Structural features and modifiers of islet amyloid polypeptide: implications for type II diabetes mellitus

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
STRUCTURAL FEATURES AND MODIFIERS OF ISLET AMYLOID
POLYPEPTIDE: IMPLICATIONS FOR TYPE II DIABETES MELLITUS

by
Sahar Bedrood

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BIOCHEMISTRY AND MOLECULAR BIOLOGY)

August 2009

Copyright 2009 Sahar Bedrood

ii
Acknowledgements

I would like to thank and acknowledge my advisor, Dr. Ralf Langen for his
support, suggestions and advice during this journey. I would also like to thank my
committee members Dr. Ian Haworth, Dr. Jeannie Chen and Dr. Bob Chow for
their confidence in me and their suggestions. Both Dr. Langen and Dr. Haworth
were an integral part of helping me complete my largest project and for that I am
very appreciative. I would also like to thank Dr. Mario Isas for both his friendship
and his mentorship during my PhD years. His personal and scientific advice was
invaluable to me.
I would like to thank my husband, Sevan Aratounians, for his unwavering
support and confidence in me not only in my scientific endeavors, but in
everything I do. I would like to acknowledge my parents for always instilling in me
the value of education. Lastly, I would like to especially acknowledge my father,
Parviz Bedrood. His difficult battle with Alzheimer disease over the last few years
was a continual reminder of the importance of studying and finding a cure for
amyloid diseases.

iii
Table of Contents
Acknowledgements ii

List of Tables vi

List of Figures vii

Abstract ix

Chapter 1: Introduction 1
1.1 Amyloid Proteins in Human Disease 1
1.2 Islet Amyloid Polypeptide: Functional Features 6
1.3 Islet Amyloid Polypeptide: Structural Features 8
1.4 Modifiers of Amyloid Misfolding 11
1.5 EPR and DEER : Method for studying hIAPP structural features 14
1.6 C. elegans: Method for studying Annexin V and amyloids in vivo 17

Chapter 2: Identifying the three-dimensional structure of hIAPP fibrils 21
2.1 Introduction 21
2.2 Materials and Methods 24
2.2.1 Chemicals and Peptides 24
2.2.2 Spin Labeling 24
2.2.3 hIAPP Fibril Formation 25
2.2.4 Electron Microscopy 25
2.2.5 EPR spectroscopy 26
2.2.6 Pulsed EPR and Distance Analysis 26
2.2.7 Molecular Modeling 27
2.3 Results 28
2.3.1 EPR spectra of spin-diluted hIAPP fibrils 28
2.3.2 Structural Features of hIAPP from R1 mobility and accessibility 32
2.3.3 Intramolecular Distances from Four-Pulse DEER Experiments 32
2.3.4 Structural Refinement via Computational Modeling 38
2.3.5 Structural Models of hIAPP fibrils 40
2.4 Discussion 48

Chapter 3: Annexin A5 Reduces Human IAPP Induced Beta-Cell Apoptosis by
Directly Interacting With Human IAPP 56
3.1 Introduction 56
3.2 Materials and Methods 58
3.2.1 Human islet tissue 58
3.2.2 Vesicle leakage assay 59
3.2.3 Thioflavin T fluorescence assay 60
3.2.4 Western Blotting 61
3.2.5 IAPP stock solutions 61
iv
3.2.6 Electron paramagnetic resonance (EPR) and site directed spin
labeling 62
3.2.7 Electron Microscopy 63
3.2.8 Calculations and statistical analysis 64
3.3 Results 64
3.3.1 Annexin A5 protects human islet tissue from hIAPP toxicity 64
3.3.2 Annexin A5 reduces hIAPP-dependent membrane perturbation 67
3.3.3 Annexin A5 reduces thioflavin T fluorescence of hIAPP and fibril
formation (EM) 71
3.3.4 Site-directed spin labeling and EPR spectroscopy reveal that IAPP
fibril formation is strongly altered in the presence of annexin A5 71
3.4 Discussion 74

Chapter 4: The effect of annexin A5 on Aβ and α-synuclein toxicity and fibril
morphology 77
4.1 Introduction 77
4.2 Methods 77
4.2.1 Chemicals and Peptides 77
4.2.2 Generation of α-Synuclein Single Cysteine Mutants 78
4.2.3 Spin Labeling and Fibril Assembly of 115ter α-Synuclein 78
4.2.4 Spin Labeling and Fibril Assembly of hIAPP and Aβ 79
4.2.5 EPR spectroscopy 80
4.2.6 Electron Microscopy 80
4.2.7 Thioflavin T Assays 80
4.2.8 Western blotting for annexin A5 expression in human islets 81
4.2.8 Transgenic C. elegans strains 82
4.2.9 Western Blotting for synuclein expression in annexin A5 transgenic
C. elegans 82
4.2.10 Western blotting for GFP expression in annexin A5 transgenic C.
elegans 83
4.3 Results 84
4.3.1 EPR spectroscopy shows annexin A5 alters amyloid protein
misfolding 84
4.3.2 Annexin A5 reduces fibril formation (EM) and thioflavin T
fluorescence of α-synuclein 89
4.3.3 Annexin A5 expression reduces α-synuclein inclusions in vivo 92
4.4 Discussion 96

Chapter 5: The effect of Curcumin on hIAPP 98
5.1 Introduction 98
5.2 Materials and Methods 99
5.2.1 Chemicals and Peptides 99
5.2.2 Spin Labeling and Fibril Formation for EPR 100
5.2.3 EPR spectroscopy 101
5.2.4 Fibril Formation for Electron Microscopy 101
v
5.2.5 Thioflavin Assay 101
5.2.6 Cell culture 102
5.2.7 Adenovirus generation and transduction experiments 102
5.2.8 hIAPP preparation and toxicity assays 103
5.2.9 MTT assay 104
5.2.11 Propidium iodide staining (PI) 105
5.2.12 Western Blot analysis 105
5.2.13 Statistical analysis 106
5.3 Results 106
5.3.1 Curcumin alters IAPP misfolding 106
5.3.2 The effect of curcumin on preformed fibrils 110
5.3.3 Curcumin partially protects INS cells against exogenous hIAPP
toxicity 115
5.3.4 Curcumin does not prevent endogenous hIAPP-induced apoptosis 117
5.3.5 Curcumin does not protect HIP rat islets against apoptosis 118
5.4 Discussion 120

References 122
vi
List of Tables

Table 1.1: Summary of Amyloidogenic Proteins and Associated Diseases 2

Table 2.1: Comparison of four-pulse DEER and Model Distances 37

vii
List of Figures

Fig. 1.1: Amyloid Fibril Assembly and structural features 3
Fig. 1.2: Ilustration of IAPP amyloid pathology 7
Fig. 1.3: Illustration of proposed models of IAPP fibrils 10
Fig. 1.4: Electron Paramagnetic Resonance 16
Fig. 1.5: Schematic of chimeric transgene construction in C. elegans 19
Fig. 2.1: EPR spectra of hIAPP 21 R1 23
Fig. 2.2: Electron microscope image of hIAPP fibrils 29
Fig. 2.3: EPR spectra of hIAPP fibrils containing single labels at positions 31
2-36

Fig. 2.4: Graph of hIAPP residue mobility 33
Fig. 2.5: Explanation of DEER data analysis 35
Fig. 2.6: Intramolecular distances from 4-pulse DEER experiments. 36
Fig. 2.7: Initial PRONOX model of hIAPP in fibrillar form 41
Fig. 2.8: Illustration of simulated annealing protocol 43
Fig. 2.9: Illustration of Right-handed and Left-handed hIAPP fibrils 45
Fig. 2.10: Predicted Backbone Angles (Phi and Psi) 46
Fig. 2.11: Model of hIAPP fibril 47
Fig. 3.1: Annexin A5 protects islet tissue against hIAPP toxicity 65
Fig. 3.2: Western blot of Islet cell lysates 66
Fig. 3.3: Membrane Leakage Assays of hIAPP in the presence of annexin A5 68
Fig. 3.4: Thioflavin T and EM of hIAPP in the presence of annexin A5 70
Fig. 3.5: EPR spectra of hIAPP in the presence and absence of annexin A5 72
viii

Fig. 4.1: EPR spectra of R1-labeled h-IAPP, Aß and a-synuclein in the 85
presence of annexin A5

Fig. 4.2: EPR Spectra of amyloids with various annexin proteins 88
Fig. 4.3: Electron Micrographs of amyloid fibrils with and without annexin A5 90
Fig. 4.4: Thioflavin T curves of amyloid fibrils with and without annexin A5 91
Fig. 4.5: Annexin A5 expression decreases a-synuclein inclusions in vivo 93
Fig. 4.6: Abundance of a-synuclein in annexin A5 transgenic C. elegans 94
Fig. 5.1: EPR Spectra and EM images of 16R1 hIAPP in the presence of 108
curcumin

Fig. 5.2: Thioflavin T Assay for hIAPP in the presence and absence of 109
curcumin

