other
fibril-forming proteins are not well understood, and identification of interacting
domains in the aggregates continues to be the subject of much research.[40,41]
Insulin’s ability to readily aggregate in solution, forming fibrils, combined with the
wealth of biochemical and structural information available on the protein, make it an
ideal model protein to study the fibrillation process. We have developed a method to
control the aggregation process in vitro through use of the azoTAB surfactants, and
characterize the prefibrillar intermediates through a combination of techniques,
including dynamic light scattering (DLS), small angle neutron scattering (SANS),
atomic force microscopy (AFM), and circular dichroism (CD). The formation of
fibrils in insulin is inhibited in the presence of the azoTAB cis- isomer, while
aggregation proceeds by nucleation and growth in the presence of the trans- isomer.
Furthermore, aggregates formed in the presence of the cis isomer appear to be
structurally distinct from the characteristic fibrils formed by the pure protein, or in the
presence of the trans isomer.

32

2.3 Experimental Methods
2.3.1 Materials
The azobenzene trimethylammonium bromide surfactants synthesized via an
azo-coupling reaction of phenol with 4-ethlyalaniline, then reactions with 1,2-
dibromoethane and trimethylamine, according to published methods [9,11]. The
surfactants have a hydrophilic trimethylammonium head group and a hydrophobic
alkyl tail separated by a photo-responsive azobenzene group (figure 2.1).

Figure 2.1. Structure and isomerizatioin of the azoTAB surfactants

When illuminated with ultraviolet light (350 nm wavelength), the surfactant
undergoes a rapid, reversible photoisomerization, from the energetically favorable,
planar trans form to the bent cis form (>90% conversion). The trans isoform is
largely recovered upon exposure to visible light (75% conversion at 434 nm), and
completely recovered in the dark (100% conversion at ~24 hrs). Interconversion
between the surfactant isomers was achieved with a 200W mercury arc lamp (Oriel,
model number 6283). For the cis isomer, an incident wavelength of 365 nm was used,
obtained through combination of a 320 nm band-pass filter (Oriel model number
33

59800) and a heat absorbing infra-red filter (Oriel model number 59060). To revert
to the predominantly trans form, a long pass filter (400nm, Oriel model number
59472) was used. For experimental samples containing the trans form of the
surfactant, solutions were kept in the dark to minimize contributions from the cis
isomer. For samples containing the cis isomer, conversion of the surfactant was
performed prior to combination with the protein solution, and the cis form was
maintained with an 85W long wave UV lamp (365 nm Spectroline, model number
XX-15A). Control assays confirmed that bombardment by these wavelengths had no
measurable effects on the solution structure of insulin. Surfactant conformation was
determined by UV/Vis absorbance spectra.
Insulin from Bovine pancreas was purchased from Sigma-Aldrich (catalog
number I5500) and used as received. Protein concentrations between 1 mg/ml and 10
mg/ml were used. While higher protein concentrations improve light and neutron
scattering statistics, they were not studied as the aggregation process in insulin is
highly concentration dependent, and to better approximate the infinite dilution
assumption in the determination of the diffusion coefficient. Protein solutions were
freshly prepared for each fibril formation experiment in a buffer containing 0.1 M
NaCl and adjusted to pH 1.6 by addition of HCl. For small angle neutron scattering
(SANS) assays, solutions were prepared in D
2
O rather than H
2
O to minimize
incoherent scattering from hydrogen atoms.

34

2.3.2 Dynamic Light Scattering
DLS measurements were performed on a Brookhaven BI-200SM instrument
(Brookhaven Instrument Corporation). The incident beam was a 632.8 nm, 35mW
helium neon (HeNe) laser (Melles Griot, model Number 05-LHP-928). This
wavelength, while in the visible range, is long enough that it does not convert the
surfactant from the cis to the trans form. The scattering was measured at 90
o
and 45
o

in order to increase the Q-range and observe scattering contributions of small species
such as the insulin monomer and dimer. Data was analyzed using the non-negative
least squares (NNLS) routine by a Brookhaven Instrument Corporation BI-9000AT
digital correlator. Prior to data collection the protein solutions were filtered through a
200 nm Anotop filter to remove large structures which could act as nucleation
centers. UV/vis spectroscopy was used to confirm protein concentration after
filtration. As aggregation is a dynamic process, a time dependent DLS analysis was
performed, with data points collected every 3 - 5 minutes during the early stages of
the fibrillation pathway, and a relatively high protein concentration (10 mg/ml)
ensuring sufficient scattering intensity to fit the auto-correlation curve. The
temperature was maintained at 60
o
C, and the illumination state could be changed in
situ through the use of a fiber-optic light guide. The hydrodynamic radius, R
H
, was
calculated assuming a spherical shape, according to the Stokes-Einstein equation,

where k
B
is Boltzmann’s constant, T is the temperature, n is the solvent viscosity, and
D is the experimentally determined diffusion coefficient.
35

2.3.3 Atomic Force Microscopy
AFM was performed using a Park Scientific instruments (now Bruker
corporation) Autoprobe CP atomic force microscope, mounted on a pneumatic
vibration isolation table. Equipped with a piezo-electric scanner, the instrument has a
maximum scan range of 100um
2
in the x-y plane, and ~5um in the z-direction (height
of surface structures). The AFM was operated in non-contact mode, as the soft nature
of biological samples makes them susceptible to damage from contact with the probe
tip. The scanning probe was a silicon tip incorporated into a rectangular cantilever
(Veeco probe model number RTESPA). The radius of curvature of the tip was below
10 nm, and the cantilever oscillation frequency was ~350kHz. Samples were
prepared by depositing a small aliquot of protein solution (20µl) on freshly cleaved
mica, and wicking the majority of it away with filter paper, creating a thin film of
protein solution. The remaining solvent was driven off by evaporation, and imaging
occurred at ambient conditions, <24hrs after sample preparation.

2.3.4 Small Angle Neutron Scattering
Small angle neutron scattering experiments were performed at Oak Ridge
National Laboratories (ORNL), on the CG3-BioSANS instrument at the High Flux
Isotope Reactor facility (HFIR). The cold neutron wavelength was 6 Å, and the two
detector distances of 1.3 and 7 meters yielded a Q range of 0.007 - 0.3 Å
-1
. The net
intensities were corrected for the background (D
2
O based buffer) and empty cuvette.
Variations in the response efficiency of individual pixels in the 2-dimensional
36

detector were accounted for by considering the scattering from an isotropic scatterer
with uniform scattering intensity across the entire Q-range, and converted to an
absolute differential cross section per unit sample volume (in units of cm
-1
) using an
attenuated empty beam. Samples were prepared in a D
2
O based buffer to minimize
incoherent scattering from hydrogen. Due to the lengthy nature of SANS data
collection, a lower protein concentration of 1 mg/ml was used to slow the aggregation
process. Additionally, during data collection the samples were ‘quenched’ to halt
further association, and the temperature of the sample chamber was set to 15
o
C.
Lower temperatures were not used to avoid ambient H
2
O condensation on the cuvette,
which significantly affects scattering.
Three analysis techniques were employed to consider the scattering data. To
calculate the radii of gyration of scattering species, the Guinier approximation was
used.

where Q is the measurement vector in inverse space, related to the real-space length
scale by Q = 2π/L and having units of Å
-1
, I(Q) is the scattering intensity, I(0) is the
extrapolated intensity at Q = 0 and R
g
is radius of gyration, valid for globular
species.[44] For samples where the protein had begun to form pre-fibrillar oligomers,
a modified Guinier analysis for rod-like structures was applied. Here the
approximation takes the form,

37

where I
c
(0) is the extrapolated intensity at Q = 0 and R
c
is the cross-sectional
radius of gyration, related to the geometric radius by R
c
2
= R
2
/2 which can be
calculated as the slope of a plot of ln(Q*I(Q)) vs. Q
2
. [45]
Pair distance distribution functions (PDDFs) were calculated assuming a
monodisperse system of scatterers.[109] PDDFs provide information on the
maximum dimension (D
max
) within the protein or protein assembly by calculating the
probability (P(r)) of finding two scattering centers separated by a distance r, through
the following equation.

D
max
, the maximum dimension within the particle was selected to ensure a smooth
return of the PDDF curve to zero at D
max
.
Shape reconstruction was performed using GA_STRUCT.[110] The program
fits the experimental data by starting with an initial estimate of the protein structure (n
randomly oriented scattering centers) and rearranging them until the synthetic
scattering data fit the experimental data. The number of scattering centers used in
each run was selected in order to approximate the theoretical number of atoms in the
protein or protein assembly. The program produces ten independent models,
compares them for structural similarity, and averages them to give a consensus
envelope.[110]

38

2.3.5 Circular Dichroism
CD measurements were performed at the University of Southern California
Bio-imaging Center on a Jasco J815 instrument. Solutions were prepared with final
protein concentrations between 0.1 and 1 mg/ml, and the spectra were collected using
a 1 mm pathlength cuvette at a temperature of 60
o
C.
Each spectrum was obtained from the average of four scans, and were
background-subtracted and normalized to molar ellipticity. The resulting spectra were
then analyzed using JFIT to quantify the contributions of α-helical, β-sheet, and
random coil secondary structural elements. Due to the azoTAB surfactant’s strong
ultraviolet absorption, a maximum surfactant concentration of 0.5mM was used, and
the effective spectral range was 196 nm - 240 nm. Data were collected for the protein
in the presence and absence of azoTAB trans and cis isomers.

2.4 Results and Discussion
During the rate-limiting nucleation phase of the fibrillation process, the
formation of early oligomeric species was tracked using DLS, which can provide
nanometric resolution of the size distribution profile of solution structures
(figure.2.2a). UV/Vis absorbance at 600 nm (outside the wavelength range in which
the protein or surfactant absorb) gives a measure of solution turbidity, and was used
to track the formation of large fibrils during the growth phase (figure. 2.2b).
39

(a) (b)
Figure 2.2. Time dependent DLS data (a), and UV/vis absorbance (b), for
insulin, and insulin in the presence of azoTAB surfactants DLS

The data show that in the presence of the azoTAB trans isomer, the rate of
insulin aggregation increases compared to the pure protein. However, under
ultraviolet illumination, in the presence of the cis isomer, the rate of aggregation
decreases relative to pure insulin. The high protein concentration (10 mg/ml), high
temperature (60
o
C) and low pH (1.6) used in these assays create conditions in which
insulin aggregation is highly energetically favorable. As a result, though the cis
isoform acts to extend the lag phase, all systems are seen to aggregate eventually. At
significantly lower temperatures (<25
o
C) or protein concentrations (<1 mg/ml), the
cis isomer indefinitely inhibits this process. While DLS is capable of deconvoluting
the scattering from different sized structures in solution and quantifying their relative
contributions, for the sake of clarity, the DLS data plotted in figure 2.2a represents
the largest species detected by DLS for each experimental condition as a function of
40

time. Figure 2.3, on the other hand, illustrates the early events in the fibrillation
pathway for pure insulin followed by DLS, showing all species detected in solution.

Figure 2.3. Time dependent DLS data for pure insulin (10 mg/ml) incubated at 60
o
C
and pH 1.6. Inset diagrams illustrate potential oligomeric intermediates along with their
hydrodynamic radius calculated using the modified Stokes-Einstein equation for the
infinite dilution diffusion coefficient of a cylindrical species.

The data show that initially, the dominant scattering species are the dimer and
possibly the monomer, with 100% of scatterers having a hydrodynamic diameter
between ~1.5 and 3 nm (in agreement with published values).[21] After about 15
minutes another species becomes detectable between 5 and 10 nm. By the time larger
41

aggregates begin to form, after around 20 minutes, the equilibrium is shifted further
away from the monomer/dimer, and only around 1 - 2% of the scattering is attributed
to these small species. A similar analysis of the DLS data for the protein in the
presence of the cis isomer (not plotted) indicates that not only is the formation of
early aggregates delayed, but even after these oligomers form, scattering from the
monomer remains significant. The plotted hydrodynamic diameter in the DLS data in
figures 2.2a and 2.3 assumes a spherical scattering species. While this may be
appropriate for the early globular species, once the rod-like protofibrils and larger
fibrillar structures begin to form, the assumption becomes less valid. The diagrams
inset in figure 2.3 show possible oligomers along the aggregation pathway. The
calculated hydrodynamic diameters quoted along with the diagrams are obtained
using the expression of the infinite dilution diffusion coefficient D
0
of a cylinder of
length L and diameter d,

Where k
B
is the Boltzmann constant, T is the temperature, δ = L/d, γ = 0.312 + 0.565/δ
- 0.1/δ
2
, and η is the solvent viscosity. After modification of the monomer structure
(partial unfolding) inducing aggregation, the first stable protein assembly detected by
DLS is consistent with a hexameric protofilament. These protofilaments then
associate, either laterally or end-to-end. While DLS is useful for tracking the
kinetics of the prefibrillar events, it cannot provide information on the morphology of
the oligomers.
42

Atomic force microscopy was used to characterize the oligomeric
intermediates as well as the resultant fibrils (figure 2.4). Initially, the only features
visible in the AFM images are ~1 - 3 nm globular species, consistent with the
monomer/dimer. After 12 minutes the first signs of ordered association appear. The
images suggest that the first stable intermediate is a rod-like assembly, consisting of 6
- 7 repeat units. These likely hexameric protofibrils have dimensions similar to those
of the first observed aggregates in the DLS data (figure 2.3). By 35 minutes these
protofibrils have begun to aggregate into the protofilaments that precede formation of
the full fibrils. Two modes of protofibril attachment are observed in the pure protein.
The first is lateral aggregation, forming bundles with a height of ~10-12 nm in the
AFM images. However, some fraction of these nuclei instead take part in end-to-end
association, forming very long, thin protofilaments with a height similar to that of the
monomer/dimer (1 - 3 nm). As aggregation continues, the elongation and broadening
of these aggregates leads to the assembly of the vast networks of large fibrillar
structures. This period is characterized by a decrease in concentration of monomer,
and the various intermediates, as more of the protein is incorporated into the fibrils.
There is also evidence of significant lateral association of the larger, fully formed
fibrils.

43

(a) (b)

(c) (d)

(e) (f)

Figure 2.4. Atomic Force Microscope (AFM) images of 10mg/ml insulin
incubated at 60
o
C at pH 1.6 for (a) t = 0 minutes (b) t = 12 minutes (c) t =
35 minutes (d) t = 65 minutes (e) t = 120 minutes (f) t = 500 minutes
44

(a)

(b)

(c)

Figure 2.5. Height profiles of prefibrillar and fibrillar species in the
amyloidosis pathway for 10 mg/ml pure insulin after incubation at 60
o
C in
pH 1.6 solution for (a) 12 min (rod-like oligomer height ~ 1.5 nm), (b) 35
min (oligomer height ~ 8 nm), and (c) 500 min (fibril height ~ 8 - 10 nm)

45

The line height profile for three AFM images at different times along the
aggregation pathway are shown in figure 2.5. The earliest rod-like intermediates to
form have a z-height of ~1.5-3 nm, similar to the diameter of the insulin
monomer/dimer. However, after 35 minutes, a laterally aggregated species become
visible with a thickness (height) of ~ 8 nm, in spite of having length dimensions
comparable with the early protofibril. This thickness is also consistent with that of
the fully formed amyloid fibrils.
In the presence of the azoTAB surfactants, not only is the fibrillation rate
affected, but there are also morphological differences in the resultant aggregates. In
the presence of the trans isomer, the short rod-like protofibrils form earlier and in
larger numbers than in pure insulin. In figure 2.6, (a) and (b), there is almost no
monomer/dimer visible in the images, whereas after comparable times in the pure
protein sample (figure 2.3 (c) and (d)), much of the protein still exists as a small
globular species (~1.5 nm - 3 nm). This is likely a result of the strong protein-binding
tendency of the surfactant molecule, and its ability to unfold peptides, a prerequisite
for fibrillation in insulin. However, the protofilaments then appear to aggregate
laterally into bundles more readily than in the end-to-end fashion leading to a more
disordered, truncated and thicker final fibril morphology.

46

(a) (b)

(c) (d)

Figure 2.6. AFM images of insulin + azoTAB surfactants (a) trans isomer, t
= 30 minutes, (b) trans isomer, t = 60 minutes. (c) cis isomer, 12 minutes,
(d) cis isomer, 65 minutes. In the presence of the trans form of the
surfactant, short rod-like intermediates are observed although they show a
tendency to lateral aggregation. In the presence of the cis form, fibrils are
replaced by amorphous/globular aggregates.
47

Other insulin studies have reported similar behavior, possibly an effect of
solvent interactions or ionic strength.[36] A result of this preferred association is that
more of the protein has aggregated, or begun to aggregate by the time the very long
helical fibrils become visible in the AFM images. The cis isomer has a more
dramatic impact on aggregate structure. The characteristic rod-like aggregates are not
observed under these conditions, and instead the early assemblies have a more
globular structure, eventually yielding larger seemingly amorphous aggregates (figure
2.6), suggesting a more non-specific mode of interaction.
CD was used to investigate the changes in secondary structure associated
aggregation, and quantify differences in the protein conformation under the various
experimental conditions (figure 2.7). X-ray, nuclear magnetic resonance (NMR) and
FT-IR studies have shown that insulin’s native solution structure, which is essentially
the same as the crystalline structure, consists of 45-55% α-helical, 14-32% β-sheet,
8% β-turn, and 12-27% unordered structure.[46, 47, 48, 49] CD spectra of our
freshly prepared insulin solutions show a double minima at 208 and 222 nm
indicative of substantial α-helical structure (figure 2.4). The secondary structural
elements were calculated to be ~57% helical, ~15% β-sheet and ~ 26% random coil,
in fairly good agreement with reported values. Initial unfolding of the protein is seen
as a decrease in ellipticity at 222 nm, which can be explained by a loss in α-helical
content. This is accompanied by a shift of the 208 nm peak to lower values,
indicating an increase in unordered or unfolded structures which are characterized by
a single minimum at 200 nm.
48

Figure 2.7. Circular Dichroism spectra for (a) Pure insulin at 1h, 4h, 7h, 9h,
12h and 22h, (b) Insulin + 0.5mM azoTAB cis isomer at 1h, 4h, 7h, 9h, 12h
and 22h, (c) Pure insulin, Insulin + 0.5mM azoTAB cis isomer, and Insulin
+ 0.5mM azoTAB trans isomer at t = 1h, (d) Pure insulin at t = 22h, Insulin
+ 0.5mM azoTAB cis isomer at t = 22h, and Insulin + 0.5mM azoTAB
trans isomer at t = 4h (after extensive aggregation).