Fig. 5.3: Curcumin does not change the fibrillar structure of preformed 111
hIAPP

Fig. 5.4: Thioflavin assay for hIAPP preformed fibrils in the presence and 112
absence of curcumin

Fig. 5.5: Curcumin fails to prevent apoptosis induced by h-IAPP 114
overexpression in INS cells

Fig. 5.6: Curcumin does not protect HIP rat islets overexpressing h-IAPP 116
from apoptosis

Fig. 5.7: Curcumin prevents INS cell cytotoxicity 119

ix
Abstract

Protein misfolding is a common motif in a number of human diseases,
including Alzheimer disease, Parkinson disease and type II diabetes mellitus
(TTDM). In TTDM, over 90% of patients are found to have pancreatic amyloid
deposits upon autopsy. These deposits are primarily composed of a 37-residue
human islet amyloid polypeptide (hIAPP). Evidence suggests an association
between these amyloid plaques and pancreatic beta-cell dysfunction. Elucidating
the structure of these deposits and the effects modifiers have on the misfolding
pathway can help further the understanding of its toxicity and could be of use in
the design of drug inhibitors for amyloid diseases. I used site-directed spin-
labeling and electron paramagnetic resonance (EPR) spectroscopy to analyze
spin-labeled derivatives of hIAPP to determine structural features of the peptide
in its fibrillar form. Using continuous-wave EPR and four-pulse DEER coupled
with computational modeling, I determined a detailed structural model of hIAPP
fibrils. The N-terminal and C-terminal regions are less ordered and more mobile
and there is a turn region between residues 19-30 located between two β-strands
(13-18 and 31-36). My findings also show that the β1 and β2 strands from one
molecule of hIAPP are staggered in relation to one another. The staggered
peptide stacks directly on another staggered peptide, forming a protofilament
with a twist that I aptly call a β-spiral motif. In addition to these structural findings,
I have studied the effects of modifiers, such as curcumin and annexin proteins, to
the misfolding of amyloid fibrils. Using EPR, ThT and EM, I found that these
molecules or proteins seem to alter hIAPP, Aβ and α-synuclein fibril formation
x
and fibril morphology. Additionally, studies of these modifiers in tissue culture
and animal models showed reduced toxicity and protein aggregation. Both the
identification of hIAPP structure and the role of amyloid modifiers can give way to
therapeutic intervention of amyloid diseases.
1

Chapter 1: Introduction

1.1 Amyloid Proteins in Human Disease
The hallmark feature of diseases such as Alzheimer disease (AD),
Parkinson disease (PD) and type II diabetes mellitus (TTDM) is a pathological
accumulation of misfolded Amyloid-βeta (Aβ), α-synuclein and human islet
amyloid polypeptide (hIAPP), respectively. These proteinaceous deposits
accumulate in tissues and cause cytotoxicity to the surrounding cells (DeMager
PP 2002; Massimo Stefani 2003; Ross CA 2004). In fact, these amyloidogenic
deposits are not unique to just these diseases, but a number of human diseases
(Table 1). The table shows a myriad of different diseases involving different
organ systems that all result in the pathogenesis of misfolded protein plaques
made of amyloidogenic proteins. Although the primary structure of amyloidogenic
proteins involved in the aforementioned diseases differ and the organ in which
the toxicity occurs also differs, they share the propensity to undergo a multi-step
misfolding process in which the monomeric species changes to an oligomeric
form, and then forms protofibrils which ultimately come together to form fibrils
(Fig. 1A). It is not known what specific intermediate of the misfolding pathway
causes the toxicity, but there is increasing evidence that the intermediate
oligomeric species of misfolding likely plays an important role in its toxicity
(Butler, Janson et al. 2003; Butler, Jang et al. 2004; Ritzel, Meier et al. 2007;
Haataja, Gurlo et al. 2008) while the fibrillar state is seen at the end stage of
disease progression.
2

Table 1.1: Summary of Amyloidogenic Proteins and Associated Diseases

3
Fig. 1.1: Amyloid Fibril Assembly and structural features.
A) Amyloid fibril assembly takes a multi-step pathway. Monomeric peptides
become oligomers which then assemble to form protofilaments. Fibrils are
made up of several protofilaments. B) Structural features of IAPP on electron
microscope shows long, unbranched fibrils that are between 50-200 Å in
diameter. Each individual protofilament are made up of IAPP molecules that
are stacked on top of each at a distance of 4.7 Å in a parallel and in-register
fashion.
4
It is clear that amyloid fibril formation is a multi-step process with intermediate
structures. The pathogenesis of amyloid diseases involves the misfolding of the
protein conformations, thus it is crucial to derive detailed molecular
understanding of these proteins during the different stages of misfolding as it will
pave the way towards therapeutic intervention. We use site-directed spin labeling
and EPR spectroscopy together with other biophysical methods to provide an
understanding of the misfolding process. We generated a detailed structural
model of hIAPP fibrils (chapter 1). We then characterized the mechanism and
effect of inhibitors of amyloid misfolding (chapters 3-5).
Amyloid plaques are comprised of proteins that have similar morphological
characteristics. They have similar staining characteristics, such as the ability to
be stained by Congo Red dye and Thioflavin T (Puchtler 1961; Khurana R 2005).
Taking images with the technique of electron microscopy, people have shown
that amyloid fibrils also have a similar ultrastructure and morphology, consisting
of long, straight, unbranched fibrils approximately composed of several
protofilaments (Shirahama T 1967; Soto 2003). Cryo-electron microscopy and
Fourier-transform infrared spectroscopy revealed a repeat distance of 4.7 Å
which corresponds to the spacing between hydrogen-bonding β-strands (Serpell
LC 2000). X-ray diffraction studies of amyloid fibrils have shown a common
cross β structure that forms β strands that are perpendicular to the fibril axis (Fig.
1B) (Sumner Makin and Serpell 2004). Additionally, circular dichroism
spectroscopy revealed the conformational change during the process of fibril
formation is from random coil to β-sheet (Goldsbury C 2000). In the β-sheet
5
model, there is evidence of a highly ordered core region in amyloid proteins in
which individual β-strands are arranged in a parallel, in-register fashion (Fig. 1B)
(Jayasinghe and Langen 2004; Margittai and Langen 2006). These insoluble
amyloid fibrils can be as small as 50 Å in diameter, but lateral aggregation of
these protofilaments can increase the overall diameter of the fibril to several
microns (Goldsbury CS 1997). Fig. 1B shows a schematic of the common
structural features found in amyloid proteins.
One example of a disease involving the misfolding of amyloid proteins is
type II diabetes mellitus (TTDM), which is a progressive disease characterized by
a decline in the body’s ability to regulate blood glucose levels due to insulin
resistance, a progressive β-cell failure and a subsequent decline in insulin
secretion. Several factors can lead to the development of TTDM. Some of the
major factors include an environmental one in which increased caloric intake and
a sedentary lifestyle places pressure on the body to process the excess glucose,
leading to insulin resistance. There are also genetic forms of the disease, one
example being a mutation in hIAPP that causes early-onset TTDM. Lastly, the
factor of primary interest for my studies is an endogenous deterioration caused
by islet amyloid polypeptide (hIAPP) deposition in the pancreatic β-cells. It is
unlikely that only a single one of these factors will lead to the progression of
TTDM. Instead, it is thought that TTDM is a multi-factorial disease with each
factor (environment, endogenous and genetic) being a part of the overall disease
progression.

6
1.2 Islet Amyloid Polypeptide: Functional Features
The pancreatic deposits associated with TTDM are primarily composed of
the 37-residue, human islet amyloid polypeptide (hIAPP), which is secreted by
the β-cells in the pancreatic islets of Langerhans (Westermark, Wernstedt et al.
1987; Westermark, Wilander et al. 1987; Clark, de Koning et al. 1995). hIAPP is
normally co-expressed and co-secreted with insulin and is thought to play a role
in carbohydrate metabolism (Butler 1990; Kahn, D'Alessio et al. 1990; Ahren,
Oosterwijk et al. 1998; Gebre-Medhin, Mulder et al. 1998) and gastric emptying
(Young 1995). While hIAPP in its physiological state is monomeric, it is found as
a misfolded and aggregated protein in the diseased state. In fact, more than
95% of TTDM patients exhibit extracellular amyloid deposits around the
pancreatic β-cells (Clark, Cooper et al. 1987). Evidence found in tissue culture
(Butler, Janson et al. 2003) and animal models of hIAPP in mice (de Koning,
Hoppener et al. 1994; Janson, Soeller et al. 1996; MacArthur, de Koning et al.
1999), rat (Butler, Jang et al. 2004), diabetic felines (Johnson 1973; O'Brien
2002) and the diabetic monkey, Macaca nigra, (Howard 1986) suggests an
association between hIAPP misfolding and pancreatic beta-cell dysfunction
(Cooper, Willis et al. 1987; Rocken, Linke et al. 1992; Lorenzo, Razzaboni et al.
1994; Clark, de Koning et al. 1995; Kahn, Andrikopoulos et al. 1999). Figure
1.2A shows an illustration of a pancreas, focusing specifically on the histology of
the islets of Langerhans. The islets contain the endocrine (i.e. hormone-
secreting) cells. The βeta cells of the islets secrete insulin and hIAPP, thus the
7
Fig. 1.2 Illustration of IAPP amyloid pathology.
A) Cartoon of pancreas showing an islet of Langerhan containing
alpha, beta and delta cells responsible for the endocrinological
function of the pancreas. B) Pathology slide of an islet of Langerhan
from a T2DM patient showing extensive amyloidosis (pink amorphous
material) and a decrease in Beta cells. C) Electron Microscope image
of the main constituent of pancreatic amyloid plaques, hIAPP fibrils.
The image shows morphologically characteristic appearance of
unbranched, twisting fibrils.
8
destruction of these cells can lead to severe deficiency in these hormones,
thereby leading to the diabetic state mentioned above. Figure 1.2B shows a
pathology slide from a patient with TTDM. One can see large areas of
amorphous pink amyloid within the islet along with a decrease in βeta cells. A
major component of these amyloid plaques is hIAPP. Electron Microscopy of
hIAPP within these plaques shows long, twisted and unbranched fibrils (Fig.
1.2C).
The pathogenesis of amyloid diseases involves aberrant protein
conformation, thus it is very important to understand the molecular details and
structure of these conformations. The information garnered from the structural
details of these protein conformations can help with the discovery and design of
therapeutic agents that can inhibit these toxic species. While it is unclear exactly
what the toxic species is in the misfolding pathway, it is clear that more structural
information about all these intermediates is needed. Chapter 2 focuses
specifically on the structure of hIAPP fibrils, while chapters 3-5 discuss how small
molecules and proteins are capable of altering amyloid misfolding.

1.3 Islet Amyloid Polypeptide: Structural Features
Aside from the common characteristics found in all amyloid fibrils as
discussed earlier, more detailed analyses of the hIAPP fibril structure has been
done using biophysical methods, such as EPR, NMR and cryo-EM. Analysis of
the EPR spin mobility and comparison to the amyloid found in Alzheimer’s
patients (Aβ) indicates a parallel β-sheet structure with a high degree of order
9
throughout the fibrillar peptide, but the N-terminal and C-terminal regions are less
ordered, more mobile and not part of the ordered core region that gives rise to
fibril formation (Jayasinghe and Langen 2004). Other studies done on the
structural features of hIAPP fibrils suggest a relative closeness in aromatic side-
groups, specifically Phe15 and/or Phe23 to Tyr37 that may help create a
hydrophobic core (Padrick and Miranker 2001; Jack, Newsome et al. 2006). With
respect to mass per unit length, Goldsbury et al show it to be 10kD/nm, which
corresponds to 2.6 molecules of hIAPP packed into 10 Å of protofilament
(Goldsbury CS 1997). These structural features are important to take into
consideration when finding the 3-D structure of hIAPP and will be further
discussed in Chapter 2.
More defined, but theoretical models of hIAPP fibrils have also been
suggested. The most recent proposed model by Luca et al uses solid-state NMR
and electron microscopy to identify hIAPP peptide conformations and structural
models. Some key outcomes from this study showed that in a morphologically
homogenous fibril sample, residues 8-13 have greater amplitude of motion than
residues 14-37. They also show residues 18-27 is a bend between two β-strands,
8-17 and 28-37 (FIG. 1.3A) (Luca, Yau et al. 2007). A different model by Kajava
and Steven et al used computational methods to show hIAPP fibrils as a planar
S-shaped fold with three β-strands, stacked in register (Fig. 1.3B) (Kajava AV
2005). They hypothesize that residues 20-27 and 30-37 are parallel and residues
30-37 and 8-16 are antiparallel.
10
Fig. 1.3: Illustration of proposed models of IAPP fibrils.
A) Luca and Tycko et al show the IAPP model for the
protofilament. In A.1there is a ribbon representation of one cross-
Beta-molecular layer, with – and C-terminal strands in red and
blue, respectively. A.2 shows the cross sectional view of two IAPP
molecules in the protofilament. B) Kajava and Steven et al
propose the Beta-serpentine fold for IAPP as in B.1 and the ball
and stick model as in B.2 C) Riek et al shows the Amyloid-Beta
model.
11
An Amyloid-βeta (Aβ) model by Riek et al has often been used to draw
comparisons to what the hIAPP fibril model is expected to be. In this model, they
use NMR, cryo-electron microscopy and computational modeling to show that Aβ
residues, 18-42, forms a β-turn-β motif with two parallel in-register β-sheets (Fig.
1.3C) (Lührs T 2005).
While these models propose structural information about hIAPP fibrils,
they are all preliminary and some are derived solely with computational methods,
which make them more theoretical. They also do not address detailed structural
motifs, such as the exact locations of the β-sheets, their orientation towards one
another and the intermolecular distances between residues that can help create
a detailed three-dimensional structure of the fibril. Chapter 2 of this thesis
elucidates the details of the hIAPP fibril structure by answering the following
questions: (1) What regions of hIAPP form the core of the fibril and how do those
regions interact with one another? (2) Where are the proposed β-sheet, turn and
loop regions and how do they come into contact with one another? (3) What are
the intramolecular distances between amino acid residues of one molecule of
hIAPP? (4) What is the exact size of the cross-β core region of a protofilament
and how does it relate to the overall size and morphology of a fibril?

1.4 Modifiers of Amyloid Misfolding
The finding that misfolded proteins cause a large number of human
diseases has prompted the study and discovery of modifiers of the misfolding
pathways. By modifying the pathway, perhaps one can avert the formation of
12
toxic species or discover physiologic molecules that expedite such toxic
pathways. Examples of such modifiers include β-sheet breakers (Wisniewski T
2008) proteosome activators (Casas, Gomis et al. 2007) that help induce the
ubiquitin-proteosome pathway, heat shock proteins (Wilhelmus MM 2006; Luo W
2008) that interact with incorrectly folded proteins and small molecule inhibitors
of amyloid misfolding, such as curcumin and rifampicin (Tomiyama, Kaneko et al.
1997; Lim, Chu et al. 2001; Ono and Yamada 2006). The modifiers of misfolding
that are of interest to my studies are annexin A5 protein, an endogenous protein,
and curcumin, a small molecule. While detailed discussions of both of these
modifiers are in Chapters 3-5, they are briefly introduced below.
A modifier of misfolding in particular interest to my study is annexin A5
protein. The annexins are a group of proteins traditionally thought of as calcium-
dependent phospholipid-binding proteins. They have a wide range of cellular
functions including association with components of cytoskeletal proteins or
extracellular matrix, inflammatory response, or neovascularization (Rescher U
2004; Gavins FN 2005; Gerke V 2005; Sharma MR 2006). There are several
annexin proteins with annexin A5 being one of them. It plays several roles,
including forming a shield around certain phospholipid molecules that blocks their
entry into blood coagulation reactions. The lack of proper annexin A5 protein
leads to a condition called anti-phospholipid syndrome, which quickens the
coagulation cascade and is known to be a leading cause of diagnosed
miscarriages in pregnant women. Another important role is its ability to bind
phosphatidylserine and therefore be an apoptotic marker (Koopman G 1994).
13
Additionally, annexin A5 interacts with a number of proteins, including collagen
type 2, vascular endothelial growth factor, integrin B5, G-actin and helicase
(Moss SE 2004).
In addition to the wide range of cellular functions, it has also been shown
that annexin A5 has a protective effect on Aβ toxicity (Lee G 2002). The study
done by Lee et al hypothesized that annexin A5 protects from Aβ toxicity by
competitively inhibiting Aβ interaction on phospholipid. In this study, they
explained that the cellular target of Aβ peptide is phosphatidylserine (PS) and
annexin A5 is a ligand for PS that blocks Aβ from binding it. This inhibition of Aβ
binding thus prevents its cytotoxicity. I took this work a step further in chapters 3-
4 by studying how annexin A5 affects the misfolding of hIAPP, Aβ and α-
synuclein using both biophysical and biological methods. Additionally, we studied
the in vivo effect of overexpressing annexin A5 in the presence of the α-synuclein
phenotype, using transgenic C. elegans.
The other modifier of interest is curcumin (diferulomethane), which is a
biphenolic small molecule and the main constituent of the rhizome C. longa
(turmeric). It is known to have many therapeutic effects including anti-
inflammatory, anti-oxidant and anti-HIV effects (Ammon HP 1991; Mazumder A
1995; Aggarwal BB 2003). Recent studies have also shown that curcumin has
anti-amyloidogenic effects on Aβ (Yang, Lim et al. 2005; Garcia-Alloza, Borrelli et
al. 2007), α-synuclein (Pandey, Strider et al. 2008) and prion proteins (Hafner-
Bratkovic, Gaspersic et al. 2008) as it relates to Alzheimer disease, Parkinson
disease and transmissible spongiform encephalopathies, respectively. While
14
these studies have shown the anti-amyloidogenic effects of curcumin, it is still
poorly understood how curcumin interacts with amyloid proteins and what
mechanism it uses to inhibit misfolding. This prompted us to investigate the
molecular interaction of curcumin with hIAPP and test its effect on β-cells
(chapter 5).