As fibrillation proceeds, the spectrum evolves into one typical of extensive β-
structured molecular organization, with a single minimum at 216 nm. This
corresponds to the formation of intermolecular β-sheets which stabilize the structure
of fibrillar species. For pure insulin this process took >12 hrs to occur. For insulin in
49

the presence of the azoTAB surfactant under visible light (trans isoform), the
conversion of the largely α-helical monomer to fibrils exhibiting strong B-character
happened much faster, taking <4hrs. By the end of the experiment, when the samples
were extensively aggregated, the spectrum for this trans-azoTAB sample was
essentially identical to that of the pure protein, suggesting that in spite of differences
in fibrillation rate, and potentially fibril morphology, the secondary structural
elements of the resultant aggregates are very similar. Spectra for insulin in the
presence of the cis isomer show deviations from those of the pure protein once
significant aggregation begins. After the initial unfolding of the protein, reducing the
α-helical content, the appearance of β-structures is delayed compared to pure insulin,
and even in largely aggregated cis-azoTAB samples, there is a lower β-sheet content,
and a higher contribution from random coil elements.
Small angle neutron scattering (SANS) allows a thorough investigation
of the structural changes associated with protein aggregation by simultaneously
collecting data over a wide range of particle solution sizes. Scattering at different Q
values indicates structures of different length scales, by the relationship L = 2π/Q,
where L is the approximate length scale of the scattering species. Our Q range of
0.007 - 0.3 Å
-1
probes real-space length scales on the order of L = 20 Å – 900 Å,
which is ideal for the analysis of prefibrillar structures. The lower protein
concentration used in these assays effectively extends the lag phase, allowing the
scattering of such prefibrillar intermediates to be better studied. The SANS data in
Figure 2.8 show that, initially, all samples exhibit scattering typical of the small
50

structures, as the curves plateau at high-Q values (~ 0.1 Å
-1
). This corresponds to
structures ~ 5 nm, consistent with the monomers, dimers or early nuclei. The weight-
average molecular weight of the protein can be estimated through the relationship

Where δ
s
and

δ
p
are the scattering length densities of the solvent and the protein
respectively, v is the protein-specific volume, and c is the protein concentration. At
these early times, the molecular weight of the scattering species was calculated to be
~ 10 kDa, slightly below the value expected for the dimer. As SANS is an additive
technique, this may indicate an equilibrium between the monomer and dimer.

Figure 2.8. SANS scattering curves for 1 mg/ml insulin, incubated 1.5h, 5h,
10h and 20h in a pH 1.6 solution with 0.1 M NaCl, in the absence and
presence of 0.5 mM cis and trans azoTAB surfactants.
51

For the sample containing the surfactant under visible light, the scattering
curve deviates from pure insulin after ~5hrs. At this time, the small structures give
way to fibrillar aggregates, which is seen as a shift in scattering to low-Q values,
associated with large structures. This conversion to larger structures progresses in the
trans-isomer system, with increased scattering at low-Q in the 10 hr sample. The
sharp bend in the curve near Q values ~ 0.02 - 0.025 Å
-1
corresponds to length scales
between 20 nm and 30 nm. The steep slope at values lower than Q = 0.02 Å
-1
is
consistent with the formation of large fibrils and fibril-fibril interactions.[50] 10 hours
after preparation of the samples, the pure protein, and protein in the presence of the
cis-isomer remain in the lag phase, however, after 24 hrs, all of the samples exhibit
scattering associated with fibril-fibril interactions (an I(Q) dependence of Q
-2
) which
lead to the formation of a loose 3D network of fibrils, similar to the semi-dilute
phases of flexible polymers solutions [12,50,51] (figure 2.8). Another method of
Small angle scattering analysis is the calculation of a pair distance distribution
function (PDDF)[52]

where P(r) is the probability of finding two scattering centers at a distance r apart, and
D
max
is the maximum dimension within the scattering species. A summary of the
results of a PDDF analysis of the SANS data are presented in the table below, and
selected systems are plotted in figures 2.9 and 2.10.

52

(a) (b)

Figure 2.9. PDDF for pure insulin (10 mg/ml) at t = 0.5h illustrating the
how R
g
and D
max
can be obtained from the plot. For a spherical scatterer, R
is the diameter. For a dimer, R describes the inter-molecular distance.
Adjacent table lists R
g
and D
max
for insulin-azoTAB surfactant systems.

Figure 2.10. PDDFs from (a) Insulin + 0.5 mM azoTAB (trans) after 20 hrs
of incubation, (b) Insulin + 0.5 mM azoTAB (cis) after 20 hrs of incubation

Time
(hrs)
R
g

(Å)
D
max
(Å)
Pure
Insulin
0.5 30.2 100
5 57.3 170
10 63.5 200
20 207.4 660
Insulin +
azoTAB
(Trans)
0.5 52 165
5 173 465
10 229.4 670
20 232.9 850
Insulin +
azoTAB
(cis)
0.5 50.6 165
5 66.5 220
20 195 670
53

Figure 2.11. PDDFs from the pure insulin assay after 20 hrs of incubation
and insulin + 0.5 mM azoTAB (trans) after 20 hrs of incubation, assuming
an extended rod-like scattering species. The D
max
in these graphs
corresponds to the diameter of the scattering rod, and the most probable
dimension, R, to the cross-sectional radius.

Over the course of the aggregation process, both D
max
and R
g
increase in all
systems, even during the lag phase. The Initial size of the species in the pure insulin
assay (~3 nm) is consistent with the insulin dimer. Both conformations of the
surfactant may induce the formation of early nuclei, as D
max
and R
g
are higher for the
samples containing azoTAB than in the pure protein, though the effect is more
pronounced in the trans sample, as seen in the AFM and CD data. The slightly higher
values for the surfactant assays may also be a reflection of the transient nature of the
monomer in those systems. For pure insulin, it is thought that the unfolding of the
monomer is largely a temperature mediated process. If the conformation of the
monomer is sufficiently sensitive to the binding of azoTAB molecules, then
surfactant induced unfolding and the subsequent formation of higher order structures
54

may make the monomer a very short lived species. By the 20 hour time point, the
aggregated cis sample shows a different pair distance distribution than the pure
protein or insulin in the presence of the trans isomer (figure 2.10). The curve for the
trans isomer system is characteristic of rod-like structures, where the first peak is
related to the cross-sectional radius, and the extended tail (highlighted by red inset
curve in figure 2.10a) is a result of the length of the structures. For the cis isomer, on
the other hand, there is a large peak in the ~ 300 Å range, which may indicate a less
fibrillar and more amorphous structure. In addition to the model independent
analysis, the PDDFs for the fully aggregated 20 hr samples were also calculated
assuming a rod-like scattering species. Here, the probability function, P(r), is
modeled as,

where
c
the characteristic thickness function. The D
max
in these graphs corresponds
to the diameter of the scattering rod, and the most probable dimension, R, to the cross
sectional radius. The fibril thickness calculated for the pure protein were consistent
with published values (8 – 10 nm), but were higher for the fibrils formed in the
presence of the azoTAB trans isomer (~ 12-15 nm). This may indicate a greater
degree of lateral aggregation.
Guinier analysis and calculation of the pair distance distribution functions can
provide information on the physical dimensions, general form and aggregation state
of the solution structures, however these analytical approaches are relatively low
resolution, and cannot provide information on atomic positions. Thus, shape
55

reconstruction of the SANS data was used to obtain high-resolution structural
information on the scattering species. The program GA_STRUCT produces 3-
dimensional models that approximate the structure of the scattering species by fitting
the experimental data. Starting with an initial system of randomly oriented scattering
centers, the program rearranges them using a genetic algorithm to achieve the best fit.
In this study, the data is fit 10 independent times, after which the program considers
the structural similarities of the models, determining the probability of a scattering
center existing at a given location and placing a given number of centers at the most
probable locations. Thus each structure represents a ‘consensus envelope’ of a family
of 10 ab initio fits. Figure 2.12 displays the shape-reconstructed structures obtained
early in the fibrillation process. The number of scattering centers was chosen to
approximate the number of atoms in the scattering species. A fit was attempted using
the number of atoms in the insulin monomer, but a better fit was achieved using the
value for the dimer. The experimental solution structures are in close agreement with
the crystal structure of the low-pH insulin dimer. At acidic pH, the shift in the
equilibrium from hexameric insulin to the dimer and monomer is accompanied only
by modest conformational changes. Much of the folding in the peptide chains is
preserved, including the C-terminal dimer-forming contacts of the B chain. However,
there is a significant difference in the orientation of the N-terminal residues of the B
chain (B1-B4), which extend away from the molecule, unlike in the compact native
dimer in which they are positioned closer to the molecular center, facilitating packing
in the hexamer. [106]
56

Figure 2.12. Structure of the low-pH insulin dimer. Shown from three
orthogonal views are the ribbon diagram and CPK space filling models
from X-ray crystallographic studies, along with the shape reconstructed
consensus envelope in blue, and best fit model (red insets). The consensus
envelope is based on 10 independent, ab initio runs.

The shape reconstructed dimer has a radius of gyration of 30.2 ± 0.172 Å and
a maximum dimension of 102 Å. This corresponds to the earliest nuclei visible in the
atomic force microscopy images, as well as the first species observable by dynamic
light scattering. The fact that the structures obtained from data collected at the
earliest time points indicate a dimeric rather that monomeric structure, is likely a
result of the solution environment. While acidic pH leads to dissociation of the
57

native hexamer, the low-pH protein structure and final morphology of the fibrils is
heavily dependent on the type of acid employed, though the complex interactions
responsible for the observed behavior are not yet fully understood. It has been shown
that sulfuric acid leads of the formation of shorter, more laterally aggregated and
bunched aggregates than the long, distinctive fibrils generally associated with
amyloidosis. Furthermore, the use of mineral acids such as HCl and HNO
3
result in a
predominantly dimeric low-pH structure, while acetic acid shifts the equilibrium
towards the monomer.[36, 106, 39] In this report, the acidic pH was achieved using
HCl, and aggregation was induced by elevated temperature. Under this experimental
scheme, dimers are initially favored, and destabilization of the small fraction of
monomeric protein results in the growth of characteristic, long, uniform amyloid
fibrils in pure protein assays (no surfactant).
Shape reconstruction of the samples that exhibit scattering associated with the
formation of early oligomers (Insulin + 0.5mM trans azoTAB 5hr, 10hr), indicate a
rod-like intermediate is formed along the aggregation pathway (figure 2.13). The
dimensions of this structure are in relatively good agreement with those of the
hypothetical extended hexamer discussed earlier in this report. It has been suggested
that this rod-like hexamer is the critical nucleus that initiates the rapid growth of
larger, extended aggregates and eventually, fibrils.[21]

58

Figure 2.13. Shape reconstruction of the early insulin rod-like oligomer.
Dimensions of the structure are consistent with a slightly compressed,
extended hexamer. Hypothetical ribbon diagram and space-filling model
illustrate potential monomeric arrangement in the pre-amyloid assembly.

It is likely that in this rod-like structure, the dimer interface formed by the C-
terminal end of the B chain of one monomer and the corresponding peptide section of
another monomer molecule is preserved. This structural feature is present in the
native ring-shaped hexamer, the low pH dimer, as well as insulin mutants designed to
be resistant to aggregation.[106] However, the interactions between these dimer
structures that leads to the rod-like hexamer rather that the native toroid hexamer are
unclear. For simplicity, the structure is often schematically represented as three
dimers arranged in an end-to-end fashion, or six monomers aligned along a single
axis.[21, 108] While this model is convenient, it is unclear what interactions would
stabilize such a structure. As discussed earlier, the most significant structural change
observed in going from the native state to the low pH conformation is the extension
59

of the N-terminal residues of the B chain away from the main body of the
molecule.[106] These unstructured sections are oriented anti-parallel to one another,
though they are too spatially separated to interact within a single dimer unit.
However, these exposed peptide segments may interact with corresponding segments
of other dimer units, stabilizing the structure of the rod-like intermediate, and perhaps
giving rise to the increased β-sheet character that is first observed in this structure,
and characteristic of amyloid species. The low pH conformation may also induce
steric hindrance that prevents the formation of a ring-like hexamer as in the native
state. In order for fibrillation to occur, the monomer must partially unfold and loosen
its tertiary contacts, however this is a reversible structural change associated with a
very modest volume expansion, perhaps an indication that the unfolding event does
not involve significant structural rearrangements.[39] Figure 2.9 includes 3
orthogonal views of the shape-reconstructed rod-like oligomer. The hypothetical
ribbon diagram and associated CPK space-filling model assume a similar mode of
interaction between the dimer units as is observed in the native hexamer, which is
dominated by hydrophobic interactions between the side chains of residues B25 Phe,
D25 Phe, A4 Glu and C4 Glu on adjacent dimer units. The dimensions of this
theoretical model agree well with those of the shape-reconstructed oligomer from the
SANS data.
The results of this study illustrate how small molecules can interact with pre-
amyloid peptides and interfere with the fibrillation process. The photo-responsive
azoTAB surfactants join a growing list of molecules identified as β amyloid
60

aggregation inhibitors, which have been shown disrupt the amyloidosis pathway in
different ways and to different extents. These compounds may inhibit the formation
of certain oligomeric assemblies without preventing fibrillation, inhibit the formation
of fibrils but not protein oligomerization, or inhibit both oligomerization and
fibrillation to some extent.[120] While the cis-isomer is seen to slow the rate of
aggregation, and perhaps inhibit amyloid fibrillation by forming more amorphous
aggregates, the mechanism of this action is unclear. The azoTAB surfactants are
capable of a number of modes of interaction with protein molecules, including polar
and electrostatic interactions with the charged head group, or hydrophobic
interactions with the carbon tail. It is possible that the surfactant interacts with those
portions of the modified monomer that are involved in protein-protein quaternary
contacts, disrupting their ordered association, perhaps through steric hindrance as a
result of the bent cis structure. Figure 2.14 compares the structure of the azoTAB
surfactant with a number of molecules that are known to interact strongly with
amyloid fibrils, interfering with their formation, and in some cases blocking
neurodegeneration in-vivo. All of the compounds in figure 2.14 are thought to
interact with the amyloid or pre-amyloid species in the cleft formed between protein
monomers. Molecular dynamics simulations suggest that the planar, cyclic structural
motifs common to these molecules are the source of their strong and relatively
specific interactions with fibrils.[121]

61

Figure 2.14. Structure of a number of molecules known to interact with amyloid
fibrils with high affinity and relative specificity. Presence of the molecules during
the aggregation process leads to structural disruption of amyloid fibrils, in some
cases blocking neurotoxicity.

It is thought that these portions of the molecule satisfy hydrophobic contacts and
interact with peptide side-chain carbonyl groups, while polar interactions between the
compound and the peptide destabilize inter-strand hydrogen bonds. Congo red, for
example, another azobenzene-based molecule, is known to interact with preformed
amyloid fibrils, and the birefringence of this dye is used as a laboratory test for the
presence of such fibrils. Both congo red and the azoTAB surfactant have benzene
rings with an azo-linkage. The dye is known to have binding sites in the vicinity of
the intermolecular β-sheets that form in amyloid species and it has been proposed that
it actually intercalates between the individual monomers. It has been shown that in
some fibrillation-prone protein systems, the presence of congo red during the
aggregation process blocks the formation of fibrils and inhibits neurotoxicity.[15, 53-
62

57] The diffuse plaques that are formed in these systems are not associated with
pathological degenerative changes in surrounding brain parenchyma, unlike the
compact fibrillar plaques observed in the disease state[58]. Similarly, 9,10-
Anthraquinone introduces structural perturbations which have a profound effect on
the ordering of oligomers, hindering β-sheet formation.[121] Thioflavin T is a
benzothiazole salt that is also widely used to test for the presence of or quantify
amyloid aggregates. This fluorescent dye exhibits a red-shift in its emission spectrum
upon binding to the β-rich intermolecular structures that are characteristic of amyloid
fibrils.[122]

2.5 Conclusions
Through dynamic light scattering, small angle neutron scattering, circular
dichroism and atomic force microscopy studies, we have demonstrated the ability to
control aspects of insulin fibrillation using photo-responsive azoTAB surfactants. The
trans form of the surfactant, obtained under visible light, accelerates the aggregation
process, while the cis form, stable under UV light, slows aggregation. AFM images
suggest that fibrils formed in the presence of the trans isomer show a greater
tendency to lateral aggregation than those formed in pure protein solutions, while the
cis isomer inhibits formation of the characteristic rod-like fibrillar structures, yielding
amorphous aggregates. The increased fibrillation rate in the presence of the more
hydrophobic trans isomer is likely a result of a greater degree of protein unfolding
due to enhanced interactions between the surfactant and the protein. This leads to
63

earlier formation of the unstable nuclei that trigger the fibrillation process. The
amorphous nature of the aggregation in the presence of the cis isomer may be due to
disruption of the specific protein-protein interactions required for fibrillation. Data
suggest that this effect interferes with the formation of protofibrils, some of the
earliest pre-fibrillar species identified, and perhaps the critical nuclei.[21] This study
illustrates the utility of small molecules such as the azoTAB photo-responsive
surfactants in the study of protein-based processes, through their ability to control of
protein secondary, tertiary and quaternary structure.

64

Chapter 3
Photocontrol of Catalase Quaternary Structure and
Activity
3.1 Abstract
The subunit dissociation of tetrameric catalase, regulated by photo-responsive
surfactants, is studied using small angle neutron scattering (SANS), dynamic light
scattering (DLS), circular dichroism (CD) and fluorescence spectroscopy, and related
to the enzyme’s biochemical activity. The azoTAB (azobenzene trimethylammonium
bromide) family of surfactants undergo a trans- to cis- conformational change on
exposure to ultraviolet light. The hydrophobic trans isomer (favored under visible
light) shows a greater tendency to bind with proteins. The ability to reversibly bind
the surfactant to the protein by changing light conditions allows for photocontrol of
various protein properties. In the presence of these surfactants, catalase dissociates
first into a super-active dimer, then at higher concentrations into an aggregation prone
monomer. Shape reconstruction is applied to the SANS data to give the solution
structure of the catalase tetramer and active dimer, and CD is used to track the
changes in secondary structure associated with subunit dissociation.

65

3.2 Introduction
Almost all biological processes require enzymes in order to occur at
significant rates. These proteins catalyze biochemical reactions by lowering their
activation energy.[64] As with all proteins, the three-dimensional folding of the
peptide backbone, and the correct formation of stable tertiary contacts in the enzyme
controls its function. The protein folding process is particularly important in
enzymes, as the specificity of binding and catalytic power of this class of molecule
depends on the unique geometry of the active site.[65] This part of the molecule
contains binding pockets for the substrate, and is the location of the catalytic reaction.
The ability to directly measure enzymes’ biochemical activity in vitro, and
simultaneously characterize their solution structure makes them ideal for studying the
form-function relationship in proteins, and their biological and industrial importance
makes control of these features very desirable. There are a number of environmental
factors that may cause conformational changes in the native enzyme structure and
activity, including pH, temperature, or the presence of chemical denaturants and
surfactants.[66-70] In this study we explore the effects of the photo-responsive,
azobenzene trimethylammonium bromide (azoTAB) family of surfactants on the
structure and activity of bovine catalase. The azoTAB surfactants have a hydrophilic
trimethylammonium head group and a hydrophobic alkyl tail separated by a photo-
responsive azobenzene group. The planar trans isomer, predominant under visible
light (433 nm), is more hydrophobic than the bent cis isomer, preferred under UV-
light (350 nm), a result of a lower dipole moment across the molecule (0.5 for trans
66

isomer, 3.1 for cis). Thus, the trans isomer binds to proteins more readily, and
induces a higher degree of protein unfolding than the cis isomer. Depending on the
enzyme’s secondary structure, interactions with photo-responsive azoTAB surfactants
may provide control over enzyme activity and binding specificity through photo-
reversible changes in secondary, tertiary and quaternary structure, essentially giving
us a photo-reversible switch for protein solution conformation.[59-63]
Catalase is a heme containing enzyme found in the peroxisomes of eukaryotic
cells that protects the cell from the toxic effects of hydrogen peroxide by catalyzing
its decomposition to oxygen and water. It is a vital enzyme occurring in almost all
aerobically respiring organisms.[22] In its native state, the protein exists as a
hydrophobically bound tetramer of identical polypeptide subunits, each over 500
residues long, with a molecular weight of approximately 60 kDa.[25] The molecule
has an isoelectric point of 5.4 making it negatively charged at neutral pH. In spite of
the presence of 16 cysteine residues, tetrameric catalase does not form any disulfide
bridges.[72] The heme moiety and active site are buried in a hydrophobic pocket at
the center of each monomer, ~ 20 Å below the molecular surface and the center of the
tetramer.
Each Catalase monomer is a four-domain protein; the first is a globular
domain, containing two α-helical arms involved in making quaternary contacts with
other subunits. The second is a β-barrel, consisting of two anti-parallel four-stranded
β-sheets, that interact with the heme group. This corresponds to the most
conformationally rigid region of the polypeptide chain.[71] The third is a largely
67

helical domain that contributes to the formation of a hydrophobic channel that
controls the accessibility of the active site. The fourth domain is referred to as the
wrapping domain, as it extends around the outside of each subunit. Though the
wrapping domain shows little discernible secondary structure, one segment of this
domain forms a short anti-parallel β-sheet with the corresponding structural element
on another monomer molecule, stabilizing the native quaternary structure.[22,73]
The instability of this oligomeric protein has been studied under a number of different
conditions, such as pH extremes and elevated concentrations of denaturants such as
urea, GdmCl or SDS surfactant. Native catalase is known to dissociate into a number
of observed protein assemblies in solution, including the highly ellipsoid dimer and
aggregation prone monomer.[26] While the monomer is considered biochemically
inactive, certain forms of the dimer have shown increased activity relative to the
tetramer.[25]

3.3 Experimental Methods
3.3.1 Materials
Catalase from bovine liver was purchased from Sigma (catalog number C40-
500) and used as received. Solutions were prepared in a pH 8.3 phosphate buffer. For
light and neutron scattering experiments, a final protein concentration of 1 mg/ml was
used. For fluorescence, activity and circular dichroism experiments lower
concentrations of 0.1 - 0.3 mg/ml were used.
68

The azoTAB surfactants were synthesized according to published methods,
via an azo-coupling reaction of phenol with 4-ethlyalaniline, then reactions with 1,2-
dibromoethane and trimethylamine (chemical structure illustrated in figure 3.1) [9,
11]. The trans isomer is more energetically favorable, and is obtained by
illumination with visible light (>75% conversion at 433 nm) or in the dark (~100%
conversion in the dark after 24hrs). Rotation of the azo linkage to the cis form was
achieved by illumination with UV light (>90% conversion to cis at 350 nm). Rapid
interconversion between the surfactant isomers was done with a 200W mercury arc
lamp (Oriel, model number 6283). For an incident wavelength of 365 nm (yielding
cis isoform), a combination of a 320 nm band-pass filter (Oriel model number 59800)
and a heat absorbing infra-red filter (Oriel model number 59060) was used. To revert
to the trans isomer, a long pass filter (400nm, Oriel model number 59472) was used.
The surfactant conversion was performed prior to preparation of protein solutions,
and surfactant concentrations from 0.1 mM to 10 mM were employed. All chemicals
were purchased from Aldrich at the highest purity.

3.3.2 Circular Dichroism
CD measurements were performed at the University of Southern California
Bio-imaging Center on a Jasco J815 instrument. Spectra were collected using a 1 mm
pathlength cuvette at a temperature of 25
o
C. Due to the azoTAB surfactant’s strong
ultraviolet absorption, a maximum surfactant concentration of 0.25 mM was used,
corresponding to final catalase concentration between 0.1 and 0.5 mg/ml. Data was
69

collected over an effective spectral range of 196 nm - 240 nm. Each spectrum was
averaged over four scans, and was background-subtracted and normalized to molar
ellipticity. The spectra were then analyzed using JFIT and CDFIT to quantify the
contributions of α-helical, β-sheet, and random coil secondary structural elements.
Data were collected for catalase in the presence and absence of azoTAB trans and cis
isomers at various concentrations.

3.3.3 Small Angle Neutron Scattering (SANS)
Small angle neutron scattering experiments were performed on the CG3-
BioSANS instrument at the High Flux Isotope Reactor facility (HFIR) at Oak Ridge
National Laboratories (ORNL) in Oakridge, TN. Samples were prepared in a D
2
0
based phosphate buffer to minimize incoherent scattering from hydrogen. A Q range
of 0.007 - 0.3 Å
-1
was probed by a neutron wavelength of 6 Å (cold neutrons), and
two detector distances of 1.3 m and 7 m were used. Variations in the response
efficiency of individual pixels in the 2 dimensional detector were accounted for by
considering the scattering from an isotropic scatterer with uniform scattering intensity
across the entire Q-range. The net scattering intensities of the empty cuvette and
buffer solution were subtracted, and the data was converted to an absolute differential
cross section per unit volume (in units of cm
-1
) using an attenuated empty beam. To
calculate the radii of gyration of scattering species, the Guinier approximation was
used.
,
70

where Q is the measurement vector in inverse space, related to the real-space
length scale by Q = 2π/L and having units of Å
-1
, I(Q) is the extrapolated intensity at
Q = 0 , and Rg is the radius of gyration.[44] Pair distance distribution functions
(PDDFs) were calculated using GNOM assuming a monodisperse system of
scatterers, providing information of the maximum dimension (D
max
) within the
protein or protein assembly by calculating the probability (P(r)) of finding two
scattering centers separated by a distance r, through the following equation. [109]

The PDDFs were calculated over a Q -range of 0.02-0.3 Å
-1
, in order to exclude
protein intermolecular interactions at low Q. The maximum particle dimension
(D
max
) was chosen in order to ensure a smooth return of the PDDF curve to zero at
D
max
. Shape reconstruction was performed with the program GA_STRUCT.[110]
The program starts with an initial system of randomly oriented scattering centers, and
uses a genetic algorithm to optimize the positions of the scattering centers until the
calculated data fit the experimental data. For the shape reconstruction analysis, the
data range from Q = 0.01 - 0.3 Å
-1
was considered in order to exclude intermolecular
interactions at low Q, and avoid length scales too small for protein continuity at high
Q. The number of scattering centers was chosen to roughly approximate the
theoretical number of atoms in the scattering species. Ten individual, independent
runs are preformed, and the reconstructed models are compared for similarity, and
averaged to give a consensus envelope.[110]
71

3.3.4 Dynamic Light Scattering
DLS measurements were performed on a Brookhaven BI-200SM instrument
(Brookhaven Instrument Corporation). The incident photons were 632.8 nm (outside
the range of wavelengths that lead to conversion of the surfactant), produced by a
35mW helium neon (HeNe) laser (Melles Griot, model Number 05-LHP-928). The
detector was at an angle of 90
o
to the incident beam and data was collected at a
temperature of 25
o
C. Analysis was done using the non-negative least squares (NNLS)
routine by a Brookhaven Instrument Corporation BI-9000AT digital correlator. The
hydrodynamic radius, R
H
, was calculated assuming a spherical shape, according to the
Stokes-Einstein equation,

where k
B
is Boltzmann’s constant, T is the temperature, n is the solvent viscosity, and
D is the experimentally determined diffusion coefficient.
Fresh protein solutions were prepared for each DLS trial, and prior to data
collection, they were filtered through a 200nm Anotop filter to remove large
structures which are strong scatters. UV/vis spectroscopy was used to confirm protein
concentration after filtration, and the illumination state could be changed in situ
through the use of a fiber-optic light guide.

72

3.3.5 Fluorescence Spectroscopy
Flourescence spectroscopy was performed on a Photon Technology
International spectroflourometer (model number QM-4). The flourescent dye, Nile
red (sigma catalog number N3013), was used as a micropolarity indicator to probe the
accessibility of the hydrophobic protein core. A small aliquot of a concentrated
solution of nile red in ethanol (1mM) was added to a 0.1 mg/ml catalase solution
giving a final nile red concentration of 0.4µM. Samples were stirred for 20 minutes
prior to data collection. For samples containing azoTAB, the surfactant was photo
converted prior to sample preparation to avoid potential unwanted effects of UV
illumination on nile red or catalase. Surfactant concentrations in the range 0 – 10 mM
were considered. Nile red was excited at 590 nm, with the emission monitored
between 600 and 800 nm (wavelengths which do not overlap with the absorption
profile of the surfactant). The hydrophobicity of the solution environment is related to
conformational changes in the enzyme which expose hydrophobic peptide segments,
thus, an increase in fluorescence intensity indicates unfolding of the native protein
structure.

3.3.6 Activity measurements
Amplex Red Catalase Assay Kit (Invitrogen catalog number A22180) was
used to determine the activity of catalase. In this approach, a catalase solution (4
U/ml) and H
2
O
2
solution (40 µM) are combined in equal quantities (25 µl each) and
allowed to react for 30 minutes, resulting in the breakdown of H
2
O
2
to water and
73

oxygen. Then a 100 µM solution of Amplex red reagent containing 0.4 U/ml horse
radish peroxidase (HRP) is added bringing the total volume to 100 µl. In the
presence of HRP, the unreacted H
2
O
2
reacts with Amplex Red creating the
fluorescent molecule resorufin. The sample is then excited at 530 nm, with the
emission monitored between 550 nm and 800 nm. The slope of the plot of the
fluorescence emission versus time is then a measure of the relative activity of the
enzyme. For these assays, a maximum azoTAB surfactant concentration of 2 mM
was used because at concentrations higher than this the surfactant begins to have an
effect on HRP, potentially affecting the accuracy of the activity measurement.

3.4 Results and Discussion
Figure 3.1 shows nile red fluorescence as a function azoTAB surfactant
concentration. The sharp increase in fluorescence intensity between 3 mM and 5 mM
in the presence of the azoTAB trans isomer suggests that significant changes in
native conformation are occurring in this experimental regime. In the presence of the
cis isomer, this extensive unfolding is not observed with increasing azoTAB
concentration. This is in agreement with the theory that the trans isomer is more
likely to bind to, and unfold native protein structures due to its relatively hydrophobic
nature compared to the cis.

74

Figure 3.1. Nile red fluorescence as a function of azoTAB concentration,
for the trans and cis isomer. The trans conformation has a marked effect
on the protein structure as seen by the sharp increase in nile red intensity
(hence, increase in hydrophobicity).

Dynamic light scattering was used to better understand the nature of the
changes in solution structure due to interactions with azoTAB surfactants, and
activity measurements were made to determine how these conformational changes
affect catalase biochemical function. The DLS data (figure 3.2) for pure catalase is in
fairly good agreement with published values for the hydrodynamic diameter of the
tetramer.[77,78] However, in the presence of azoTAB surfactants, the data suggest
that the tetramer dissociates into smaller species. At low surfactant concentrations,
the tetramer gives way to a ~ 6 nm oligomer, likely the catalase dimer. At higher
surfactant concentrations, the dimer dissociates further into the monomer, though size
determination of monomeric catalase is difficult due to its propensity to aggregate.
75

Increasing azoTAB concentration beyond this point appears to accelerate aggregation,
likely a result of continued unfolding the monomer destabilizing it further.

Figure 3.2. Hydrodynamic diameter of solution structures in catalase-
azoTAB system. Inset diagrams illustrate possible protein assemblies
associated with each distribution profile

76

Though dissociation of the native quaternary structure is seen in the presence
of the cis isomer, the effect is reduced compared to the trans isomer at the same
surfactant concentration. This provides the potential to reversible switch between the
various solution states available to catalase by simply altering the illumination state.
An interesting feature of the DLS data (figure 3.2) is that the important changes in
catalase quaternary structure appear to take place at concentrations significantly lower
than the dramatic conformational changes observed as an increase in nile red
fluorescence intensity (figure 3.1). This suggests that the increase in hydrophobicity
seen at azoTAB trans isomer concentrations above 3 mM is due to progressive
unfolding of the monomer subsequent to subunit dissociation.
The activity assays (figure 3.3) show that, relative to native catalase, the dimer
obtained at lower azoTAB concentrations exhibits super-activity (120%). Activity
enhancements of a similar magnitude have been reported for catalase dimers attained
by denaturation of the enzyme in <0.3 M GdmCl.[79,80] At higher azoTAB
concentrations, the activity is seen to decrease again, but the sharp drop-off in activity
reported in the presence of >0.3 M GdmCl is not observed in the azoTAB
concentration range studied. It is worth noting that the accessible azoTAB
concentration range for activity measurements is below the concentration at which
nile red signals large scale conformational changes (figure 3.1). It is possible that
these dramatic structural shifts would be linked with the sharp loss in activity seen in
systems in which catalase is denatured by other approaches.

77

Figure 3.3. Catalase (4 U/ml ) activity in the presence of azoTAB
surfactants. The increase in activity suggests the presence of a super-active
dimer.

Small angle neutron scattering was used to generate high-resolution structural
data on the catalase-azoTAB system. Figure 3.4 shows the SANS data for catalase,
and catalase in the presence of azoTAB surfactants at various concentrations and
illumination states. The weight-average molecular weight of the protein can be
estimated through the relationship

Where δ
s
and

δ
p
are the scattering length densities of the solvent and the protein
respectively, v is the protein-specific volume, and c is the protein concentration. For
the pure protein, the molecular weight was estimated to be ~200 kDa. The fact that it
78

is slightly lower than the value expected for the tetramer (~240 kDa) is likely a result
of dissociation of some fraction of the tetrameric catalase into dimers and monomers
that has been observed under physiological conditions and in-vitro preparations of the
protein.[118] At low surfactant concentration (0.2 mM), the scattering shows reduced
intensity (indicating a transition to smaller structures in solution) with an I(0) roughly
half that of the pure enzyme (~100 kDa). This is consistent with dissociation of the
tetramer to the dimer.

Figure 3.4. SANS scattering data for the catalase - azoTAB system at
various concentrations and after various incubation times.

79

When the surfactant concentration is increased (0.5 mM), although the
scattering at high Q remains lower than that for pure catalase, there is a sharp increase
in scattering at low-Q, likely indicative of the formation of aggregates. At the highest
surfactant concentration considered (1 mM), there is a considerable shift in scattering
to lower Q values. The steep slope at low Q, associated with extensive aggregation,
is accompanied by a decrease in high Q intensity, suggesting that the concentration of
tetrameric and dimeric catalase is very low, and that much of the protein exists in the
form of aggregates. The scattering curve for the sample containing the 1 mM
azoTAB in the cis form is very similar to that of the 0.5 mM trans sample, illustrating
the trans isomer’s enhanced ability to manipulate protein structure, as observed in a
number of previously studied systems.[12, 81, 82]
Pair distance distribution functions (PDDFs) were calculated for the different
experimental assays. PDDFs are a model-independent approach to determining the
probability, (P(r)), or finding two scattering centers separated by a distance r.