1.5 EPR and DEER: Method for studying hIAPP structural features

The field of structural biology employs many different techniques for the
structural analysis of proteins, such as X-ray crystallography, NMR and cryo-
electron microscopy. These conventional methods, however, have proven
difficult in studying misfolded amyloid diseases because of their insoluble nature
and their difficulty in forming crystals. Thus, our lab performs site-directed spin
labeling (SDSL) and Electron Paramagnetic Resonance (EPR) because this
technique does not require the formation of crystals and is not limited to a
specific protein size or solubility.
Overall EPR is a technique for studying chemical species with one or
more unpaired electrons. Using EPR coupled with SDSL, information about the
structure, conformational changes and environment of a protein can be attained.
It has been used to determine the structure of globular, membrane and amyloid
proteins (Hubbell WL 1998). The EPR method involves mutating a residue of
interest to a cysteine and labeling that residue with a nitroxide spin label via a
disulfide bond (Fig. 1.4A) to form a side chain R1 (all mutated proteins are
denoted as R1 along with the residue number to which it is attached). Many
15
studies have shown that the addition of the spin label onto the protein does not
perturb the protein structure (Mchaourab HS 1996; Hubbell WL 1998). We
confirm this by employing various methods such as seeding with wildtype protein
and taking EM images to confirm the presence of the desired structure. The
derivitized protein is analyzed by EPR and the resulting spectrum gives
information about the mobility and environment of the spin label (Fig. 1.4B).
Mobile sites (i.e. loop sites) have sharp lines, large amplitudes, and narrow
central line widths (Fig. 1.4B). Less mobile sites, such as helix surface sites and
tertiary contact sites, have intermediate amplitudes and broad central lines.
Buried sites have decreased amplitudes and very broad central lines.
Measurement of the central line width can provide a mobility parameter that
describes whether that particular residue is in the interior, more buried regions or
in the highly mobile loop regions (Fig. 1.4B) (Mchaourab HS 1996). To further
characterize whether particular regions are more buried or more towards the
exterior of the protein structure, oxygen accessibility assays can be used. We
can measure the accessibilities (Π) of the R1 sidechain to the paramagnetic
collider, O
2.
For soluble proteins, oxygen accessibility is used to obtain
information about which amino acids are more accessible to oxygen, and thus
more surface-exposed. If the Π value is high, the residue is in more contact with
oxygen which means it is more likely on a more outer portion of the β-sheet (C
Altenbach 1994).
Double Electron Electron Resonance (DEER) spectroscopy is another
technique used to describe the structure of proteins and to map their
16
Fig. 1.4: Electron Paramagnetic Resonance
A) MTSL nitroxide spin label used in site-directed spin labeling and EPR.
B) Illustration of EPR spectra readouts for residues having various
mobilities. The central line width indicated by the red lines is a parameter
for residue mobility.
17
intramolecular distances. This technique utilizes the site-directed incorporation of
two R1 groups into a protein in order to estimate the inter-residue distances
through magnetic interactions between the spin labels. It can measure spin-spin
interactions between spin labels in the 15-80 Å range, nicely complementing the
8-20 Å range of continuous wave (CW) experiments. The interaction of the two
lone electrons on these labels will give information about the distance between
the two residues. By labeling appropriate residues, one can map out the
distances of all the residues in the peptide. We will be using this technique along
with mobility and accessibility assays to formulate a 3-dimensional model of
hIAPP fibrils, as described in Chapter 2.

1.6 C. elegans: Method for studying Annexin V and amyloids in vivo
Transgenic Caenorhabditis elegans models have been established for a
number of age-associated and neurodegenerative diseases, including
Alzheimer’s (CD 1995) and Parkinson’s (Lakso M 2003) disease. The C. elegans
model has many advantages including the ability to undertake genetic screens to
identify genes involved in pathological processes. Another advantage of C.
elegans is simply their short lifespan that allows for a quick assessment of
genetic interventions, especially age-related processes that take significantly
longer in mouse models. Furthermore, C. elegans are multi-cellular organisms,
yet simple enough to study genetic interventions in detail. For example, knowing
this organism has 302 neurons and knowing the exact fate of these neurons, it is
convenient to map transgenic changes in these cells. The ability to visualize GFP
18
expression in living C. elegans can offer another interesting approach to
monitoring the effects of transgenic expression.
As mentioned previously, hIAPP is a peptide that is released from and
plays a role in the endocrine system of humans. While the C. elegans lack an
endocrine system, they do contain a neural system whose biology is likened to
that of an endocrine system, making it an optimal choice for studying secretion,
signaling, etc. Thus, the introduction of hIAPP into this system can give us
information about secretion of the peptide out of the cells, localization in cells and
changes in behavior with over-expression. There have not been studies involving
C. elegans and hIAPP, but there have been some studies showing the effects of
both Aβ and α-synuclein. Immunohistochemical analysis shows when transgenic
C. elegans were engineered to express Aβ, these animals produced muscle-
specific deposits of the protein that were reactive with anti-Aβ antibodies and
stained with amyloid-specific Congo Red and thioflavin T dyes (Fig. 1.5, white
arrows) (CD 1995; Wu Y 2005). There was also a phenotypic result showing a
progressive paralysis, indicating that amyloid-β shows toxic effects in muscle
cells of C. elegans (Koopman G 1994). Overexpression of α-synuclein with a
pan-neuronal promoter in the C. elegans model also showed significant motor
deficits (Lakso M 2003). This not only shows the toxicity of such proteins, but
also shows the promise of using C. elegans as an animal model for studying the
toxicity of amyloid proteins.
19
Fig.1.5: Schematic of chimeric transgene construction in
C. Elegans
A chimeric transgene is created using a C. elegans promoter
and an artifical signal peptide attached to b-amyloid peptide.
The transgene is injected into the C. elegans gonad.
Immunohistochemical analysis shows the presence of β-
amyloid overexpression.
20
In our studies, we expressed human α-synuclein fused to yellow fluorescent
protein (YFP) in C. elegans under control of the Unc-54 promoter, which drives
expression to the body wall muscle cells. Muscle expression rather than neuronal
expression was chosen for several reasons. This particular promoter is strong
and muscle cells are large, allowing for visual detection of α-synuclein expression
and its sub-cellular localization. Additionally, muscle expression has been used
successfully in many previous studies to model protein misfolding diseases (Link
1995; Morley 2002; Nollen 2004). Chapter 4 of this thesis discusses how C.
elegans were used as a model for α-synuclein cytotoxicity and what in vivo role
annexin A5 has in inhibiting the aggregation of α-synuclein.
In the series of studies presented in this dissertation, I have attempted to
explain more about the amyloid misfolding pathway, with a specific focus on
hIAPP as it relates to TTDM. I elucidate key structural features and a detailed
three-dimensional model of hIAPP fibrils, a characteristic and final component of
amyloid misfolding, using EPR, EM and computational modeling (Chapter 2). I
then discuss how modifiers of amyloid misfolding can alter the pathway. I focus
on how annexin A5 protein alters hIAPP (chapter 3), Aβ and α-synuclein fibril
formation and cytotoxicity in C. elegans (Chapter 4). Finally, I sought to
determine if curcumin alters the misfolding and toxicity of hIAPP, as it relates to
TTDM, and whether it does so at concentrations that are non-toxic to pancreatic
β- cells (Chapter 5).