The GNOM program [83] was used to calculate the PDDFs over a Q range of
0.007-0.3 Å
-1
. The maximum particle diameter, D
max
, was selected to be the lowest
value that gave a smooth return of the function to zero.
80

0
10
20
30
40
50
60
70
80
0 20 40 60 80 100 120 140
r (A)
P(r)
Pure catalase
Catalase + 0.2 mM azoTAB

Figure 3.5. Pair distance distribution functions for tetrameric and likely
dimeric catalase

The radius of gyration for the pure protein was found to be 39.6 A, in excellent
agreement with published values. In figure 3.5, the 0.2 mM azoTAB sample shows a
decrease in R
g
, as well as a decrease in D
max
, the maximum dimension within the
molecule compared to the pure protein, consistent with subunit dissociation. The
calculated R
g
= 30.2 is slightly higher than expected for the dimer, and may indicate
the presence of an equilibrium with the tetramer.
One of the advantages of neutron scattering techniques is that the incident
neutrons, which have no electric field, do not interact with the electron cloud
associated with large protein molecules and are instead scattered by the atomic nuclei,
potentially providing high resolution structural information on atomic position. In
order to determine the size and general molecular profile of the species present in our
81

system under the various experimental conditions, shape reconstruction analysis of
the SANS data was performed with the program GA_STRUCT.[110] Starting with
an initial system of n randomly oriented scattering centers, the program calculates the
theoretical scattering pattern, and compares the synthetic data with the experimental
data. The position of each scattering center is then iteratively optimized to achieve
the best fit with the experimental data. Figure 3.6 displays the results of the shape
reconstruction analysis of pure catalase in buffer (ie. native catalase tetramer
structure), along with the ribbon diagram and CPK space-filling model from X-ray
crystallography.[75]

Figure 3.6. Structure of bovine liver catalase native tetramer. X-ray
crystallography based ribbon diagram and CPK space-filling model [PDB
7CAT] are displayed along with the shape reconstructed model from SANS
data. The consensus envelope, in blue, is based on 10 independent runs.
The best fit individual model is shown in the red inset.
82

The shape and dimensions of the tetramer agree well with the crystal structure, with a
maximum dimension of 130 Å, and a radius of gyration of 39.6 Å. The best-fit model
is illustrated in the red inset diagrams in figure 3.6.
Figure 3.7 displays the shape-reconstructed model of the dimer, obtained in
the presence of 0.2 mM azoTAB surfactant. The model has a maximum dimension of
100 Å, and consists of two lobes of about equal size, consistent with the theoretical
dimer structure. The radius of gyration R
g
= 27.2 ± 0.5 Å. For higher surfactant
concentrations (1.0 mM azoTAB), the shape reconstruction was carried out on the
data range Q = 0.015 Å
-1
– 0.3 Å in order to exclude intermolecular interactions
leading to aggregation and increased scattering at low Q values. The model obtained
from this fit produces a lopsided dimer-like structure, in which one of the lobes is
larger than the other. Since SANS is an additive technique, in systems where a
number of well-defined protein conformations or assemblies are present at the same
time in solution, the shape-reconstructed model is often a type of weighted
superimposition of the possible solution conformations.[116] The non-symmetrical
dimer in figure 3.7 may then indicate that in this system, though the dimer is still
present, it has begun to give way to the monomer.
83

Figure 3.7. Shape reconstruction of catalase in the presence of azoTAB
surfactants. The dimer is formed at a surfactant concentration of 0.2 mM.
At the higher concentration (1.0 mM), the lopsided dimeric structure may
indicate the existence of monomeric catalase. X-ray crystallography based
ribbon diagram and CPK space-filling model [PDB 7CAT] illustrate how
separation along different planes in the tetramer results in different dimer
structures. The consensus envelopes in these diagrams are based on 10
independent runs.

84

While the presence of active and inactive forms of the dimer have been
reported during denaturation and subunit dissociation induced by guanidinium
chloride, urea or pH destabilization, it is unclear along which plane the tetramer
separates, and which contacts in quaternary structure are lost to give way to the
dimer.[25, 26] The use of a surfactant denaturant such as azoTAB to induce
dissociation and relaxation of the native state may involve an unfolding pathway not
present during denaturation from more chaotropic agents such as guanidinium
chloride. At concentrations significantly below the critical micelle concentration
(CMC), surfactants may have preferred, and often relatively specific binding sites on
proteins depending on the protein’s structure.[119] While shape reconstruction can
confirm the existence of the dimer as a solution quaternary structure under certain
solution conditions, it cannot give insights to the changes in secondary structure
associated with the shift of equilibrium.
Circular dichroism was used to analyze the effect of azoTAB surfactants on
the secondary structural elements of catalase.(figure 3.8) Native catalase consists of
26% α-helices and 12% β-structures, while the remainder consists of irregular
structures and extended sections which are involved in stabilizing the tetrameric
assembly.[72] The CD spectra for catalase in the presence of azoTAB illustrates the
tendency of the surfactant to preferentially attack α-helical structures.
85

Figure 3.8. Circular Dichroism spectra of catalase in the presence of
various azoTAB concentrations and illumination states. An unfolding of α-
helical secondary structures is observed.

There is a progressive loss in the α-helical character of the enzyme with
increasing surfactant concentration. This is seen as a decrease in intensity of the peak
at 222 nm accompanied by a shift of the 208 nm peak to lower values (characteristic
of the formation of random coil structures). The α-helical content fell from ~ 25 % in
the pure protein to < 14 % at the highest azoTAB concentration investigated (1 mM).
Again, this effect is enhanced for the trans isomer with respect to the cis. Although
there is considerable unfolding of the secondary structural elements with increasing
concentrations of azoTAB surfactants, the subunit dissociation occurs at relatively
low surfactant concentrations (0.2 mM azoTAB). The CD data show that at this low
concentration, only modest changes in secondary structure have taken place, and so
86

suggest that the separation of the tetramer into dimer units happens before significant
unfolding of the protein. The first change to occur at low azoTAB concentrations is a
slight loss in α-helical content (~ 2 %). Although circular dichroism cannot provide
any insight into the site of unfolding, by considering the structure of catalase, this
provides a possible mechanism for the action of the surfactant. Contacts between
catalase subunits in the tetramer involve contributions from a largely helical arm of
the globular domain spanning 75 residues at the N-terminus of the molecule.
Relaxation of the secondary structure in this region may disrupt the subunits’ ability
to maintain these contacts inducing dissociation of the tetramer. The tetrameric
assembly is additionally stabilized by hydrophobic interactions between irregular and
extended peptide chain segments in the monomer units;[72] dissociation to the dimer
may be energetically facilitated by replacing these protein-protein interactions with
protein-surfactant interactions. Another potential target for azoTAB binding is the
domain spanning residues 321- 436, which contains two helical sections, and creates
a ‘funnel’ to the catalytic core of the enzyme. Secondary structural changes to this
domain could mediate the accessibility of the active site to the ligand, affecting
activity. Furthermore, the fact that the active site is buried in the protected, β-rich
core of the molecule, may account for the retention of activity after subunit
dissociation. Given the wide variety of interactions that these surfactants are able to
take part in, as well as the sheer number of potential binding sites on the surface of
large proteins like the catalase tetramer, their effect in solution is likely the result of a
number of overt and subtle interactions. In this system, the azoTAB surfactants are
87

used to demonstrate the utility in selective binding of small molecules to proteins and
protein assemblies, yielding control over biological activity.

3.5 Conclusion
The dissociation of multimeric catalase may be regulated by binding of the
azoTAB molecule. In the presence of these surfactants, the tetramer dissociates into
smaller units. What is thought to be an active dimer is observed at low surfactant
concentration, which gives way to an aggregation prone monomer at higher
concentrations. This effect is more pronounced in the presence of the trans isomer
than the cis, presumably a result of the increased hydrophobicity of that form of the
surfactant. Small angle neutron scattering determined that for the tetramer, R
g
= 39.6
Å, and D
max
= 130 Å, and for the active dimer, R
g
= 27.2 Å, and D
max
= 100 Å.
Circular dichroism revealed that while increased surfactant concentrations lead to
progressive unfolding of the protein, the subunit dissociation occurs before significant
changes in secondary structure take place. The ability to photo-isomerize the
surfactant in solution provides the possibility to make this control of structure and
activity reversible by changing illumination conditions, though we are yet to observe
reversibility in this system.

88

Chapter 4
Structural Changes Papain induced by Photo-
responsive Surfactants
4.1 Abstract
Papain is a cysteine-protease expressed in the latex of carica papaya, which
catalyzes the hydrolysis of peptide, amide, ester, thiol ester and thiono ester bonds.
The protein is folded into two domains of similar size and generally contiguous
peptide segments. The active site rests in the cleft between the two domains. The
denaturation of papain is thought to be a biphasic process in which the domains
unfold independently due to differences in secondary structure and stability. We have
used small angle neutron scattering (SANS) to track changes in the protein’s solution
structure associated with surfactant induced unfolding, and circular dichroism (CD) to
consider the secondary structural aspects of these changes. It was found that binding
of azoTAB surfactants leads to a progressive loss in papain secondary structure,
primarily in α-helical content. Pair distance distribution functions calculated from
SANS data show an increase in the radius of gyration from R
g
= 19.5 Å for the native
protein to R
g
=26.3 Å at a protein:surfactant ratio of 1:25. Shape reconstruction was
applied to the SANS data to determine the non-native solution structures of the
partially folded papain intermediates. The data suggest that the denaturation process
involves a separation of the domains along with the losses in secondary structure. The
89

azoTAB surfactants undergo a trans- to cis- conformational change on exposure to
ultraviolet light. The more hydrophobic trans isomer (favored under visible light)
shows a greater tendency to bind with proteins. At low surfactant concentrations, the
ability to reversibly control the extent of binding of the surfactant to the protein by
changing light conditions can be used to reversibly control the extent of unfolding.

4.2 Introduction
A common protein structural motif involves the formation of multiple
relatively rigid domains connected by usually flexible linkers. Each domain is
constructed of often non-contiguous segments of the polypeptide chain which fold
into distinct, generally globular units.[83] For these proteins, the flexibility of the
domains with respect to one another, and resultant motions of the molecule are often
important for the proper biological function of the protein.[116] A result of
structurally independent domains is that each domain may have a different stability,
and tendency to unfold. In the study of denaturation, it has been shown that some
proteins such as α-lactalbumin, form folding intermediates which maintain a
substantial degree of secondary structure while the tertiary structure is almost
completely disrupted.[84, 85] On the other hand, many multi-domain proteins may
follow folding or unfolding pathways in which the individual domains’ secondary and
tertiary structures are altered seemingly independently. In order to study the effect of
azoTAB surfactants on domain structure, as well as better understand how domain
structure may affect activity, we have applied it to the analysis of the enzyme papain.
90

Papain is a cysteine protease expressed in the latex of the papaya tree (carica
papaya), where it plays and important role in the defense of the plant against
lepidopteran larvae and polyphagous pests.[111] Papain is considered to be the
archetype of the cysteine protease family of enzymes, which catalyze the hydrolysis
of peptide, amide and ester bonds.[112] The papain molecule contains a total of 212
residues. One interesting biophysical feature of the protein is its two-domain
structure. Papain is comprised of a single polypeptide chain that is folded into two,
distinct, non-symmetrical domains, with different secondary structural elements, but
with similar volumes and masses. Each domain is formed from largely contiguous
sections of the peptide backbone, with the exception of the 10 terminal residues at
both the C- and N-terminals, which are each buried in the complementary domain.
Thus, Domain 1 spans residues 10 to 111, and also includes residues 208 through
212, while Domain 2 spans residues 112 to 207, and includes N-terminal residues 1
through 9 (figure 4.1).

Figure 4.1. Structure of papain protein. Domain 1 is primarily helical,
whereas domain 2 is characterized by a strong β-structured content. [PDB
1PPP]
91

Domain 1 has a secondary structure characterized by 3 α-helices. The first,
and longest, is formed by residues 24 through 42, and includes the catalytically
relevant Cys25 and its associated sulfhydryl group. This helix is imperfect due to
hydrogen bond disruption and distortion, creating polar and non-polar sides of the
helix, which are involved in stabilizing the domain-domain interface and contribute to
forming the hydrophobic core of domain 1. The second is a short helix, spanning
residues 50 to 57, and is largely buried. The third helix (residues 67-78), oriented
perpendicular to the second, exhibits partial 3
10
character near its N-terminus, and has
one side that is exposed to solvent. Domain 2 is characterized by its anti-parallel β-
sheet structure, but also contains two short α-helical sections on opposite ends of the
surface of the domain, comprised of residues 117 to 127 and 138 to 143. The β-
structured residues are buried in the interior of this domain. The central β-strand
includes residues 158 to 167, and has an elbow of around 120
o
. The main chain atoms
of the other strands form a barrel like cavity in the core of the domain, which is
occupied by hydrophobic side chains.
The catalytic action of the enzyme is due to the thiol group of the Cys25
residue on domain 1, which forms a catalytic dyad with a histadine residue (His159)
from domain 2, deprotonating it and allowing nucleophilic attack of the substrate
carbonyl group by the cysteine’s anionic sulfur. The molecule’s remaining six
cysteine residues are involved in forming disulfide bridges.[86] This protein
represents an interesting target for the azoTAB surfactants. The denaturation process
in papain is thought to be biphasic, suggesting the domains may be capable of folding
92

or adopting non-native conformations independently of one another.[83] The
tendency of azoTAB surfactants to favor interactions with α-helical structures, or to
have preferred binding sites on a molecule, means that it may be possible to
structurally alter one domain, while preserving tertiary contacts in the other domain.
In this study, we use small angle neutron scattering to study the structural changes
associated with papain’s interaction with photo-responsive azoTAB surfactants, and
circular dichroism to track changes in the protein’s the secondary structure. The
results suggest that the unfolding of the protein in the presence of azoTAB surfactants
involves an increase in the radius of gyration, and a partial spatial separation of the
domain units. The extent of unfolding increases with an increase in surfactant
concentration and is more pronounced in the presence of the trans isomer than the cis
isomer. The changes in secondary structure associated with this unfolding primarily
involve a loss in α-helical content and an increase in random coil character.

4.3 Experimental Methods and Materials
Papain protein from carica papaya was purchased from Sigma Aldrich
(catalog number P4762) and used as received. Sample solutions were prepared with
final protein concentrations between 1 and 3 mg/ml for small angle neutron scattering
experiments, and 0.1 and 1 mg/ml for circular dichroism assays. The azoTAB
surfactants were synthesized via an azo-coupling reaction of phenol with 4-
ethlyalaniline, then reactions with 1,2-dibromoethane and trimethylamine, according
to published methods [9, 11]. The trans isomer is obtained under illumination with
93

visible light (>75% conversion at 433 nm) or in the dark (~100% conversion in the
dark after 24hrs), and is the more energetically favorable isomer. Conversion to the
cis isoform is achieved by illumination with UV light (>90% conversion to cis at 350
nm). Rapid interconversion between the surfactant isomers was done with a 200W
mercury arc lamp (Oriel, model number 6283). For an incident wavelength of 365 nm
(yielding cis isoform), a combination of a 320 nm band-pass filter (Oriel model
number 59800) and a heat absorbing infra-red filter (Oriel model number 59060) was
used. To revert to the trans isomer, a long pass filter (400nm, Oriel model number
59472) was used. The surfactant conversion was performed prior to preparation of
protein solutions, and surfactant concentrations from 0.01 mM to 1 mM were
employed. All chemicals were purchased from Aldrich at the highest purity.

4.3.1 Small Angle Neutron Scattering
Small angle neutron scattering samples were prepared with a final protein
concentration of 1 mg/ml in a D
2
O based buffer to minimize incoherent scattering
from hydrogen. Data were collected on the CG3-BioSANS instrument at the High
Flux Isotope Reactor facility (HFIR) at Oak Ridge National Laboratories (ORNL).
The wavelength of the incident beam of cold neutrons was 6 Å, and the two detector
distances of 1.3 and 7 meters yielded a Q range of 0.007 - 0.3 Å
-1
. Variations in the
response efficiency of individual pixels in the 2 dimensional detector were accounted
for by considering the scattering from an isotropic scatterer with uniform scattering
intensity across the entire Q-range, and converted to an absolute differential cross
94

section per unit sample volume (in units of cm
-1
) using an attenuated empty beam.
The net intensities were corrected for the background (D
2
O based buffer) and empty
cuvette.
A number of data analysis approaches are available when considering SANS
data. The three complementary techniques employed in this study are the Guinier
approximation, calculation of pair distance distribution functions (PDDFs), and shape
reconstruction. Guinier analysis gives an estimate of the radius of gyration of the
scattering species through the following approximation,

where Q is the measurement vector in inverse space, related to the real-space
length scale by Q = 2π/L and having units of Å
-1
, I(Q) is the scattering intensity, I(0)
is the extrapolated intensity at Q = 0 and R
g
is radius of gyration, valid for globular
species.[44] Pair distance distribution functions (PDDFs) were calculated using
GNOM,[109] which calculates the probability (P(r)) of finding two scattering centers
separated by a distance r, through the following equation, where D
max
is the
maximum dimension within the protein.

The PDDFs were calculated over a Q-range of 0.02 - 0.3 Å
-1
, in order to exclude
protein intermolecular interactions at low Q. For these calculations, the system is
assumed to be monodisperse, and the maximum particle dimension (D
max
) was chosen
in order to ensure a smooth return of the PDDF curve to zero at D
max
. Shape
95

reconstruction was performed with the program GA_STRUCT.[110] Starting with an
system of randomly oriented scattering centers, the program uses a genetic algorithm
to optimize the positions of the centers until synthetic scattering data fit the
experimental data. Only the data range Q = 0.01-0.3 Å
-1
was considered in order to
exclude intermolecular interactions at low Q, and avoid length scales too small for
protein continuity at high Q. The number of scattering centers was chosen to roughly
approximate the theoretical number of atoms in the scattering species. Ten ab initio
runs are preformed, and the reconstructed models are compared for similarity, and
averaged to give a consensus envelope.[110]

4.3.2 Circular Dichroism
Each CD spectrum was obtained from the average of four scans, and was
background-subtracted and normalized to molar ellipticity. The resulting spectra were
then analyzed using the programs JFIT and CDFIT to quantify the contributions of α-
helical, β-sheet, and random coil secondary structural elements. Due to the azoTAB
surfactant’s strong ultraviolet absorption, a maximum surfactant concentration of
0.5mM was used, and the effective spectral range was 196 nm - 240 nm. Protein
concentration were decreased accordingly to preserve the protein:surfactant mole
ratio. Data were collected for the protein in the presence and absence of azoTAB
trans and cis isomers.
CD measurements were performed at the University of Southern California
Bio-imaging Center on a Jasco J815 instrument. Solutions were prepared with final
96

protein concentrations between 0.1 and 1 mg/ml, and the spectra were collected using
a 1 mm pathlength cuvette at 25
o
C.

4.4 Results and Discussion
Circular dichroism is an excellent method for rapidly evaluating the secondary
structure, conformation or folding of proteins in solution. For pure papain at neutral
pH, the secondary structural elements were determined to be 25% α-helix, 16% β-
sheet, and 59% random coil, in good agreement with published values.[87, 88, 115]
The high content of irregularly structured main chain sections is seen in the CD data
(figure 4.2) as the peak below 200 nm, characteristic of random coil structures. The
double minima at 222 nm and 208 nm are consistent with α-helical elements.

Figure 4.2 CD spectra of (1mg/ml) papain secondary structural changes
induced by photo-surfactants
97

In the presence of azoTAB surfactants, a progressive unfolding of the α-
helical regions of the protein is observed as a decrease in the intensity of the 222 nm
peak, and an increase in the contribution of irregular structures to the spectra. While
there is a substantial decrease in the helical character of the protein, the β-sheet
structures are not as readily disrupted. The α-helical content is seen to decrease from
~ 25 % to ~ 12 % at a surfactant concentration of 1 mM, while the decrease in β-
structures is < 4 %. While the surfactant exhibits a higher binding affinity for α-
helices than β-structures, the dramatic unfolding of the α-helical structures relative to
the β-sheets in papain is likely also a result of the protein’s tertiary structure. The
small β-rich portion of domain 2 is protected from attack from the surfactant by an
outer shell of unstructured and α-helical segments on one side, and its proximity to
domain 1 on the other. Presumably, significant conformational changes would be
required before this portion of the molecule becomes easily accessible to the solvent.
SANS data (figure 4.2) shows an increase in scattering intensity across the
entire Q range, indicating an increase in the size of solution structures. At high-Q
values, this is consistent with unfolding of the protein monomer. The increase in low-
Q scattering, as well as the increase in the slope at low-Q may be an effect of protein
self-association subsequent to structural destabilization.

98

Figure 4.3. Papain SANS scattering curves of 1 mg/ml papain solution

Guinier analysis was used to calculate the radius of gyration through the
relationship

where Q is the measurement vector in inverse space, related to the real-space length
scale by Q = 2π/L and having units of Å
-1
, I(Q) is the extrapolated intensity at Q = 0,
and R
g
is the radius of gyration.[44] The data show that as the concentration of
surfactant is increased, so the radius of gyration increases significantly, from ~21 Å
to over 26 Å. This suggests there is swelling or unfolding of some segments of the
protein associated with the loss in helical structure observed in the CD data.
99

Pair distance distribution functions (PDDFs) were calculated from the SANS
scattering data for pure papain, and papain in the presence of azoTAB surfactants (0.5
mM and 1 mM trans surfactant), and are plotted in figure 4.4.

Figure 4.4. Model independent PDDFs calculated for pure papain, and
papain in the presence of 0.5 mM and 1.0 mM azoTAB. An increase in the
most probable distance, r, is accompanied by a shift of D
max
to higher
values.

The most common dimension (corresponding to the radius of a spherical
scattering species), r = 19 ± 1.5 Å for pure papain agrees with the Guinier analysis
and literature values. The maximum dimension within the protein, D
max
was found to
be 48 Å. As the concentration of azoTAB increases, the value of most common
dimension increases along with the maximum diameter. In the presence of 0.5 mM
trans azoTAB the most probable separation of scattering centers increases to 23.1 ±
1.6 Å, and for 1 mM azoTAB, r = 26.2 ± 1.9 Å. D
max
also increases substantially in
100

the presence of the surfactant. For the 0.5 mM trans surfactant assay, D
max
= 55 Å,
and for 1 mM trans azoTAB D
max
= 62 Å. The slight shoulder at higher r values for
the samples containing azoTAB, as well as the substantial increase in D
max
compared
to the increase in r may indicate an extension of the molecule along one of its axes.
Shape reconstruction analysis was applied to the SANS data using the
program GA_STRUCT, in order to obtain high-resolution structural information on
the protein during this unfolding event.

Figure 4.5. Solution structure of native papain. The ribbon diagram and
CPK space filling models from X-ray crystallographic studies [PDB 9PAP],
along with the shape reconstructed consensus envelope in blue, and best fit
model (red insets) are shown from three orthogonal views. The consensus
envelope is based on 10 independent, ab initio runs
Figure 4.5 shows the shape-reconstructed model of the native papain molecule (no
surfactant, neutral pH buffer) from three orthogonal perspectives, along with the
101

ribbon and CPK space filling models obtained through X-ray crystallography.[PDB
9PAP] The consensus envelope is displayed in blue, while the individual model with
the best fit to the experimental data in displayed as an inset in red. The consensus
envelope structure agrees well with the dimensions and molecular shape of the X-ray
crystallographic structure. The number of scattering centers in the model is chosen to
approximate the number of atoms in a single papain molecule (~1860), and ten
independent runs are performed. The program then compares the individual models
from the runs and generates a consensus envelope by placing the 1860 scattering
centers in the optimal positions.
Figure 4.6 shows the results of the shape reconstruction analysis of samples
containing azoTAB surfactants. The advantage of shape reconstruction over
traditional approaches, such as X-ray crystallography, is seen in these diagrams,
which display the solution structure of non-native, partially unfolded papain in the
presence of azoTAB surfactants. The consensus envelope as well as the best-fit
model are included for pure papain, and papain in the presence of trans azoTAB at
two concentrations. From these models, the unfolding of the protein due to
interactions with azoTAB appears to involve a separation of the two domains. At 0.5
mM azoTAB, the molecule is slightly longer than the native protein along the axis
perpendicular to the domain-domain interface, and the cleft between the two domains
is more pronounced. At the higher surfactant concentration of 1 mM, there is a
prominent separation of the domains, and it appears the inter-domain contacts are
102

loosened, as the two appear linked by a thin neck, as opposed to the tight interface
present in the native state.

Figure 4.6. azoTAB surfactant induced unfolding of papain tracked by
SANS shape reconstruction. The shape reconstructed consensus envelope
(in blue) and best fit model (red insets) are shown from three orthogonal
views for three different surfactant conditions.

In the native protein, three covalent bonds span the interface between the two
domains. Each domain is a generally contiguous peptide segment, with the two
domains divided between residues 111 and 112. However, the ten terminal residues
on each end of the peptide chain cross the interface, becoming part of the opposite
domain, and creating stabilizing contacts across the interface.[86] The narrow linker
between the domains seen at higher surfactant concentrations may indicate a loss in
the inter-domain contacts as tertiary associations are disrupted.
103

Correlating these results with the circular dichroism data, it is clear that the
observed unfolding events are associated with changes in secondary structure, and are
not due to losses in tertiary structure alone. The CD data indicate that as azoTAB
concentration increases, α-helical content is lost at the greatest rate, transitioning to
unstructured random coil. The modest loss in β-structured peptide suggests that the
core of domain 2, characterized by its anti-parallel β-sheet structure, is still intact.
The β-structures are further protected from solvent interactions by a shell of
unstructured or α-helical peptide. Previous studies have shown that the thermal
denaturation of papain exhibits two distinct phases.[83, 16] It has been proposed that
this is due to quasi-independent unfolding transitions in the individual domains. The
preservation of the β-core of domain 2 suggested by the CD and SANS data may be
the source of the observed biphasic denaturation pathway observed in these previous
studies of papain unfolding. [83]

4.5 Conclusions
The tendency of surfactants to bind to proteins with a high affinity at low
concentrations, and work co-operatively to unfold tertiary and secondary structural
elements allows for the controlled manipulation of protein solution conformation.
The unfolding of papain may be induced by binding to azoTAB surfactants, with
observable structural changes beginning at protein:surfactant ratios ~1:10. Circular
Dichroism revealed that the unfolding pathway primarily involves a progressive loss
in a-helical content with increasing surfactant concentration, along with an increase in
104

random coil character. Small angle neutron scattering (SANS) was used to obtain
high resolution structural data on the non-native protein conformations. Guinier
analysis and pair distance distribution calculations determined that the radius of
gyration increases from R
g
= 19.5 Å for the native protein, to R
g
=26.3 Å in the
presence of azoTAB surfactants (~1:25). Likewise, the maximum dimension within
the protein increases from 48 Å for the pure protein, to ~ 62 Å at elevated surfactant
concentrations. Shape reconstruction, applied to the SANS data using the program
GA_STRUCT, suggested that the unfolding process involves a disruption of the two-
domain structure of the protein. Surfactant binding leads to a pronounced separation
of the two domains, along with partial unfolding of one of the domains.

4.5.1 Future work
It would be interesting to relate these conformational changes to changes in
biochemical activity and substrate specificity. To this end, an enzymatic assay could
be used to measure papain activity under various environmental conditions, and in the
presence of azoTAB surfactants. A common approach is a titrimetric rate
determination based on the reaction:
Na-Benzoyl-L-Arginine Ethyl Ester (BAEE) + H
2
O → Na-Benzoyl-L-Arginine + Ethanol,
which is catalyzed by papain. The solution is maintained at pH = 6.2 by addition of
small volumes of a NaOH solution of known concentration. The volume of NaOH
required to maintain this pH is then plotted against the reaction time giving a measure
of the protein’s activity.[89, 90]
105

Chapter 5
Future Studies
5.1 Introduction
The unique features of azoTAB surfactants allow photo control of electrical
conductivity, surface tension and hydrophobicity. This provides the possibility for
many interesting applications in systems of biological macromolecules. Through
their effects on protein secondary, tertiary and quaternary structure, these surfactants
may give reversible control over a host of material properties. In this section,
preliminary data from the self-association of β-amyloid peptide fragment (1-42) is
discussed in a continuation of the fibrilogenesis study in insulin. Finally, the effects
of azoTAB surfactants on protein dynamics, and the ability to observe these dynamics
will be discussed in a proposal for future work.

5.2 Amyloid peptide fragment (1-42)
In order to complement our work on insulin aggregation, and expand our
analysis of the amyloid fibrillation process, we have begun studying the self-
association process in the Amyloid-β peptide fragments. Amyloid-β peptides (Aβ) are
the focus of much research stemming from the fact that the aggregated peptides are
the major component of extracellular senile plaques, which form in the diseased brain
and are required for Alzheimer’s diagnosis.[91, 92]
106

Recent studies have shown that the neuronal toxicity of the peptide depends
on its aggregation state [93], and that soluble oligomers affect synaptic activity well
before any amyloid has aggregated into compact plaques. Aβ peptide is a cleavage
product of amyloid precursor protein (APP), a type 1 transmembrane protein.[94]
Proteolytic processing of APP results in amyloid-β peptides of various lengths, and
numerous cleavage products are created in the brain.[95] While Aβ(1–40) and Aβ(1–
42) have been the focus of much interest, some of the other fragments exhibit similar
neurotoxicity. [96] The kinetics of aggregation of Aβ are heavily dependent on
peptide length and which residues are included in the fragment.[97] The Aβ(1–28)
fragment, for example, was found to aggregate very slowly with respect to the full
length peptide, taking ~ 10 - 12 days instead of the few minutes required by Aβ(1–40)
and Aβ(1–42) to complete the process. The 11 residue fragment Aβ(25–35) however,
has been shown to aggregate with time, forming β-structured fibrils and retaining the
toxicity of the full length peptide Aβ(1–42). It has been proposed that this Aβ(25–35)
fragment represents the biologically active portion of the peptide. [98] Studies on
Aβ(25–35) have shown that like the full length fragment, toxicity is dependent on the
aggregation state of the peptide. However, unlike the full length peptide, the fragment
forms fibrils immediately in solution and aging is not required for aggregation. [99]
Additionally, while the Aβ(25–35) fragment does form oligomers, it does not form
the dense core amyloid plaques [100]
Little is known about the solution structure of the early amyloid species, and
in order to complement previous work in our research group on the Aβ(1–40) system
107

[12], and extend the application of experimental techniques developed using the
model insulin system, we are studying the self-association of the Aβ(1–42) fragments
and the effect of our photo-responsive surfactants.

5.2.1 Preliminary Data and Future Experiments
At this point, we have conducted small angle neutron scattering experiments
(figure 5.1), and aim to complement the data with techniques such as dynamic light
scattering (DLS), Fourier transform infrared spectroscopy (FT-IR) and atomic force
microscopy (AFM). SANS data were collected at Oak Ridge National Laboratories
(ORNL) on the CG3-BioSANS instrument at the High Flux Isotope Reactor facility
(HFIR). A neutron wavelength of 6 Å was used, and two detector distances of 1.3 m
and 7 m gave a Q range of 0.007 - 0.3 Å
-1
as in the previous chapters of this report.
Aβ(1-42) was purchased from rPeptide (catalog number A-1165) and used as
received. The lyophilized peptide was resuspended in 1 % NH
4
OH in D
2
O (to
minimize incoherent scattering from hydrogen), at a final protein concentration of 1
mg/ml, then sonicated for ~ 3 minutes. The Aβ(1-42) fragment is very aggregation
prone, and as observed in the Aβ(1-40) system [12], is believed to form aggregates
almost immediately in solution. This is indicated in the SANS data by substantial
low angle scattering. In the presence of the trans isomer, the y-intercept extrapolated
at Q = 0 is much higher than that of the pure protein, consistent with an increase in
the molecular weight of solution species, and exhibits a slope at low-Q of -2,
associated with a loose three dimensional fibril network. This increase in the slope at
108

low Q values suggests that, in this sample, the extent of fibrillation is increased with
respect to the pure peptide. Initially, the cis isomer appears to have little effect on the
aggregation state of the protein, with early scattering curves closely resembling those
of the pure protein. Over the experimental time frame, however, the cis isomer
induces further aggregation in the protein, and after 12 hours of incubation the
scattering curve deviates from the pure protein and approaches that of the trans
isomer sample. While the extent of fibrillation continues to increase in the 20 hour
sample, at this point the cis sample has not yet reached the aggregation state achieved
by the trans isomer in the time prior to collection of the first SANS data.

109

(a) (b)

(c) (d)

Figure 5.1. SANS scattering curves for Pure Aβ(1-42) , Aβ(1-42) in the
presence of cis azoTAB, and trans azoTAB.
110

Since it may be assumed that the peptide has formed a significant amount of
fibrillar material in all of the samples analyzed, a modified Guinier analysis for rod-
like structures was applied to the SANS data in addition to a classic Guinier
analysis.[45]

where I
c
(0) is the extrapolated intensity at Q = 0 and R
c
is the cross-sectional radius of
gyration, which can be calculated as the slope of a plot of ln(Q*I(Q)) vs. Q
2
.[44]
Pair distance distribution functions for Aβ in the presence of azoTAB are
plotted in figure 4.5. The difference in the shape of the function may reflect a
difference in the structure of aggregates. In the presence of the trans isomer, the
PDDF is typical of large fibrillar structures, with an R
g
of under 20 nm and a tail
extending over 70 nm. In the assay containing the cis isomer, while the values of R
g
and D
max
are comparable, there is another peak around intermediate length-scales (35
nm).
(a) (b)

Figure 5.2. PDDFs for Aβ(1-42) in the presence of (a) 0.5 mM trans
isomer, (b) 0.5 mM cis isomer.
111

In addition to the techniques discussed in this report, our research group recently
obtained an FT-IR spectrometer, which can provide detailed secondary structural
information.

5.3 Protein Dynamics and Activity
For large multi-domain proteins, the flexibility of the domains with respect to one
another, and the resultant motions of the molecule are often important for the proper
biological function of the protein.[116] Depending on the flexibility of the peptide
structure, these motions may include hinge-bending, molecular breathing or
conformational changes in addition to the atomic vibrations and backbone and side
chain motions which take place on a sub-picosecond to nanosecond time scale.

5.3.1 Neutron Spin Echo Spectroscopy (NSE)
NSE can provide information on protein dynamics and domain motion with
picosecond time resolution. Neutrons have no charge, but have a spin, and as such are
subject to Larmor precession in the presence of an external magnetic field.[101] In
NSE, the incident neutron beam is polarized and passed through a strong magnetic
field. Due to the different velocities of polarized neutrons in the beam, there is
dephasing of the beam polarizations as the neutrons precess to differing extents. The
polarity of the beam is then flipped by an angle of 180 degrees about an axis
112

perpendicular to the neutron flux, reversing the precession direction and rephasing the
neutron beam polarization. If there is no change in neutron kinetic energy, the beam
rephases completely, and neutrons arriving at the detector all share the same polarity.
However, since there are inelastic scattering events taking place when the beam
passes through the sample, there is a spread in polarization of the neutron beam at the
detector. The detector, which measures neutron polarity is mounted on a movable
arm, accessing a Q range of ~ 0.04 - 0.25 Å
-1
. The experimental Q-dependence of the
effective diffusion coefficient, D
eff
, is compared to simulated values, obtained by
models that make assumptions on the nature of the linkages between protein domains.
The three models considered are the rigid body model, in which no internal domain
motions are possible, the freely jointed model, in which the domains are treated as
beads connected by freely rotating bonds, and the soft-linker model, where the protein
domains are modeled as rigid structures connected by soft spring linkers.

5.3.2 Phosphoglycerate kinase (PGK)
PGK is a transferase enzyme that is integral to the glycolysis process, where it
transfers a phosphate group from 1,3-biphosphoglycerate to adenosine diphosphate
(ADP), forming adenosine triphosphate (ATP) and 3-Phosphoglycerate.[102] The
protein is 416 residues long, giving it a molecular weight of 44.5 kDa. One of the
interesting structural features of PKG is that the single polypeptide chain is folded
into two domains of nearly identical size, which are widely separated.[103] The
catalytic site sits at the interface of the two domains, and it is thought that the each of
113

the two substrates binds to a domain, and the hinge bending motion of the enzyme
brings them into close proximity.[104] The domains generally correspond to the N-
and C-termini of the chain, although a helix comprised of the final 12 residues of the
C-terminus (404 - 416) is buried in the N-terminal domain. This means that there are
two chain sections forming the long flexible linker between the domains; one α-helix,
and one irregularly structured section. Although there are no obvious similarities in
the amino acid sequence of the two domains, they have very similar structures. Each
is comprised of a central six-stranded β-sheet surrounded by helices, in a (β + α)
motif common in glycolytic enzymes (figure 5.3). [105]

Figure 5.3. Structure of phosphoglycerate kinase molecule showing binding
of substrate in active site

The azoTAB family of surfactants, which preferentially bind to a-structures, have the
potential to control the catalytic action of this enzyme by reversibly controlling
domain structure and dynamics. The largely α-helical hinge region represents a target
for our surfactant and a mechanism for controlling domain motions and activity.
Additionally, the domain structure, with buried β-sheets and exposed helices,
114

suggests reversible binding of the azoTAB surfactants in these regions could have
substantial effects on substrate binding affinity and specificity. Recently small-angle
neutron scattering (SANS) and neutron spin echo (NSE) spectroscopy were used by
our research group to analyze the tertiary structure and internal dynamics of
lysozyme. [81] Lysozyme is another two domain protein that undergoes hinge-
bending motions. It was found that binding of the trans isomer of the azoTAB
surfactant resulted in swelling of the hinge region, away from the active site,
enhancing molecular flexibility and inducing super-activity (an 8-fold increase in
enzymatic activity was observed).[81] The combination of the neutron based
techniques used in the study (SANS and NSE) allow examination of the solution
structure of partially folded proteins in non-native conformations, and can provide
information on protein dynamics and domain motion with picosecond time resolution.
We propose a similar analysis of PGK, relating domain structure and dynamics to
biochemical activity.