21
Chapter 2: Identifying the three-dimensional structure of
hIAPP fibrils
2.1 Introduction
In type 2 diabetes mellitus (T2DM), the amyloid deposits found in the
pancreas of 95% of patients are primarily composed of the misfolded 37-residue
human islet amyloid polypeptide (hIAPP) (Clark, Cooper et al. 1987). hIAPP is
co-expressed and co-secreted with insulin by the β-cells in the pancreatic islets
of Langerhans (Westermark, Wernstedt et al. 1987; Westermark, Wilander et al.
1987; Clark, de Koning et al. 1995). Its functions have not been fully elucidated
but it is thought to play a role in carbohydrate metabolism (Butler 1990; Kahn,
D'Alessio et al. 1990; Ahren, Oosterwijk et al. 1998; Gebre-Medhin, Mulder et al.
1998) and gastric emptying (Young 1995). While hIAPP in its physiological state
is monomeric, it is found as a misfolded and aggregated protein in the diseased
state.
There is a propensity for hIAPP, as well as other amyloid proteins, to
undergo a multi-step misfolding process in which the monomeric species
changes to an oligomeric form and then ultimately forms fibrils. Studying the
fibrillar form of hIAPP is important to understanding the overall process of
amyloid formation. While a complete three-dimensional structure of IAPP fibrils
has yet to be identified, several approaches have yielded information about the
structural features of hIAPP. Fourier transform infrared and circular dichroism
showed hIAPP fibrils have β-sheet structure (Kayed, Bernhagen et al. 1999;
22
Higham, Jaikaran et al. 2000). X-ray and electron diffraction studies concluded
that the fibrils are 4.7 Å apart and make up cross-β strands that are
perpendicular to the fibril axis (Sumner Makin and Serpell 2004).
Previous SDSL studies on hIAPP have shown that residue mobility of
spin-diluted fibrils is significantly lower than the mobility of hIAPP as a random
coil monomer in solution (Jayasinghe and Langen 2004). Figure 2.1 shows the
spectra of both monomeric (black line) and fibrillar (green line) hIAPP. The three
sharp lines are indicative of fast motions and shows that freshly dissolved hIAPP
adopts a random coil structure in solution. After a three day incubation period,
the monomers form fibrils and gives rise to spectral broadening and low-signal
amplitude. Analysis of the spin mobility and comparison to the amyloid found in
Alzheimer’s patients (Aβ) indicates a high degree of order throughout the fibrillar
peptide, but the N-terminal and C-terminal regions are less ordered and more
mobile (Jayasinghe and Langen 2004). Evidence suggests that these regions are
not part of the ordered core region that gives rise to fibril formation. This work
done by Jayasinghe et al gave us insight into the ordered, structured features of
hIAPP fibrils.
In the present study I expanded on this structural information by using
electron paramagnetic resonance (EPR) coupled with site-directed spin-labeling
(SDSL) to identify detailed structural features of hIAPP fibrils. In SDSL, individual
cysteine mutants are made at each residue of hIAPP selected for study and a
23
Fig. 2.1: EPR spectra of hIAPP 21 R1
EPR spectra of hIAPP 21R1 in its soluble (black line) and fibrillar (red
line) states. For comparison, each spectrum was normalized to
represent the same number fo spins. To visualize line shapes, the
fibrillar spectrum is re-plotted at 10X magnification (green line). The
solution spectrum is characteristic of a highly mobile R1 side chain;
implying a high degree of disorder and motion of hIAPP in solution. The
fibrillar spectrum contained spectral lines that were significantly broader
and of much lower amplitude.
24
nitroxide spin label is introduced at those sites, resulting in a nitroxide-labeled
side chain (R1). We used continuous-wave EPR to obtain information about the
local environment of the spin label, such as its mobility and oxygen accessibility.
We then coupled this information with double-mutant distance data from double
electron-electron resonance (DEER) and four-pulse EPR to map a three-
dimensional model of hIAPP fibrils. The EPR and EM data are then used as
constraints for simulated annealing molecular dynamics modeling.

2.2 Materials and Methods
2.2.1 Chemicals and Peptides
Hexafluoroisopropanol (HFIP) was obtained from Sigma-Aldrich. Synthetic
wild-type human

IAPP was obtained from Bachem Bioscience Inc. (King of
Prussia, PA). The spin

label, 1-oxyl-2,2,5,5-tetramethyl- 3-pyrroline-3-methyl
methanethiosulfonate

(MTSL), was obtained from Toronto Research Chemicals
(Toronto,

Canada). Single cysteine and double cysteine mutants of full-length
hIAPP were purchased from Biomer Technology (Pleasanton, CA). The native
cysteine residues at position 2 and 7 are substituted with an alanine (Jayasinghe
and Langen 2004). Peptides arrived 90% purified and lyophilized and were
stored at -80 ° C until use.

2.2.2 Spin Labeling
For spin labeling, purified

peptides were reacted with 3-fold molar excess
of MTSL for 1

h at room temperature. Unreacted MTSL was removed using a
25
Toyopearl cation exchange column. The spin-labeled peptide was then desalted
using a C18 reverse phase SpinColumn from Harvard Apparatus (Holliston,
Massachusetts) and eluted with 100% HFIP. Peptide concentrations were
calculated by UV absorbance

at 280 nm in 6 M guanidine HCl using an extinction
coefficient

of 1400 M
-1
cm
-1
. Labeled peptides were stored at 4 ° C until use.

2.2.3 hIAPP Fibril Formation
Stock solutions of spin-labeled hIAPP and wildtype hIAPP were
lyophilized. For preparation of stock solutions the lyophilized peptides were
reconstituted with deionized water containing 0.5 % acetic acid. hIAPP fibrils
were formed in a spin-diluted state (Jayasinghe and Langen 2004) by mixing
stock solutions of spin-labeled hIAPP with unlabeled hIAPP wildtype peptide (1:4
ratio) in 1.5 mL eppendorf tubes. For fibril formation, aliquots of stock hIAPP
were added to 10mM Phosphate buffer with 100mM NaCl (pH 7.4) for a total
volume of 100uL and 100uM concentration. To encourage the growth of amyloid
fibrils, WT hIAPP fibrils that were grown for one week and then sonicated on ice
in 1 minute intervals using a tip sonicator for 10 minutes. The sonicated seeds
were then added to each spin-labeled sample in a 10.0 wt %. The samples were
incubated at room temperature for 3-7 days to allow for fibril formation.

2.2.4 Electron Microscopy
To confirm or examine the presence of fibril growth, 10uL of hIAPP fibril
samples were adsorbed onto carbon and formavar-coated copper grids and
26
negatively stained with 2% (w/v) uranyl acetate solution for 5 min. The stained
grids were examined and photographed using a JEOL JEM-1400 electron
microscope at 80-100 kV.

2.2.5 EPR spectroscopy
For EPR Spectroscopy, fibril samples were pippetted into a glass capillary
(0.6 mm inner diameter, 0.84 outer diameter, VitroCom, mt. Lks., New Jersey)
that had been sealed at one end. The sample was centrifuged and all
supernatant was removed to eliminate smaller aggregates and free spin label.
EPR spectra of fibrils were recorded on a Bruker (Billerica, MA) EMX
spectrometer with a HS resonator at 12 mW incident microwave power with a
scan range of 150. Mobility was calculated using the inverse of the central line
width, which is the distance between the peak of the central line and the trough
of the third line. The O
2
accessibility was obtained using a power saturation
method (Borbat P 2006) in the presence of ambient oxygen.

2.2.6 Pulsed EPR and Distance Analysis
The intramolecular distance determinations were made using four-pulse
DEER experiments. Fibril samples were prepared by diluting 3% fully spin-
labeled hIAPP, containing two spin labels per protein, was mixed with unlabeled
wild-type protein and seeded with 10.0 wt % sonicated wild-type hIAPP fibrils.
Samples (300uL) were then incubated at room temperature for 3-7 days to allow
for fibril formation. Samples were ultra-centrifuged in order to remove the
27
supernatant containing smaller aggregates from fibril pellet and collected into a
quartz capillary (1.5mm inner diameter, 1.8 mm outer diameter). The pellet was
flash-frozen and data were acquired at 78K. Four-pulse DEER experiments were
performed using a Bruker Elexsys E580 X-band pulse EPR spectrometer fitted
with a 3-mm split ring (MS-3) resonator, a continuous-flow helium cryostat
(CF935, Oxford Instruments), and a temperature controller (ITC503S, Oxford
Instruments). The data were fit using Tikhonov regularization (Chiang YW 2005)
as implemented in DEERAnalysis2008 packages (Jeschke G 2006).

2.2.7 Molecular Modeling
A peptide of the wildtype hIAPP sequence was built using the algorithm
PRONOX. The initial peptide was constructed as a linear beta sheet (φ = -120.0° ,
ψ = 120.0° , ω = 180.0° ). Fourteen spin labels were added with torsi on angles t1
(Sγ-Cβ-Cα-N) = -60° and t2 (Sδ-Sγ-Cβ-Cα) = 180° . After a series of experimental
variations and trials, we used PRONOX to construct 2 separate sets of β-sheets
containing 48 β-strands in each sheet (Fig. 2.8A). The β-strand of sheet 1
contained residues 12-18 and the β-strand of sheet 2 contained residues 31-36.
Every strand contained spin-labeled residues. The loop region was disconnected
initially between I26 and I27 in order to allow the strands the freedom to move in
the SAMD simulation. Experimental data for 14 inter-label distances were used to
define restraints for the SAMD calculations. Using PRONOX, these data were
converted into an appropriate format for use as restraints in AMBER8. Interlabel
distances were defined between the N atoms of each label. The SAMD
28
calculations were carried out in AMBER8 with a parm98 force field. After a brief
minimization of the starting structure (Fig. 2.8A) SAMD was performed in cycles
of 60 ps, including a heating phase from 0K to 1000K in 4 ps using a step of
0.002 ps, during which the force constants for the restraints were increased from
0.1 to 10.0; maintenance of the temperature at 1000K for 10 ps; and then cooling
to 0K over 20 ps. This approach was based on the standard recommended
protocol for simulated annealing calculations in the AMBER8 manual (Pearlman
D 1995). After this initial SAMD cycle, the two strands were connected and the
spin labels were replaced with the original amino acid residues. This structure
then underwent SAMD with the same conditions as above.

2.3 Results
2.3.1 EPR spectra of spin-diluted hIAPP fibrils
Our objective was to obtain information about the local environment of
each residue of hIAPP using site-directed spin labeling and continuous-wave
EPR spectroscopy. In order to obtain this information I generated 28 singly-
labeled hIAPP derivatives and used nitroxide scanning to evaluate the relative
mobility of each residue of hIAPP in its fibrillar form. In accordance with recent
studies showing the polymorphic nature of hIAPP fibrils (Luca, Yau et al. 2007), I
generated homogenous fibrils of the “twisted” morphology by seeding with
homogenous wildtype hIAPP and creating an environment that is conducive to
the formation of these types of fibrils (see methods section). EM images of all
29
Fig. 2.2: Electron microscope image of hIAPP fibrils
A) EM image of spin-labeled hIAPP fibrils.The image illustrates the
homogenous “twisted” morphology of hIAPP fibrils obtained by growing
them in the presence of wildtype monomer and wildtype seeds at room
temperature without agitation. This is a representative example of the
fibril morphology obtained for all the continuous-wave and four-pulse
DEER experiments. B)Histogram of hIAPP fibril pitch as measured from
EM.
A
B

30
mutants were taken to confirm morphology (Fig. 2.2A). We also observed the
average pitch of the fibril (Fig. 2.2B) as recent studies have also shown the
average pitch of fibrils twisting around one another to be approximately 150-560Å
(Goldsbury CS 1997; Malinchik 1998; Goldsbury and Aebi 2000). Figure 2.2 not
only confirms that our fibrils are of the homogenous “twisted” morphology but
also shows that the average distance of the pitch of twisted fibrils falls within
normal limits of previously studied fibrils. We used a spin dilution approach (1:4
labled:wildtype hIAPP) in order to decrease the effects of spin-spin interactions
seen in fully labeled samples. In the absence of spin-spin interactions, EPR line
shapes provide information on the dynamics of R1, allowing us to focus on the
mobility of each residue. Figure 2.3 shows the spectra of all the spin-diluted
samples used for the mobility analysis. The spectra show significant differences
between spin-labeled residues, specifically in the N-terminal region, residues 2-8.
The spectra for these residues show pronounced changes from the more mobile
“core residues” as indicated by the larger signal amplitude and the less narrow
spectral lines, suggestive of a less immobilized region. This same spectral
pattern can also be seen in the C-terminal residues, albeit to a lesser extent.
Additionally, it is evident that the residues between residues 10-30 show
decreased amplitude and broad spectral lines that are seen in more immobile
residues.

31
Fig. 2.3: EPR spectra of hIAPP fibrils containing single labels at
positions 2-36.
Spectra from fibrils labeled with a mixture of 25% R1 and 75% wildtype
peptide. All spectra were obtained using a scan width of 150 Gauss and
were normalized to the same number of spins.
32
2.3.2 Structural Features of hIAPP from R1 mobility
The spectra for all of the spin-labeled residues yielded a mobility
parameter, expressed by the inverse line width (ΔH
o
-1
), shown in Fig. 2.4. It is
evident from this graph that the N-terminal region, residues 2R1-10R1 (Fig. 2.4,
boxed region), show very high mobility indicative of a region that is more
unstructured and in a random coil. It is also clear that while the C-terminal region,
residues 31-36 shows an overall higher mobility, the periodicity between the
residues suggests a structured region. The periodicity pattern of two between
residues 31-36 specifically suggests a β-sheet pattern in which the residues on
the exterior of the protein have a higher mobility (32R1, 34R1 and 36R1) and
residues on the inside have a lower mobility (31R1, 33R1 and 35R1). The higher
overall mobility may suggest that while it remains in a structured conformation,
this overall region might be more exposed and mobile. This same β-sheet pattern
of periodicity is also seen between residues 14-18 with residues 14R1, 16R1 and
18R1 being more mobile and 15R1 and 17R1 being more immobile or on the
inside of the β-sheet. The residues in the region between residues 20-28 do not
seem to show a definitive periodicity pattern. The mobility data in our studies as
well as previous studies (Kayed, Bernhagen et al. 1999; Higham, Jaikaran et al.
2000) indicate there are areas of β-sheet content in hIAPP fibrils.

2.3.3 Intramolecular Distances from Four-Pulse DEER Experiments
The mobility and accessibility data obtained from continuous-wave EPR
gave insight into some of the general structural features of hIAPP fibrils, but it
33
Fig. 2.4: Graph of hIAPP residue mobility
Mobilities of fibrils containing 25% R1 as calculated from the
inverse of the central line widths of EPR spectra (1/dHPP) is
represented (black dots).
34
does not provide enough information to create a detailed three-dimensional
model. To further our endeavors towards a detailed three-dimensional structure,
we used four-pulse DEER to measure the interlabel distances in 21 doubly-
labeled hIAPP derivatives. Samples were grown from a wildtype hIAPP seed.
The labeled samples were diluted with 98% wildtype hIAPP. Thus, the final
fibrillar sample contained 2% labeled peptide. EM images confirmed a
homogenously “twisted” morphology and continuous-wave EPR showed
complete mixing of the wildtype peptide with the labeled peptide.
Four-pulse DEER monitors the time evolution of a spin echo intensity
(Pannier M 2000). The frequency of the resulting signal is a measurement of the
distance between two spin labels with an upper distance limit of ~60 Å. The data
were fit using Tikhonov regularization (Jeschke G 2006). Distance information
from the dipolar time evolution data was generated using the DEERAnalysis2008
package. In order to explain the process of data analysis done on the DEER
samples one example from the 21 doubly-labeled peptides is shown in Fig. 2.5.
Using hIAPP 13R1/19R1 double mutant sample as an example, the first piece of
information that is obtained is the time evolution of the spin echo without
background correction (Fig. 2.5A). The background contribution from non-specific
contacts is subtracted using a three-dimensional model of hIAPP fibrils (Fig.
2.5B). The time domain signal is Fourier transformed into a frequency domain
that is depicted by the pake pattern seen in Fig. 2.5C. From there, the frequency
range is input into the equation in (Fig. 2.5E) in order to obtain a distance

35
Fig. 2.5: Explanation of DEER data analysis
Using hIAPP13R1/19R1 as an example, the first piece of information that is
obtained is the time evolution of the spin echo without background correction
(A). The background contribution from non-specific contacts is subtracted
using a three-dimensional model of hIAPP fibrils (B). The time domain signal is
Fourier transformed into a frequency domain that is depicted by the pake
pattern seen in (C). The distance distribution (D) between the two double
mutants in obtained by inputting the frequency range into equation (E).
36
Fig. 2.6: Intramolecular distances from 4-poulse DEER
experiments.
The left panels show the dipolar evolution time for each of the indicated
double-labeled derivatives of hIAPP fibrils. The black traces are
background-corrected experimental data and the red lines represent the
results of the fits using Tikhonov regularization. The right panels shows
the resulting distance distributions, whose peaks are tabulated in Table
2.1.
37
Table 2.1: Comparison of intramolecular distances
of four-pulsed DEER and models
The experimental distances are taken from the peaks
of the Tikhonov regularization-based fits.
38
This analysis was performed for all doubly-labeled residues of hIAPP fibrils (Fig.
2.6) and the fitted distances as well as the predicted distances from the models
distribution (Fig. 2.5D) between the two labeled residues in hIAPP 13R1/19R1.
discussed below are shown in Table 2.1.

2.3.4 Structural Refinement via Computational Modeling
In order to generate a three-dimensional structure of hIAPP fibrils we used
a computational approach for structural refinement based on our EPR data.
Before the computational results are discussed, I will present a brief review of the
theory of simulated annealing and molecular dynamics as it applies to modeling
of hIAPP fibrils. Simulated annealing molecular dynamics (SAMD) is a
computational method that optimizes the simultaneous agreement of an atomic
model with experimental data and with chemical information (Brunger AT 2002).
It is a method that combines force fields of the internal coordinates of a molecule
(i.e. bond length, torsion angle, etc) with the non-bonded interatomic interactions.
Molecular dynamics (MD) is a form of computational simulation that uses the
generalized coordinates (x, y, and z Cartesian positions of the atoms of a
molecule) and statistical mechanics to obtain a detailed resolution of how the
atoms are interacting with one another. The basic theory behind molecular
dynamics includes a series of equations and algorithms that depend on
calculating force fields and solving equations of motion based on accelerations
obtained from the new force. To begin with, the translational motion of a
spherical molecule is caused by force (F) exerted by some external agent. This is
39
illustrated in Newton’s second law shows that force (F) is equal to mass (m)
times acceleration (a):

F= ma

In every mechanical simulation, the intermolecular potential energy among N
atoms is a sum of isolated contribution between the atoms and is calculated by :

Additionally, molecular dynamics simulation includes a calculation of the forces
on the atoms, as described by the Lennard-Jones potential:

This potential is used to describe the interaction between two atoms which are
not chemically bonded to one another. This equation describes that there is an
attractive force at long distances (1/r)
12
and a repulsive force at short distances
(1/r)
6
. These potential functions are formulated as a sum over interactions
between atoms of the system. For our modeling system, we used AMBER8 with
PARM98 force fields that account for molecular interactions (i.e. Van Der Waals)
40
of the molecule along with DEER experimental distance constraints that force the
system into the most correct conformation.
Simulated annealing (SA) is a process in which a molecule is heated until
all particles are randomly arranged and then allowed to cool in order for all the
particles to arrange themselves in the lowest energy state (E). In terms of
searching for structural conformations, one uses SAMD to search for the
conformations of a molecule that best fit the experimental data and that also
maintain appropriate covalent and non-covalent interactions at the lowest energy
state.