115

References
1. Zhang, X.-j., Wozniak, J. A. and Matthews, B. W. Protein Flexibility and
Adaptability Seen in 25 Crystal Forms of T4 Lysozyme. Journal of Molecular
Biology 250, 527-552 (1995).

2. Hammes, G. G. Multiple Conformational Changes in Enzyme Catalysis.
Biochemistry 41, 8221-8228 (2002).

3. Boehr, D. D., Dyson, H. J. and Wright, P. E. An Nmr Perspective on Enzyme
Dynamics. Chemical Reviews (Washington, DC, United States) 106, 3055-
3079 (2006).

4 Shang, T., Smith, K. A., and Hatton, T. A. (2003) Photoresponsive Surfactants
Exhibiting Unusually Large, Reversible Surface Tension Changes under
Varying Illumination Conditions. Langmuir 19, 10764-10773.

5. Lee, C. T., Jr., Smith, K. A., and Hatton, T. A. (2004) Photoreversible
Viscosity Changes and Gelation in Mixtures of Hydrophobically Modified
Polyelectrolytes and Photosensitive Surfactants. Macromol. 37, 5397-5405.

6. Eastoe, J. and Vesperinas, A. Self-Assembly of Light-Sensitive Surfactants.
Soft Matter 1, 338-347 (2005)

7. Griko, Y. V., Freire, E., Privalov, G., Van Dael, H., and Privalov, P. L. (1995)
The unfolding thermodynamics of c-type lysozymes: a calorimetric study of
the heat denaturation of equine lysozyme. J. Mol. Biol. 252, 447-459.

8. Hamill, A. C., Wang, S.-C. and Lee, C. T., Jr. Probing Lysozyme
Conformationwith Light Reveals a New Folding Intermediate. Biochemistry
44, 15139-15149 (2005).

9. Lee, C. T., Jr., Smith, K. A. and Hatton, T. A. Photocontrol of Protein
Folding: The Interaction of Photosensitive Surfactants with Bovine Serum
Albumin. Biochemistry 44, 524-536 (2005).

10. Scheraga, H. A., Wedemeyer, W. J., and Welker, E. (2001) Bovine pancreatic
ribonuclease A: oxidative and conformational folding studies. Methods
Enzymol. 341, 189-221.

11. Wang, S.-C. and Lee, C. T., Jr. Protein Secondary Structure Controlled with
Lightand Photoresponsive Surfactants. Journal of Physical Chemistry B 110,
16117-16123(2006)
116

12. Andrea C. Hamill and C. Ted Lee, Jr., Photocontrol of β-Amyloid Peptide (1-
40) Fibril Growth in the Presence of a Photosurfactant. J. Phys. Chem. B
2009, 113, 6164–6172

13. Hamill, A. C., Wang, S.-C. and Lee, C. T., Jr. Solution Structure of an
Amyloid-Forming Protein During Photoinitiated Hexamer-Dodecamer
Transitions Revealed through Small-Angle Neutron Scattering. Biochemistry
46, 7694-7705 (2007).

14. Hayashita, T., Kurosawas, T., Miyata, T., Tanaka, K., and Igawa, M. (1994)
Effect of structural variation within cationic azo-surfactant upon
photoresponsive function in aqueous solution. Colloid Polym. Sci. 272, 1611-
1619.

15. Lorenzo A, Yankner B, β-Amyloid neurotoxicity requires fibril formation and
is inhibited by Congo red Proc. Nat.. Acad. Sci. Vol. 91, pp. 12243-12247,
1994

16. Frieden C., Nichol L. W, Protein-Protein Interactions, John Wiley and Sons,
Inc, 1981 pg. 31-100

17. Chiti, F. and C.M. Dobson, Protein misfolding, functional amyloid, and
human disease. Annual Review of Biochemistry, 2006. 75: p. 333-366.

18. Klein, W.L., W.B. Stine, and D.B. Teplow, Small assemblies of unmodified
amyloid-b protein are the proximate neurotoxin in Alzheimer's disease.
Neurobiology of Aging, 2004. 25(5): p. 569-580.

19. Ross C. Poirer M, Protein aggregation and neurodegenerative disease, Nature
med. 2004 1038-10

20. A. Ahmad, V. N. Uversky, D. Hong, Early Events in the Fibrillation of
Monomeric Insulin Journal of Bio. Chem. 280 52 2005

21. Nayak A., Sorci M, Krueger S., A universal pathway for amyloid nucleus and
precursor formation for insulin Proteins 2009; 74:556–565.

22. M N Murthy, T S I. Reid III, Structure of Beef Liver Catalase .J. Mol. Rid.
(1981) 152. 46.5-499

23. Fita I., Silva A, M. MurthyThe Refined Structure of Beef Liver Catalase at 2-
5 A Resolution Acta Cryst. (1986). B42, 497-515
117

24. Unwin, P. N. T Beef Liver Catalase Structure:Interpretation of Electron
Micrographs J. Mol. Biol. (1975) 98, 235-242

25. Prakash K, Prajapati S., Ahmad A, S.K. Jain, Unique Oligomeric
intermediates of bovine liver Catalse, Protein Sci. 2002 11(1) 46-57

26. Prajapati S, Bhakuni V, Babu KR, Jain SK. Eur J Biochem. Alkaline
unfolding and salt-induced folding of bovine liver catalase at high pH, 1998
Jul 1;255(1):178-84.

27. Valstar, A., Brown, W., and Almgren, M. (1999) The Lysozyme-Sodium
Dodecyl Sulfate System Studied by Dynamic and Static Light Scattering.
Langmuir 15, 2366-2374

28. Hammouda B., Probing Nanascale Structures: the SANS toolbox, made
available by NIST center for neutron research
(http://www.ncnr.nist.gov/staff/hammouda/the_sans_toolbox)

29. Hwang, S., Kim, Y., Jung, Y. Chung, W., Microassay for Catalase activity of
the Mycobacterium Tuberculosis from Clinical Isolates, Am. Thor. Soc. 46:
7694-7705

30. D Rugar, P. Hansma, Atomic force microscopy, Physics Today, 1990 23-30

31. Gerald D. Fasman (1996). Circular dichroism and the conformational analysis
of biomolecules Springer. pp. 3 ISBN 9780306451423

32. Whitmore L, Wallace BA (2008). "Protein secondary structure analyses from
circular dichroism spectroscopy: methods and reference databases".
Biopolymers 89 (5): 392–400

33. Brookmeyer, R; Johnson, E; Ziegler-Graham, K; Arrighi, HM (July 2007).
Forecasting the global burden of Alzheimer's disease. Alzheimer's and
Dementia 3 (3): 186–91.

34. Koo. E., Lansbury p, Amyloid diseases: Abnormal protein aggregation in
neurodegeneration Proc. Natl. Acad. Sci. Vol. 96, pp. 9989–9990, 1999

35. Jang WG, Kim EJ, Park KG, Park YB, Choi HS, Kim HJ, Kim YD, Kim KS,
Lee KU, Lee IK (2007). "Glucocorticoid receptor mediated repression of
human insulin gene expression is regulated by PGC-1alpha". Biochem.
Biophys. Res. Commun. 352 (3): 716–21.
118

36. Nielsen L, Frokjaer S, Studies of the Structure of Insulin Fibrils by Fourier
Transform Infrared (FTIR) Spectroscopy and Electron Microscopy, J. of
Pharm. Sci, Vol. 90, NO. 1, 2001

37. Chang X., Jørgensen A.M, Bardrum P., and Led J. Solution Structures of the
R6 Human Insulin Hexamer Biochemistry 1997, 36, 9409-9422

38. Victoria Sluzky, Alexander M. Klibanov, and Robert Langer Mechanism of
Insulin Aggregation and Stabilization in Agitated Aqueous Solutions
Biotechnology and Bioengineering, Vol. 40, Pp. 895-903 (1992)

39. Dzwolak w, Ravindra R., Lendermann J., Aggregation of Bovine Insulin
Probed by DSC/PPC Calorimetry and FTIR Spectroscopy Biochemistry 2003,
42, 11347-11355

40. Arrowsmith, C. H., Pachter, R., Altman, R. B., Iyer, S. B., & Jardetzky, O.
(1990) Biochemistry 29, 6332-6341.

41. Badsberg, U., Jørgensen, A. M. M., Gesmar, H., Led, J. J., Hammerstad, J.
M., Jespersen, L., & Ulstrup, J. (1996) Biochemistry 35, 7021-7031.

43. Porod, G. (1982) in Small Angle X-ray Scattering (Glatter, O., Ed.) pp 17-51,
Academic Press, London.

44. Stenstam, A., Montalvo, G., Grillo, I., and Gradzielski, M. (2003) Small
Angle Neutron Scattering Study of Lysozyme-Sodium Dodecyl Sulfate
Aggregates. J. Phys. Chem. B 107, 12331-12338

46. Blundell T, Dodson G, Hodgkin D, Mercola D. 1972. Insulin: The structure in
the crystal and its reflection in chemistry and biology. Adv Protein Chem
26:279–402.

47. Hua QX, Shoelson SE, Kochoyan M, Weiss MA.1991. Receptor binding
redefined by a structural switch in a mutant insulin. Nature 238–241.

48. Wei J, Lin Y-Z, Zhou J-M, Tsou C-L. 1991. FTIR studies of secondary
structures of bovine insulin and its derivative. Biochim Biophys Acta 1080:
29–33.

49. Vecchio G, Bossi A, Pasta P, Carrea G. 1996. Fourier-transform infrared
conformational study of bovine insulin in surfactant solutions. Int. J. Peptide
Protein Res 8:113–117.
119

50. Zhou, W.; Islam, M. F.; Wang, H.; Ho, D. L.; Yodh, A. G.; Winey, K. I.;
Fischer, J. E. Chem. Phys. Lett. 2004, 384, 185.
51. Yan, H., Saiani, A. Gough, J. E.; Miller, Thermoreversible Protein Hydrogel
as Cell Scaffold A. F. Biomacromolecules 2006, 7, 2776.

52. Glatter, O. (1979) The interpretation of real-space information from small-
angle scattering experiments. J. App. Cryst. 12, 166-175.

53. Dugas, H. (1996) Bioorganic Chemistry, 3rd. ed., Springer Verlag, New York.

54. Aune, K. C., and Timasheff, S. N. (1971) Dimerization of alpha-
chymotrypsin. I. pH
dependence in the acid region. Biochemistry 10, 1609-1617.

55. Srisailam, S., Kumar Thallampuranam Krishnaswamy, S., Rajalingam, D.,
Kathir Karuppanan, M., Sheu, H.-S., Jan, F.-J., Chao, P.-C., and Yu, C. (2003)
Amyloidlike fibril formation in an all β-barrel protein. Partially structured
intermediate state(s) is a precursor for fibril formation. J. Bio. Chem. 278,
17701-9.

56. Koren, R., and Hammes, G. G. (1976) A kinetic study of protein-protein
interactions. Biochemistry 15, 1165-1171.

57. Pallares, I., Vendrell, J., Aviles, F. X., and Ventura, S. (2004) Amyloid Fibril
Formation by a Partially Structured Intermediate State of a-Chymotrypsin. J.
Mol.Biol. 342, 321-331.

58. Yankner, B. A. & Mesulam, beta-Amyloid and the pathogenesis of
Alzheimer's disease. (1991) N. Engl. J. Med. 325, 1849-1857.

59. Bruice, T. C., and Benkovic, S. J. (2000) Chemical basis for enzyme catalysis,
Biochemistry 39, 6267-6274.

60. Daniel, R. M., Dunn, R. V., Finney, J. L., and Smith, J. S. (2003) The role of
dynamics in enzyme activity, Annual Rev. of Biophys. and Biomol Structure
32:69-92.

61. Daniel, R. M., Dunn, R. V., Finney, J. L., and Smith, J. C. (2003) The role of
dynamics in enzyme activity, Annual Review of Biophysics and Biomolecular
Structure 32, 69-92.

120

62. Olsson, M. H. M., Parson, W. W., and Warshel, A. (2006) Dynamical
contributions to enzyme catalysis: Critical tests of a popular hypothesis,
Chem.Rev. 106, 1737-1756.

63. Zhang, X., and Houk, K. N. (2005) Why enzymes are proficient catalysts:
beyond the Pauling paradigm, Acc. Chem. Res. 38, 379-385.

64. Wong, D. W. S. (1995) in Food Enzymes: Structure and Mechanism, pp 346-
355.

65. Vainshtein, B. K., Melik-Adamyan, W. R., Barynin, V. V., Vagin, A. A., and
Grebenko, A. I. (1985) Spatial organization of catalase proteins, proc. Int.
Symp. Biomol. Struct. Interactions, Suppl. J. Biosci. 8, 471-479.

66. Vainshtein, B. K., Melik-Adamyan, (1985). Symp. Biomol. Struct.
Interactions, Suppl. J. Biosci. 8, 471-479.

67. Prajapati, S., Bhakuni, V., Babu, K.R., and Jain, S. (1998) Alkaline unfolding
and salt-induced folding of bovine liver catalase at high pH, Eur. J. Biochem.
255, 178-184.

68. Samejima, T., Kamata, M., and Shibata, K. (1962) Dissociation of beef liver
catalase, J. Biochem. 51, 181-187.

69. Samejima, T. a. Y., J.T. (1963) Reconstitution of acid-denatured catalase, J.
Biol. Chem. 238, 3256–3261.

70. Jones, M. N., Manley, P., Midgley, P.J.W., and Wilkinson, A.E. (1982)
Dissociation of bovine and bacterial catalases by sodium-dodecyl sulphate,
Biopolymers 21, 1435–1450.

71. Fita, I., and Rossmann, M. G. (1985) The active center of catalase, J. Mol.
Biol. 185, 21-37.

72. Kirkman, H. N., and Gaetani, G. F. (1984) Catalase: a tetrameric enzyme with
four tightly bound molecules of NADPH, Proc. Natl. Acad. Sci. U.S.A 81,
4343–4347.

73. Hillar, A., and Nicholls, P. (1992) A mechanism for NADPH inhibition of
catalase compound II formation, FEBS Lett. 314, 179–182.
121

74. Kirkman, H. N., Galiano, S., and Gaetani, G. F. (1987) The function of
catalase bound NADPH, J. Biol. Chem. 262, 660–666.

75. Fita, I., Rossmann, M. G. (1985) The NADPH binding site on beef liver
catalase, Proc. Natl. Acad. Sci. U.S.A. 82, 1604-1608.

76. Stenstam, A., Montalvo, G., Grillo, I., and Gradzielski, M. (2003) Small
Angle Neutron Scattering Study of Lysozyme-Sodium Dodecyl Sulfate
Aggregates. J. Phys. Chem. B 107, 12331-12338.

77. Tanford, C. and Lovrien, R. 1962. Dissociation of catalase into subunits. J.
Am.Chem. Soc. 84: 1892–1896.

78. Chen X, Xie H, Kong J, Deng J. Characterization for
didodecyldimethylammonium bromide liquid crystal film entrapping catalase
with enhanced direct electron transfer rate. Biosens Bioelectron 2001, 16,
115-120.

79. Takeda, A., Hirano, K., Shiroya, Y., and Samejima, T. (1983) On the
denaturation of porcine erythrocyte catalase with alkali, urea, and guanidine
hydrochloride in relation to its subunit structure, J. Biochem. 93, 967-975.

80. Deisseroth, A., and Dounce, A. L. (1967) Nature of change produced in
catalase by lyophilization, Arch. Biochem. Biophys. 120, 671-692.

81. Mirarefi P, S.C. Wang, Faraone A, Lee C. T. L. Jr. Light Induced Protein
Dynamics observed with Neutron Spin Echo Measurements. Biochem.m

82. Mirarefi P, Lee C. T Jr., Photo Induced unfolding and inactivation of bovine
carbonic anhydrase in the presence of a photoresponsive surfactant. Proteins
and Proteomics Vol 1804, 1, 106-114

83. Obalinski T., Protein Folding: New Research 2006 Nova Sci. Publishers,
ISBN 1-59454-835-8

84. Adams, M. J., Blundell, T. L., Dodson, E. J., Dodson, G. G., Vijayan, M.,
Baker, E. N., Harding, M. M., Hodgkin, D. C., Rimmer, B., and Sheat, S.
(1969), Nature (London) 224,491.

122

85. Adams, M. J., Ford, G. C., Koekoek, R., Lentz, P. J., Jr., McPherson, A., Jr.,
Rossmann, M. G., Smiley, Schevitz, R. W., and Wonacott, A. J. (1970),
Nature (London) 227,1098.