2.3.5 Structural Models of hIAPP fibrils
We collaborated with the laboratory of Dr. Ian Haworth to develop a
computational approach that uses our EPR data to build a three-dimensional
structure of hIAPP fibrils. In our initial approach, we used PRONOX program to
generate starting structures of spin-labeled hIAPP fibrils. We added several
constraints to our initial linear β-sheet structure: (i) Intramolecular planarity (ii)
hydrogen bonding constraints to β-sheets that are presumably above and below
the spin-labeled peptide (iii) backbone and dihedral bonding constraints in
regions thought to be β-sheet and turn regions. The intramolecular distances
between the spin labels in the theoretical model was compared to experimental
EPR data. Using solely the PRONOX approach and the aforementioned
constraints, one model consistently matched the experimental results (Fig. 2.7).
This initial model, we call the “open U-shaped” model consists of a β1-strand
41
Figure 2.7: Initial PRONOX model of hIAPP in
fibrillar form.
The initial model generated from pronox is this "open U-
shaped" model. It consists of a ß1-strand between
residues 11-18 and a ß2-strand between 28-37 and a
loop region between residues 19-27. The odd numbered
residues of the ß1 and ß2 strands presumably faces
inward. The N-terminus is drawn as a ß-sheet structure
for the purposes of running the program, but is known to
be an unstructured region. Spin-labeled residues are not
shown in this model.
42
between residues 11-18 and a β2-strand between 28-37 and a loop region
between residues 19-27. The odd numbered residues of the β1 and β2 strands
presumably face inward. The N-terminus is drawn as a β-sheet structure for the
purposes of running the program, but is known to be an unstructured region
(Jayasinghe and Langen 2004). While this model satisfied most of the distance
constraints, it did not satisfy all of them. Additionally, the discussion section
presents how this structure was not in alignment with other current models and
could not account for some of the mobility data that I obtained.
The PRONOX approach was tremendously helpful in allowing us to
simulate possible models and teaching me how small changes in certain parts of
the peptide can change the overall structure. It also provided us with a starting
point from which better models could be made. However, we also found there
were certain starting constraints that were necessary inputs for the program but
we thought might create a bias towards a particular structure. In order to remove
any bias that is introduced into structural modeling (i.e. planarity), we decided to
only use DEER double-distance constraints and use them in SAMD using the
AMBER8 (Pearlman D 1995) program. In order to generate these structures, 48
peptides containing residues 12-36 were stacked on top of each other as a βeta
sheet, with a disconnection between Ile26 and Ile27 (Fig. 2.8A). The sheets were
introduced disconnected from one another with the idea that the 15 DEER
distance constraints per peptide will serve as the only restraint between the two
sheets and the disconnection may allow for more flexibility of the sheets in
relation to one another. The first structure generated contained two beta sheets
43

Fig. 2.8 Illustration of simulated annealing protocol
A) The original SAMD starting structure contained 48 peptides of hIAPP
residues 12-36 with a break between residues 26 and 27. The light blue
and magenta represent the two disconnected β-sheets. Figure (B)
represents the initial cycle of SAMD and illustrates the β-helical spiral
that begins to form.

A B
44
that were disconnected from each other but showed a clear β-helicity (Fig. 2.8B).
The labels in this structure were then replaced with the original amino acids and
the strands were connected to one another, completing each individual peptide.
Importantly, we found that the starting position of one of the β-sheets in relation
to the other β-sheet was a determinant of the handedness of the twist of the final
fibril structure. That is, if the input structure had the β-sheets perfectly in-line with
each other (what we called a one-to-one structure), then the final SAMD structure
resulted in a left-handed twist (Fig. 2.9B). If, however, the sheets were offset by 4
peptides (a one-to-four structure) then the resulting SAMD structure contained a
right-handed twist (Fig. 2.9A). Some cryo-EM and fiber diffraction studies
(Goldsbury C 2000) (Claire S. GoldsburyaGarth J. S. CooperbKenneth N.
GoldiecShirley A. MüllerdEtuate L. Saafie 1997) (Kajava AV 2005) suggest that
the predominant fibril type may contain a left-handed twist. Nevertheless, our
data and EM images from other studies (Saibil 2007) (Aaron K. Chamberlain*
2000) also suggest that the right-handed twist is possible. Thus, until one type of
twist can be definitively ruled out, then both structures are presented as
possibilities.
After several cycles of refinement, a SAMD structure was generated for
both right-handed and left-handed hIAPP fibrils (Fig. 2.9). Both structures appear
to be mirror images of each other, thus their structural characteristics and
interactions should not vary significantly. In fact, Figure 2.10 shows the plots of
the predicted backbone, phi (Φ) and psi (ψ) torsion angles for both the right-
handed and left-handed structures. This information not only shows the
45
Right-handed Left-handed
Figure 2.9 Illustration of Right-handed and Left-handed
hIAPP fibrils
A) Model of hIAPP protofilament with a right-handed twist. B)
Model of hIAPP protofilament with a left-handed twist.
46
Fig. 2.10: Predicted Backbone Angles (Phi and Psi)
A) The Phi (Φ) torsion angles as predicted by SAMD for the right-
handed structure. B) The Phi (Φ) torsion angles as predicted by
SAMD for the left-handed structure. C) The Psi (ψ) torsion angles
as predicted by SAMD for the right-handed. D) The Psi (ψ) torsion
angles as predicted by SAMD for the left-handed.
47

Figure 2.11: Model of hIAPP fibril
Representation of the fibrillar form of hIAPP using SAMD with contraints
from DEER data. A) Representation of 48 individual hIAPP peptides
stacked on top of each other after undergoing SAMD. The red peptide
represents on individual hIAPP peptide in fibrillar form. The C-terminus
(residues 1-11) are omitted from the structure. B) A closer view of the
mid-section of (B), focusing on the staggered appearance of individual
hIAPP peptides. C) View of the individual hIAPP peptide and amino
acids.
48
similarities between the two twists, but also shows that the regions that we
predict are in β-strands (residues 12-18 and 31-36) of the fibril is indeed in a
predicted β-sheet conformation according to the phi and psi values. The regions
that deviate from these angles are located in the loop regions of the peptide
(residues 25-30).
I will focus on the right-handed model as the primary structure from which
my discussion will stem (Fig. 2.11). As an overall structure, the hIAPP protofibril
appears to be in a β-spiral structure. That is, the β-strands of each individual
peptide stack parallel on top of each other yet create a twist such that the overall
protofilament forms a β-spiral. Additionally, it presents as a unique structure in
which the β2 strand is staggered 5-6 peptides away from the β1strand of the
same molecule of hIAPP (Fig. 2.11B red peptide). In this particular model, the β-
sheets twist around each other in a right-handed manner and form a pitch of
approximately 220-250 Å, which is in agreement with our EM images and pitch
data (Fig. 2.2). The discussion section below describes how this stagger and
twist allows for the long double-mutant distances seen between the β1 and β2
strands, yet also satisfies a myriad of other data gathered from previous studies
done by other groups as well as our own.

2.4 Discussion
There have been many studies on the fibrillar structure of amyloidogenic
proteins. However, hIAPP fibrils are one example of an amyloidogenic protein
whose three-dimensional structure had not been fully elucidated. I will begin with
49
a discussion of the most current theoretical models of hIAPP fibrils and then
discuss our proposed model in light of all the current data.
The most recent proposed model by Luca et al uses solid-state NMR and
electron microscopy to identify hIAPP peptide conformations and structural
models. Some key outcomes from this study showed that hIAPP fibrils are
polymorphic in nature. Using different conditions they were able to create
“twisted” and “striated ribbon” fibril morphology. They propose that in a
morphologically homogenous fibril sample (striated ribbon), C-NMR chemical
shifts reveal that residues 18-27 is a bend between two β-strands, 8-17 and 28-
37 (Chapter 1, Fig. 3A) (Luca, Yau et al. 2007). The study also reveals the amino
acids that deviate from a β-sheet pattern are: 18, 20, 21, 27, 28, 31 and 37.
Moreover, residues 8-13 have greater amplitude of motion than residues 14-37.
Without stating whether the interactions are intra- or inter-molecular, their study
also states that residue 14 or 15 is in close proximity to residue 26 or 27.
Computational modeling for the hIAPP protofilament reveals closely packed β-
strands with dimensions of approximately 50 Å by 50 Å (Chapter 1 Fig. 3A).
While this structure is the most recent hIAPP proposed model, it is important to
note that it is still a preliminary model that lacks the detailed structural information
that a final structure should contain, such as orientation of residues and exact
location of interdigitating residues.
A different model by Kajava and Steven et al used computational methods
to show hIAPP fibrils as a planar S-shaped fold with three β-strands, stacked in
register (Chapter 1, Fig. 3B) (Kajava AV 2005). They hypothesize that residues
50
20-27 and 30-37 are parallel and residues 30-37 and 8-16 are anti-parallel. They
propose that the S-shaped fold is a planar, serpentine molecule that is stacked
axially, in register and generates an array of parallel β-sheets. It is important to
note that their model does not contain any experimental data. Rather it is
computationally derived based on physical constraints introduced from previous
studies (i.e. dimensions, mass per unit length calculations, etc).
An Amyloid-Beta (Aβ) model by Riek et al has often been used to draw
comparisons to what the hIAPP fibril model is expected to be. In this model, they
use NMR, cryo-electron microscopy and computational modeling to show that Aβ
residues, 18-42, forms a β-turn-β motif with two parallel in-register β-sheets
(Chapter 1, Fig. 3C) (Lührs T 2005). Cryo-EM revealed the length of a β-strand to
be 45Å and the width to be 25 Å. Importantly, he shows a staggering of the β1
and β2 strands in relation to one another. This stagger, albeit shorter, is similar to
the one we observe in hIAPP fibrils (discussed below).
The atomic structures of two crystallized segments of IAPP have also
been determined by Wiltzius et al (Wiltzius, Sievers et al. 2008). Their data
indicates that the NNFGAIL (21-26) and SSTNVG (27-33) segments form
hydrogen bonds with identical segments 4.8 Å above and below it, in an
extended parallel β-sheet. They propose that the SSTNVG segment forms a
Class 1 steric zipper (Sawaya, Sambashivan et al. 2007), with the Ser 29 side
chains extending across the interface and forming a hydrogen-bond with one
another. The NNFGAIL segment does not form a steric zipper as the SSTNVG
segment, but appears to form a turn that leads into a steric zipper. It is also
51
important to point out that this study relies on data from short segments of the
entire protein, rather than a full-length form. Thus, the full length wildtype peptide
may not necessarily behave in the same way and form the same structures as
the short segments do.
While these models propose structural information about hIAPP fibrils,
they are all preliminary and some are derived solely with computational methods.
They also do not address detailed structural motifs, such as the exact locations
of the β-sheets, their orientation towards one another and the intermolecular
distances between residues that can help create a detailed three-dimensional
structure of the fibril.
The goal of my study was to obtain detailed structural information of
hIAPP fibrils using continuous-wave EPR and four-pulse DEER in conjunction
with computational refinement using SAMD. We generated 28 singly-labeled and
21 doubly-labeled derivatives of hIAPP fibrils and studied their structural
features. Mobility assays revealed a highly mobile N-terminus suggestive of an
unstructured region. It is clear that our model agrees with previously published
data that the β-sheets are formed from β-strand segments that run perpendicular
to the long axis of the fibril and are linked by hydrogen bonds that run parallel to
the fibril axis (Eanes 1968; Sunde 1998; Tycko 2006). The periodicity pattern
between residues 13-18 and 31-36 is highly suggestive of β-sheet regions with
the odd residues of both strands facing inward. The loop region in our staggered
model starts at residue 19 and ends at residue 29, which agrees closely with the
C-NMR data obtained from Luca et al (Luca, Yau et al. 2007).
52
The four-pulse DEER experiments combined with SAMD modeling
revealed a unique model in which the β2 strand is staggered 5-6 peptides away
from the β1 strand of the same molecule of hIAPP (Fig. 2.11B red strand is one
peptide). The interesting characteristic in this model, not previously recognized
by any other study on hIAPP, is that the β1 and β2 strands from one molecule of
hIAPP are not in the same plane, but instead staggered in relationship to each
other. In Figure 2.11B, the image focuses on the relationship of about 18 strands.
The red peptide highlights one individual hIAPP peptide. This particular image is
positioned such that one can look in between the β-strands of the individual
peptide in order to illustrate that the β1 and the β2 strands are staggered in
relation to one another. The β-sheets also appear to twist around each other
creating a right-handed β-helix-type structure that we aptly named a β-spiral. The
overall structure of the protofilament contains a pitch of approximately 220-250 Å
which is confirmed with previous studies (Sumner Makin and Serpell 2004)
(Kajava AV 2005) as well as our own EM data showing the pitch of a fibril is in
the range of 180-380 Å (Fig. 2.2). The right-handedness of our model is also
confirmed experimentally with the EM images of hIAPP fibrils (Fig. 2.2) that
shows fibrils twisting in a right-handed manner.
The intra-peptide stagger present in our model can explain many of the
structural findings not only in our studies, but previous studies done on the
structure of hIAPP fibrils. First off, the double mutant DEER data revealed that
distances between residues in the β1 strand to the β2 strand are too long to be in
the tightly packed “hairpin” model that has been the most recent suggestion. The
53
hIAPP models suggested by all of the aforementioned studies (Kajava AV 2005;
Lührs T 2005; Luca, Yau et al. 2007; Wiltzius, Sievers et al. 2008) show that the
β1 and β2 strand of one hIAPP molecule are packed in close proximity to each
other. While it is still true that a β1 strand packs closely and interdigitates with a
β2 strand, we found that this close interaction is not between strands of the same
hIAPP molecule. Rather the β1 strand of a hIAPP molecule is packed against the
β2 strand of molecule that started 5-6 peptides below it. Our four-pulse DEER
experiments are one of the only currently used methods that studies the
intramolecular distances between residues of hIAPP fibrils. The distances
obtained from this method were clearly too long to be located next to each other
(within 10Å) as other models have predicted. Thus, while these protofilaments
still appear to be tightly packed with residues in close proximity with one another,
our staggered model also satisfies our double-mutant data.
There are also several characteristics and favorable interactions between
residues that help support our model. Figure 2.11C is an illustration of one
peptide of hIAPP in its fibrillar form. First, it is interesting to point out the “u-
shape” of the structure. This “u-shape” is very similar to the original “u-shape”
model we observed using PRONOX (Fig. 2.7). The difference between these
structures is that the PRONOX model (Fig. 2.7) was a planar peptide and the
SAMD model shown in Figure 2.11C is staggered. Moreover, the mobility and
accessibility data is also satisfied in this model. The residues that face inward in
the two β-strand regions (15, 17, 31,33,35) show low mobility because they are
packing next to each other and are physically restrained in their movement. The
54
residues that are on the outside of the two β-strand regions (14,16,18,32,34,and
36) are more highly mobile because they are not packing tightly against other
residues. Additionally, Gly24 is located in the center of the loop region, thereby
allowing more flexibility for the turn to form (Fig. 2.9C).
Another interesting observation is that Phe15 and Phe23 appear to be
facing inward which is highly favorable given they are both hydrophobic (Fig.
2.9C). Moreover, the phenyl rings from different strands align with each other in
the center of the filament and create a hydrophobic core.
One may ask the question of why a peptide would prefer to take a
staggered conformation as opposed to one that is more planar. One explanation
for this may be that as hIAPP is misfolding and forming fibrils, certain peptides
may stack on top of one another. If the stacking is purely planar then the only
interactions driving this stacking are the hydrogen bonds between β-sheets and
possibly the inter-residue interactions. However, if the stacking were to be in a
more staggered manner (as they appear to be in our structure), then the stagger
exposes single sheets at the ends of the fibril. The respective exposed surfaces
are “sticky ends” that will support and induce fibril growth. Thus, there are not
only hydrogen bonds between β-sheets and inter-residue interactions, but there
is a higher propensity for hydrophobic interactions. These “sticky-ends” and the
large grooves seen in the overall twisted structure represent ideal targets for drug
inhibitors. Moreover, many studies have shown that seeding amyloid proteins
with preexisting fibrils significantly increases the fibril formation and creates more
homogenous fibril morphology (Anant K. Paravastua 2009) (Brian O'Nuallain
55
2004). According to our model, these seeds already contain this staggered “wall”
of hydrophobicity and would thus drive the fibril formation more quickly.
In conclusion, I have presented a novel structure of hIAPP in its fibrillar
form based upon SAMD computational modeling that solely used experimental
restraints from continuous-wave and pulsed-EPR. The resulting structure is one
that has the individual hIAPP peptides staggered. The resulting hIAPP β-strands
that form are parallel to each other