86. Kampuis T. G., Kalk B, Drenth, J Structure of Papain Refined at 1.65 A
Resolution, J. Mol. Biol. (1984) 179, 233-256

87. Chen Y. H, Tsi Yang J. J., Martinez H. Determination of the Secondary
Structures of Proteins by Circular Dichroism and Optical Rotatory Dispersion
Bio Chem. . 11, 22, 1972

88. Hernfindez-Arana A and Soriano-Garcia M Detection and characterization by
circular dichroism of a stable intermediate state formed in the thermal unfolding
of papain Biochimica et Biophysica Acta 954 (1988) 170-175

89. Reagent Chemicals ACS Specifications, 8th ed, 95, American Chemical Society,
Washington, DC (1993)

90. Arnon, R. (1970) Methods in Enzymology, XIX, (Perlmann, G. E. & Lorand,
L., eds.) Vol. 19, pp. 226

91. Louw, C., Gordon, A., Johnston, N., Mollatt, C., Bradley, G., Whiteley, C.G.,.
Arginine deiminases: therapeutic tools in the etiology and pathogenesis of
Alzheimer's disease. J. Enzyme Inhib. Med. Chem. 22, 121–126. (2007)

92. Robakis, N.K.,. β-amyloid and amyloid precursor protein: chemistry,
molecular biology,and neuropathology. In: Terry, R., Katzman, R., Bick, K.L.
(Eds.), Alzheimer's Disease. Raven Press, New York (1994)

93. Tamagno E., P. Bardini, M. Guglielmotto, O. Danni, M. Tabaton The various
aggregation states of β-amyloid 1–42 mediate different effects on oxidative
stress, neurodegeneration, and BACE-1 expression, Biology & Medicine 41
(2006) 202–212

94. Ling, Y., Morgan, K., Kalsheker, N., Amyloid precursor protein (APP) and
the biology of proteolytic processing: relevance to Alzheimer's disease. Int. J.
Biochem. Cell Biol. 35, 1505– 1535. (2003)

95. Maddalena, A.S., Papassotiropoulos, A., Gonzalez-Agosti, C., Signorell, A.,
Hegi, T., Pasch, T., Nitsch, R.M., Hock, C.,. Cerebrospinal fluid profile of
amyloid β peptides in patients with Alzheimer's disease determined by protein
biochip technology. Neurodegener. Dis. 1, 231–235. (2004)

123

96. Y. G. Kaminsky, M. W. Marlatt , Mark A. Smith, Subcellular and metabolic
examination of amyloid-β peptides in Alzheimer disease pathogenesis:
Evidence for Aβ25–35 Experimental Neurology 221 (2010) 26–37

97. Jarrett, J.T., Berger, E.P., and Lansbury, P.T., Jr. The C-terminus of the β
protein is critical in amyloidogenesis. Ann. N. Y. Acad. Sci. 695, 144–148
(1993)

98. Pike, C.J.,Walencewicz-Wasserman,A.J., Kosmoski, J., Cribbs,D.H., Glabe,
C.G.,Cotman, C.W.,. Structure-activity analyses of β-amyloid peptides:
contributions of the β 25–35 region to aggregation and neurotoxicity. J.
Neurochem. 64, 253–265. (1995)

99. Hensley, K., Carney, J.M., Mattson, M.P., Aksenova, M., Harris, M., Wu,
J.F., Floyd, R.A., Butterfield, D.A.,. A model for β-amyloid aggregation and
neurotoxicity based on free radical generation by the peptide: relevance to
Alzheimer disease. Proc. Natl. Acad. Sci. U. S. A. 91, 3270–3274. (1994)

100. Gengler, S., Gault, V., Harriott, P., Hölscher, C., Impairments of hippocampal
synaptic plasticity induced by aggregated beta-amyloid (25–35) are
dependent on stimulation-protocol and genetic background. Exp. Brain Res
(2006).

101. V. Bargmann, L. Michel, and V. L. Telegdi, Precession of the Polarization of
Particles Moving in a Homogeneous Electromagnetic Field, Phys. Rev. Lett.
2, 435 (1959).

102. Watson HC, Walker NP, Shaw PJ, et al. (1982). "Sequence and structure of
yeast phosphoglycerate kinase". EMBO J. 1 (12): 1635–40

103. Bank R.D., Blake C.C.F. Sequence, Structure and activity of
phosphoglycerate Kinase: a possible hinge bending protein, Nature 179, 773,
1979

104. Bernstein, B E; Michels, P A; Hol, W G, Synergistic effects of substrate-
induced conformational changes inphosphoglycerate kinase activation Nature,
ISSN 0028-0836, 01/1997, Volume 385, Issue 6613: 275 - 278

105. Sundarlingham, M. Structure and conformation of nucleic acids: Protein
Nucleic acid interactions. University park press, 1975: 487-527

124

106. Whittingham, J.L, Scott, D.J, Chance, K., Wilson, A., Insulin at pH 2:
Structural analysis of the conditions promoting insulin fibre formation. J. Mol.
Biol. 2002 318(2): 479-490

107. Lee J, Pilch P.F. The Insulin Receptor: Structure, Function, and Signaling,
Am. J. Phys. Cell Physiol. 1994, 266(2): 319-334

108. Jansen R., Dzowlak W., Winter R., Amyloidogenic self-assembly of insulin
Aggregates Probed By High Resolution Atomic Force Mocroscopy, Biophys.
Journal, 2005 88: 1344-1353

109. Svergun D.I., Petoukhov M, Koch M, Determination of domain structure of
proteins from X-ray Solution Scattering. Biophysical Journal, 2001, 80: 2946-
2953

110. Heller W.T., Krueger J.K., Trewhella J., Further Insights into Calmodulin-
Myosin Light Chain Kinase Interaction from Solution Scattering and Shape
Restoration. Biochem. 2003, 42: 10579-10588

111. Konno K., Hirayama C., Nakamura M., Tateishi Y., Hattori M. Papain
Protects Papaya Trees From Herbivorous Insects: Role of Cysteine Proteases
in Latex, The Plant Journal, 2004, 37: 370-378

112. Brocklehurst K., Willenbrock F., Salih E., Hydrolyic Enzymes, Elsevier,
Amsterdam, 1987, pg 39

113. Sreerama N., Woody R., Computation and Analysis of Protein Circular
Dichroism Spectra, Methods in Enzymology, 2004, 383: 318-351

114. Greenfeild N., Using Circular Dichroism Spectra to Estimate Protein
Secondary Structure, Nature Protocols, 2007, 1.6: 2876-2901

115. Manavalan P., Johnson C, Protein Secondary Structure from Circular
Dichroism Spectra, Proc. Int. Sym. Biomol. Struc. Interaction, Suppl. J.
Biosci. 1985, 8: 141-149

116. Heller W., Influence of Multiple Well Defined Conformations on Small Angle
Scattering of Proteins in Solution. Biological Crystallography, 2004 D61: 33-
44

117. Bu, Z., Engelman, D.M., A method for determining transmembrane helix
association and orientation in detergent micelles using small angle X-ray
scattering, Biophys. J. 1999, 77: 1064-1073.

125

118. Weimer, E.A., Ofman, R., Middelkoop, E., de Boer, M., Wanders, R.J.,
Production and Characterization of Monoclonal Antibodies against Native and
Disassembled Human Catalase, J. of Immunological Methods, 1992, 151:165-
175

119. Otzen, D., Protein-Surfactant Interactions: A Tale of Many States, Biochimica
et Biophysica Acta, 2011, 1814: 562-591

120. Necula, M., Kayed, R., Milton, S., Glabe, C., Small Molecule Inhibitors of
aggregation Indicate That Amyloid-Beta Pathways are Independent and
Distict, J. of. Biological Chem., 2007, 282: 10311-10324

121. Convertino, M., Pellarin, R., Catto, M., Carotti, A., Caflisch, A., 9,10-
Anthraquinone Hinders Beta-Aggregation: How Does A Small Molecule
Interfere with AB-Peptide Amyloid Fibrillation?, Protein Science, 2009, 18:
792-800

122. LeVine, H. III, Quantification of Beta-Sheet Amyloid Fibril Structures with
Thioflavin T, Methods in Enzymology, 1999, 309: 274-284

123. Giehm, L., Oliveira, C.L.P., Chritiansen, G., Pederson, J.S., Otzen, D.E.,
SDS-induced Fibrillation of Alpha-Synuclein: An Alternative Fibrillation
Pathway. J. Mol. Biol., 2010, 401: 115-133

124. Reynolds, J.A., Herbert, S., Polet, H., Steinhardt, J., The Binding of Divers
Detergent Anions to Bovine Serum Albumin, Biochemistry, 1967, 6: 937-943

125. Reynolds, J.A., Herbert, S., Steinhardt, J., The Binding of Some Long-chain
Fatty Acid Anions and Alchohols by Bovine Serum Albumin, Biochemistry,
1968, 7: 1357-1361

126. Otzen, D.E., Sehgal, P., Westh, P., A-Lactalbumin is Unfolded by All Classes
of Detergents but With Different Mechnisms, J. Coll. Int. Sci, 2009, 329: 273-
283

127. Jones, M.N., Protein-Surfactant Interactions, in Surface Activity of proteins,
Marcel Dekker, New York, 1996: 237-284

128. Yonath, A., Podjarny, A., Honig, B., Sielecki, A., Traub, W., Crystallographic
studies of protein denaturation and renaturation, Sodium Dodecyl Sulfate
Induced Structural Changes in Triclinic Lysozyme, Boichemistry 1977, 16:
1418-1424

126

129. Campbell, N., Biology, Fourth ed., Benjamin/Cummings Publishing
Company, ISBN 0-8053-1940-9, 1996, p. 309-310

130. Segrest, J.P., Jones, M,K,. De Loof, H,. Brouillette, C,G,. The Amiphipathic
Helix in the Exchangeable Apolipoptroteins: A Review of Secondary
Structure and Function, J. of Lipid Res. 1992, 33(2):141-166

127

Bibliography
Adams, M. J., Blundell, T. L., Dodson, E. J., Dodson, G. G., Vijayan, M.,
Baker, E. N., Harding, M. M., Hodgkin, D. C., Rimmer, B., Sheat, S.,
Structure of Rhombohedral 2 Zinc Insulin Crystals (1969), Nature (London)
224,491.

Adams, M. J., Ford, G. C., Koekoek, R., Lentz, P. J., Jr., McPherson, A., Jr.,
Rossmann, M. G., Smiley, Schevitz, R. W., Wonacott, A. J. Structure of
lactate dehydrogenase at 2.8 Å resolution. (1970), Nature (London) 227,1098.

Ahmad, A., Uversky, V. N., Hong, D., Early Events in the Fibrillation of
Monomeric Insulin Journal of Bio. Chem. 280 52 2005

Arnon, R. (1970) Methods in Enzymology, XIX, (Perlmann, G. E. & Lorand,
L., eds.) Vol. 19, pp. 226

Arrowsmith, C. H., Pachter, R., Altman, R. B., Iyer, S. B., & Jardetzky, O.
(1990) Biochemistry 29, 6332-6341.

Aune, K. C., Timasheff, S. N. (1971) Dimerization of alpha-chymotrypsin. I.
pH dependence in the acid region. Biochemistry 10, 1609-1617.

Badsberg, U., Jørgensen, A. M. M., Gesmar, H., Led, J. J., Hammerstad, J.
M., Jespersen, L., & Ulstrup, J. (1996) Biochemistry 35, 7021-7031.

Bank R.D., Blake C.C.F. Sequence, Structure and activity of
phosphoglycerate Kinase: a possible hinge bending protein, Nature 179, 773,
1979

Bargmann V., Michel L, and V. L. Telegdi, Precession of the Polarization of
Particles Moving in a Homogeneous Electromagnetic Field, Phys. Rev. Lett.
2, 435 (1959).

Bernstein, B E; Michels, P A; Hol, W G, Synergistic effects of substrate-
induced conformational changes inphosphoglycerate kinase activation Nature,
ISSN 0028-0836, 01/1997, Volume 385, Issue 6613: 275 - 278

Blundell T, Dodson G, Hodgkin D, Mercola D. 1972. Insulin: The structure in
the crystal and its reflection in chemistry and biology. Adv Protein Chem
26:279–402.

128

Boehr, D. D., Dyson, H. J. and Wright, P. E. An Nmr Perspective on Enzyme
Dynamics. Chemical Reviews (Washington, DC, United States) 106, 3055-
3079 (2006).

Brocklehurst K., Willenbrock F., Salih E., Hydrolyic Enzymes, Elsevier,
Amsterdam, 1987, pg 39

Brookmeyer, R., Johnson, E., Ziegler-Graham, K.., Arrighi, HM (2007).
Forecasting the global burden of Alzheimer's disease. Alzheimer's and
Dementia 3 (3): 186–91.

Bruice, T. C., Benkovic, S. J. (2000) Chemical basis for enzyme catalysis,
Biochemistry 39, 6267-6274.

Bu, Z., Engelman, D.M., A method for determining transmembrane helix
association and orientation in detergent micelles using small angle X-ray
scattering, Biophys. J. 1999, 77: 1064-1073.

Campbell, N., Biology, Fourth ed., Benjamin/Cummings Publishing
Company, ISBN 0-8053-1940-9, 1996, p. 309-310

Chang X., Jørgensen A.M, Bardrum P., and Led J. Solution Structures of the
R6 Human Insulin Hexamer Biochemistry 1997, 36, 9409-9422

Chen Y. H, Tsi Yang J. J., Martinez H. Determination of the Secondary
Structures of Proteins by Circular Dichroism and Optical Rotatory Dispersion
Bio Chem. . 11, 22, 1972

Chen X, Xie H, Kong J, Deng J. Characterization for
didodecyldimethylammonium bromide liquid crystal film entrapping catalase
with enhanced direct electron transfer rate. Biosens Bioelectron 2001, 16,
115-120.

Chiti, F., Dobson, C.M. Protein misfolding, functional amyloid, and human
disease. Annual Review of Biochemistry, 2006. 75: p. 333-366.

Convertino, M., Pellarin, R., Catto, M., Carotti, A., Caflisch, A., 9,10-
Anthraquinone Hinders Beta-Aggregation: How Does A Small Molecule
Interfere with AB-Peptide Amyloid Fibrillation?, Protein Science, 2009, 18:
792-800

129

Daniel, R. M., Dunn, R. V., Finney, J. L., and Smith, J. C. (2003) The role of
dynamics in enzyme activity, Annual Review of Biophysics and Biomolecular
Structure 32, 69-92.

Deisseroth, A., and Dounce, A. L. (1967) Nature of change produced in
catalase by lyophilization, Arch. Biochem. Biophys. 120, 671-692.

Dugas, H. (1996) Bioorganic Chemistry, 3rd. ed., Springer Verlag, New York.

Dzwolak, W, Ravindra R., Lendermann J., Aggregation of Bovine Insulin
Probed by DSC/PPC Calorimetry and FTIR Spectroscopy Biochemistry 2003,
42, 11347-11355

Eastoe, J. and Vesperinas, A. Self-Assembly of Light-Sensitive Surfactants.
Soft Matter 1, 338-347 (2005)

Fita, I., Rossmann, M. G. (1985) The active center of catalase, J. Mol. Biol.
185, 21-37.

Fita I., Silva A, M. MurthyThe Refined Structure of Beef Liver Catalase at 2-
5 A Resolution Acta Cryst. (1986). B42, 497-515

Fita, I., Rossmann, M. G. (1985) The NADPH binding site on beef liver
catalase, Proc. Natl. Acad. Sci. U.S.A. 82, 1604-1608.

Frieden C., Nichol L. W, Protein-Protein Interactions, John Wiley and Sons,
Inc, 1981 pg. 31-100

Gengler, S., Gault, V., Harriott, P., Hölscher, C., Impairments of hippocampal
synaptic plasticity induced by aggregated beta-amyloid (25–35) are
dependent on stimulation-protocol and genetic background. Exp. Brain Res
(2006).

Gerald D. Fasman (1996). Circular dichroism and the conformational analysis
of biomolecules Springer. pp. 3 ISBN 9780306451423

Giehm, L., Oliveira, C.L.P., Chritiansen, G., Pederson, J.S., Otzen, D.E.,
SDS-induced Fibrillation of Alpha-Synuclein: An Alternative Fibrillation
Pathway. J. Mol. Biol., 2010, 401: 115-133

Glatter, O. (1979) The interpretation of real-space information from small-
angle scattering experiments. J. App. Cryst. 12, 166-175.

130

Greenfeild, N., Using Circular Dichroism Spectra to Estimate Protein
Secondary Structure, Nature Protocols, 2007, 1.6: 2876-2901

Griko, Y. V., Freire, E., Privalov, G., Van Dael, H., and Privalov, P. L. (1995)
The unfolding thermodynamics of c-type lysozymes: a calorimetric study of
the heat denaturation of equine lysozyme. J. Mol. Biol. 252, 447-459.

Hammes, G. G. Multiple Conformational Changes in Enzyme Catalysis.
Biochemistry 41, 8221-8228 (2002).

Hamill, A., Lee, C. T. Jr., Photocontrol of β-Amyloid Peptide (1-40) Fibril
Growth in the Presence of a Photosurfactant. J. Phys. Chem. B 2009, 113,
6164–6172

Hamill, A. C., Wang, S.-C. and Lee, C. T., Jr. Probing Lysozyme
Conformationwith Light Reveals a New Folding Intermediate. Biochemistry
44, 15139-15149 (2005).

Hamill, A. C., Wang, S.-C. and Lee, C. T., Jr. Solution Structure of an
Amyloid-Forming Protein During Photoinitiated Hexamer-Dodecamer
Transitions Revealed through Small-Angle Neutron Scattering. Biochemistry
46, 7694-7705 (2007).

Hayashita, T., Kurosawas, T., Miyata, T., Tanaka, K., and Igawa, M. (1994)
Effect of structural variation within cationic azo-surfactant upon
photoresponsive function in aqueous solution. Colloid Polym. Sci. 272, 1611-
1619.

Hammouda B., Probing Nanascale Structures: the SANS toolbox, made
available by NIST center for neutron research
(http://www.ncnr.nist.gov/staff/hammouda/the_sans_toolbox)

Heller W., Influence of Multiple Well Defined Conformations on Small Angle
Scattering of Proteins in Solution. Biological Crystallography, 2004 D61: 33-
44

Heller W.T., Krueger J.K., Trewhella J., Further Insights into Calmodulin-
Myosin Light Chain Kinase Interaction from Solution Scattering and Shape
Restoration. Biochem. 2003, 42: 10579-10588

131

Hensley, K., Carney, J.M., Mattson, M.P., Aksenova, M., Harris, M., Wu,
J.F., Floyd, R.A., Butterfield, D.A.,. A model for β-amyloid aggregation and
neurotoxicity based on free radical generation by the peptide: relevance to
Alzheimer disease. Proc. Natl. Acad. Sci. U. S. A. 91, 3270–3274. (1994)

Hernandez-Arana A and Soriano-Garcia M Detection and characterization by
circular dichroism of a stable intermediate state formed in the thermal
unfolding of papain Biochimica et Biophysica Acta 954 (1988) 170-175

Hillar, A., and Nicholls, P. (1992) A mechanism for NADPH inhibition of
catalase compound II formation, FEBS Lett. 314, 179–182.

Hua Q.X., Shoelson S.E., Kochoyan M., Weiss M.A., 1991. Receptor
binding redefined by a structural switch in a mutant insulin. Nature 238–241.

Hwang, S., Kim, Y., Jung, Y. Chung, W., Microassay for Catalase activity of
the Mycobacterium Tuberculosis from Clinical Isolates, Am. Thor. Soc. 46:
7694-7705

Jang WG, Kim EJ, Park KG, Park YB, Choi HS, Kim HJ, Kim YD, Kim KS,
Lee KU, Lee IK (2007). "Glucocorticoid receptor mediated repression of
human insulin gene expression is regulated by PGC-1alpha". Biochem.
Biophys. Res. Commun. 352 (3): 716–21.

Jansen R., Dzowlak W., Winter R., Amyloidogenic self-assembly of insulin
Aggregates Probed By High Resolution Atomic Force Mocroscopy, Biophys.
Journal, 2005 88: 1344-1353

Jarrett, J.T., Berger, E.P., and Lansbury, P.T., Jr. The C-terminus of the β
protein is critical in amyloidogenesis. Ann. N. Y. Acad. Sci. 695, 144–148
(1993)

Jones, M.N., Protein-Surfactant Interactions, in Surface Activity of proteins,
Marcel Dekker, New York, 1996: 237-284

Jones, M. N., Manley, P., Midgley, P.J.W., and Wilkinson, A.E. (1982)
Dissociation of bovine and bacterial catalases by sodium-dodecyl sulphate,
Biopolymers 21, 1435–1450.

Kaminsky, Y. G., Marlatt , M. W., Smith, M., Subcellular and metabolic
examination of amyloid-β peptides in Alzheimer disease pathogenesis:
Evidence for Aβ25–35 Experimental Neurology 221 (2010) 26–37

132

Kampuis T. G., Kalk B, Drenth, J Structure of Papain Refined at 1.65 A
Resolution, J. Mol. Biol. (1984) 179, 233-256

Kirkman, H. N., Gaetani, G. F. (1984) Catalase: a tetrameric enzyme with
four tightly bound molecules of NADPH, Proc. Natl. Acad. Sci. U.S.A 81,
4343–4347.

Kirkman, H. N., Galiano, S., and Gaetani, G. F. (1987) The function of
catalase bound NADPH, J. Biol. Chem. 262, 660–666.

Klein, W.L., W.B. Stine, and D.B. Teplow, Small assemblies of unmodified
amyloid-b protein are the proximate neurotoxin in Alzheimer's disease.
Neurobiology of Aging, 2004. 25(5): p. 569-580.

Konno K., Hirayama C., Nakamura M., Tateishi Y., Hattori M. Papain
Protects Papaya Trees From Herbivorous Insects: Role of Cysteine Proteases
in Latex, The Plant Journal, 2004, 37: 370-378

Koo. E., Lansbury p, Amyloid diseases: Abnormal protein aggregation in
neurodegeneration Proc. Natl. Acad. Sci. Vol. 96, pp. 9989–9990, 1999

Koren, R., and Hammes, G. G. (1976) A kinetic study of protein-protein
interactions. Biochemistry 15, 1165-1171.

Lee, C. T., Jr., Smith, K. A. and Hatton, T. A. Photocontrol of Protein
Folding: The Interaction of Photosensitive Surfactants with Bovine Serum
Albumin. Biochemistry 44, 524-536 (2005).

Lee, C. T., Jr., Smith, K. A., and Hatton, T. A. (2004) Photoreversible
Viscosity Changes and Gelation in Mixtures of Hydrophobically Modified
Polyelectrolytes and Photosensitive Surfactants. Macromol. 37, 5397-5405.

Lee J, Pilch P.F. The Insulin Receptor: Structure, Function, and Signaling,
Am. J. Phys. Cell Physiol. 1994, 266(2): 319-334

LeVine, H. III, Quantification of Beta-Sheet Amyloid Fibril Structures with
Thioflavin T, Methods in Enzymology, 1999, 309: 274-284

Ling, Y., Morgan, K., Kalsheker, N., Amyloid precursor protein (APP) and
the biology of proteolytic processing: relevance to Alzheimer's disease. Int. J.
Biochem. Cell Biol. 35, 1505– 1535. (2003)

133

Lorenzo A, Yankner B, β-Amyloid neurotoxicity requires fibril formation and
is inhibited by Congo red Proc. Nat.. Acad. Sci. Vol. 91, pp. 12243-12247,
1994

Louw, C., Gordon, A., Johnston, N., Mollatt, C., Bradley, G., Whiteley, C.G.,.
Arginine deiminases: therapeutic tools in the etiology and pathogenesis of
Alzheimer's disease. J. Enzyme Inhib. Med. Chem. 22, 121–126. (2007)

Maddalena, A.S., Papassotiropoulos, A., Gonzalez-Agosti, C., Signorell, A.,
Hegi, T., Pasch, T., Nitsch, R.M., Hock, C.,. Cerebrospinal fluid profile of
amyloid β peptides in patients with Alzheimer's disease determined by protein
biochip technology. Neurodegener. Dis. 1, 231–235. (2004)

Manavalan P., Johnson C, Protein Secondary Structure from Circular
Dichroism Spectra, Proc. Int. Sym. Biomol. Struc. Interaction, Suppl. J.
Biosci. 1985, 8: 141-149

Mirarefi, P., Lee C. T Jr., Photo Induced unfolding and inactivation of bovine
carbonic anhydrase in the presence of a photoresponsive surfactant. Proteins
and Proteomics Vol 1804, 1, 106-114

Mirarefi P, Wang, S.C., Faraone A, Lee C. T. L. Jr. Light Induced Protein
Dynamics observed with Neutron Spin Echo Measurements. Biochem.m

Murthy, M N., Reid, T. III, Structure of Beef Liver Catalase .J. Mol. Rid.
(1981) 152. 46.5-499

Nayak A., Sorci M, Krueger S., A universal pathway for amyloid nucleus and
precursor formation for insulin Proteins 2009; 74:556–565.

Necula, M., Kayed, R., Milton, S., Glabe, C., Small Molecule Inhibitors of
aggregation Indicate That Amyloid-Beta Pathways are Independent and
Distict, J. of. Biological Chem., 2007, 282: 10311-10324

Nielsen L, Frokjaer S, Studies of the Structure of Insulin Fibrils by Fourier
Transform Infrared (FTIR) Spectroscopy and Electron Microscopy, J. of
Pharm. Sci, Vol. 90, NO. 1, 2001

Obalinski T., Protein Folding: New Research 2006 Nova Sci. Publishers,
ISBN 1-59454-835-8

Olsson, M. H. M., Parson, W. W., and Warshel, A. (2006) Dynamical
contributions to enzyme catalysis: Critical tests of a popular hypothesis,
Chem.Rev. 106, 1737-1756.
134

Otzen, D., Protein-Surfactant Interactions: A Tale of Many States, Biochimica
et Biophysica Acta, 2011, 1814: 562-591

Otzen, D.E., Sehgal, P., Westh, P., A-Lactalbumin is Unfolded by All Classes
of Detergents but With Different Mechnisms, J. Coll. Int. Sci, 2009, 329: 273-
283

Pallares, I., Vendrell, J., Aviles, F. X., and Ventura, S. (2004) Amyloid Fibril
Formation by a Partially Structured Intermediate State of a-Chymotrypsin. J.
Mol.Biol. 342, 321-331.

Pike, C.J.,Walencewicz-Wasserman,A.J., Kosmoski, J., Cribbs,D.H., Glabe,
C.G.,Cotman, C.W.,. Structure-activity analyses of β-amyloid peptides:
contributions of the β 25–35 region to aggregation and neurotoxicity. J.
Neurochem. 64, 253–265. (1995)

Porod, G. (1982) in Small Angle X-ray Scattering (Glatter, O., Ed.) pp 17-51,
Academic Press, London.

Prajapati, S., Bhakuni, V., Babu, K.R., Jain, S. (1998) Alkaline unfolding and
salt-induced folding of bovine liver catalase at high pH, Eur. J. Biochem. 255,
178-184.

Prakash K, Prajapati S., Ahmad A, S.K. Jain, Unique Oligomeric
intermediates of bovine liver Catalse, Protein Sci. 2002 11(1) 46-57

Reynolds, J.A., Herbert, S., Polet, H., Steinhardt, J., The Binding of Divers
Detergent Anions to Bovine Serum Albumin, Biochemistry, 1967, 6: 937-943

Reynolds, J.A., Herbert, S., Steinhardt, J., The Binding of Some Long-chain
Fatty Acid Anions and Alchohols by Bovine Serum Albumin, Biochemistry,
1968, 7: 1357-1361

Robakis, N.K.,. β-amyloid and amyloid precursor protein: chemistry,
molecular biology,and neuropathology. In: Terry, R., Katzman, R., Bick, K.L.
(Eds.), Alzheimer's Disease. Raven Press, New York (1994)

Ross C., Poirer M, Protein aggregation and neurodegenerative disease, Nature
med. 2004 1038-10

Rugar, D., Hansma P., Atomic force microscopy, Physics Today, 1990 23-30

135

Samejima, T. (1963) Reconstitution of acid-denatured catalase, J. Biol. Chem.
238, 3256–3261.

Samejima, T., Kamata, M., and Shibata, K. (1962) Dissociation of beef liver
catalase, J. Biochem. 51, 181-187.

Scheraga, H. A., Wedemeyer, W. J., and Welker, E. (2001) Bovine pancreatic
ribonuclease A: oxidative and conformational folding studies. Methods
Enzymol. 341, 189-221.

Segrest, J.P., Jones, M,K,. De Loof, H,. Brouillette, C,G,. The Amiphipathic
Helix in the Exchangeable Apolipoptroteins: A Review of Secondary
Structure and Function, J. of Lipid Res. 1992, 33(2):141-166

Shang, T., Smith, K. A., and Hatton, T. A. (2003) Photoresponsive Surfactants
Exhibiting Unusually Large, Reversible Surface Tension Changes under
Varying Illumination Conditions. Langmuir 19, 10764-10773.

Sluzky, V., Klibanov, A., Langer, R., Mechanism of Insulin Aggregation and
Stabilization in Agitated Aqueous Solutions Biotechnology and
Bioengineering, Vol. 40, Pp. 895-903 (1992)

Sreerama N., Woody R., Computation and Analysis of Protein Circular
Dichroism Spectra, Methods in Enzymology, 2004, 383: 318-351

Srisailam, S., Kumar Thallampuranam Krishnaswamy, S., Rajalingam, D.,
Kathir Karuppanan, M., Sheu, H.-S., Jan, F.-J., Chao, P.-C., and Yu, C. (2003)
Amyloidlike fibril formation in an all β-barrel protein. Partially structured
intermediate state(s) is a precursor for fibril formation. J. Bio. Chem. 278,
17701-9.

Stenstam, A., Montalvo, G., Grillo, I., and Gradzielski, M. (2003) Small
Angle Neutron Scattering Study of Lysozyme-Sodium Dodecyl Sulfate
Aggregates. J. Phys. Chem. B 107, 12331-12338

Sundarlingham, M. Structure and conformation of nucleic acids: Protein
Nucleic acid interactions. University park press, 1975: 487-527

Svergun D.I., Petoukhov M, Koch M, Determination of domain structure of
proteins from X-ray Solution Scattering. Biophysical Journal, 2001, 80: 2946-
2953

136

Takeda, A., Hirano, K., Shiroya, Y., and Samejima, T. (1983) On the
denaturation of porcine erythrocyte catalase with alkali, urea, and guanidine
hydrochloride in relation to its subunit structure, J. Biochem. 93, 967-975.

Tamagno E., P. Bardini, M. Guglielmotto, O. Danni, M. Tabaton The various
aggregation states of β-amyloid 1–42 mediate different effects on oxidative
stress, neurodegeneration, and BACE-1 expression, Biology & Medicine 41
(2006) 202–212

Tanford, C. and Lovrien, R. 1962. Dissociation of catalase into subunits. J.
Am.Chem. Soc. 84: 1892–1896.

Unwin, P. N. T Beef Liver Catalase Structure:Interpretation of Electron
Micrographs J. Mol. Biol. (1975) 98, 235-242

Vainshtein, B. K., Melik-Adamyan, (1985). Symp. Biomol. Struct.
Interactions, Suppl. J. Biosci. 8, 471-479.

Vainshtein, B. K., Melik-Adamyan, W. R., Barynin, V. V., Vagin, A. A., and
Grebenko, A. I. (1985) Spatial organization of catalase proteins, proc. Int.
Symp. Biomol. Struct. Interactions, Suppl. J. Biosci. 8, 471-479.

Valstar, A., Brown, W., and Almgren, M. (1999) The Lysozyme-Sodium
Dodecyl Sulfate System Studied by Dynamic and Static Light Scattering.
Langmuir 15, 2366-2374

Vecchio G, Bossi A, Pasta P, Carrea G. 1996. Fourier-transform infrared
conformational study of bovine insulin in surfactant solutions. Int. J. Peptide
Protein Res 8:113–117.

Watson HC, Walker NP, Shaw PJ, et al. (1982). "Sequence and structure of
yeast phosphoglycerate kinase". EMBO J. 1 (12): 1635–40

Wei J, Lin Y-Z, Zhou J-M, Tsou C-L. 1991. FTIR studies of secondary
structures of bovine insulin and its derivative. Biochim Biophys Acta 1080:
29–33.

Weimer, E.A., Ofman, R., Middelkoop, E., de Boer, M., Wanders, R.J.,
Production and Characterization of Monoclonal Antibodies against Native and
Disassembled Human Catalase, J. of Immunological Methods, 1992, 151:165-
175

137

Whitmore L, Wallace BA (2008). "Protein secondary structure analyses from
circular dichroism spectroscopy: methods and reference databases".
Biopolymers 89 (5): 392–400

Whittingham, J.L, Scott, D.J, Chance, K., Wilson, A., Insulin at pH 2:
Structural analysis of the conditions promoting insulin fibre formation. J. Mol.
Biol. 2002 318(2): 479-490

Wang, S.C. Lee, C. T., Jr. Protein Secondary Structure Controlled with
Lightand Photoresponsive Surfactants. Journal of Physical Chemistry B 110,
16117-16123(2006)

Wong, D. W. S. (1995) in Food Enzymes: Structure and Mechanism, pp 346-
355.

Yan, H., Saiani, A. Gough, J. E.; Miller, Thermoreversible Protein Hydrogel
as Cell Scaffold A. F. Biomacromolecules 2006, 7, 2776.

Yankner, B. A. & Mesulam, beta-Amyloid and the pathogenesis of
Alzheimer's disease. (1991) N. Engl. J. Med. 325, 1849-1857.

Yonath, A., Podjarny, A., Honig, B., Sielecki, A., Traub, W., Crystallographic
studies of protein denaturation and renaturation, Sodium Dodecyl Sulfate
Induced Structural Changes in Triclinic Lysozyme, Boichemistry 1977, 16:
1418-1424

Zhang, X., Houk, K. N. (2005) Why enzymes are proficient catalysts: beyond
the Pauling paradigm, Acc. Chem. Res. 38, 379-385.

Zhang, X., Wozniak, J. A. and Matthews, B. W. Protein Flexibility and
Adaptability Seen in 25 Crystal Forms of T4 Lysozyme. Journal of Molecular
Biology 250, 527-552 (1995).

Zhou, W.; Islam, M. F.; Wang, H.; Ho, D. L.; Yodh, A. G.; Winey, K. I.;
Fischer, J. E. Chem. Phys. Lett. 2004, 384, 185.

Abstract (if available)

Abstract The interaction of photo-responsive surfactants with proteins has been considered as a means to exert reversible control over a number of aspects of protein structure and function. The azobenzene trimethylammonium bromide (azoTAB) family of cationic surfactants undergo a photo-reversible cis to trans isomerization upon exposure to light of the appropriate wavelength. The trans form of the molecule has a lower dipole moment across its azo linkage, and is more hydrophobic than the cis isomer. This results in a higher binding affinity with proteins for the trans isomer, inducing a greater degree of unfolding of tertiary and secondary structures. The surfactant has been applied to the study of the amyloid fibrillation pathway in insulin, in which the protein self-associates into long, insoluble, rod-like structures. The fibrillation rate in insulin is enhanced in the presence of the trans- isomer while the formation of fibrils is largely inhibited in the presence of the cis- isomer, where amorphous aggregates are observed instead. Additionally early fibrillar species formed in the trans-azoTAB assays exhibit a greater tendency to lateral aggregation than do structures in the pure protein, resulting in a more truncated, bundled final aggregate morphology. Use of the surfactants as a means to control protein quaternary solution structure has also been explored in the subunit dissociation of tetrameric catalase. In the presence of azoTAB surfactants, catalase dissociates first into a super-active dimer, then at higher concentrations into an aggregation prone monomer. Finally, the structural changes associated with azoTAB-induced unfolding of the two domain protein papain are tracked. The denaturation pathway involves a progressive loss in secondary structure with increasing azoTAB concentration, along with a relaxation of the compact tertiary structure, and a spatial separation of the two domains. A number of complementary experimental techniques are combined to determine the solution structure of non-native protein conformations, including light scattering, circular dichroism and small angle neutron scattering.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Control the function, dynamics, aggregation of proteins with light illumination

PDF

Photocontrol of protein conformation and enzyme activity in the presence of light-responsive surfactants

PDF

Photocontrol of protein conformation through the use of photoresponsive surfactants, investigated by small angle neutron scattering

PDF

Controlling the form-dynamics-function relationship of proteins with light illumination

PDF

Enhancement of biofuel enzyme activity and kinetics with azoTAB surfactants

PDF

Controlling membrane protein folding with light illumination and catanionic surfactant systems

PDF

A molecular dynamics study of interactions between the enzyme lysozyme and the photo-responsive surfactant azobenzene trimethylammonium bromide (azoTAB)

PDF

Molecular dynamics simulation study of initial protein unfolding induced by the photo-responsive surfactants, azoTAB

PDF

Controlling membrane protein folding using photoresponsive surfactant

PDF

Photocontrol of DNA compaction using light-responsive surfactants

PDF

Protein-surfactant adsorption on solid surfaces

Asset Metadata

Creator Mazwi, Khiza L. (author)

Core Title The interaction of photo-responsive surfactants with biological macromolecules

School Andrew and Erna Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Materials Science

Publication Date 08/21/2012

Defense Date 07/19/2012

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Amyloid,azoTAB,enzyme,neurodegeneration,OAI-PMH Harvest,Photo-responsive,protein,structure-function,surfactant

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Lee, C. Ted, Jr. (committee chair), Goo, Edward K. (committee member), Langen, Ralf (committee member)

Creator Email khiza.mazwi@gmail.com,mazwi@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-91149

Unique identifier UC11288817

Identifier usctheses-c3-91149 (legacy record id)

Legacy Identifier etd-MazwiKhiza-1158.pdf

Dmrecord 91149

Document Type Dissertation

Rights Mazwi, Khiza L.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA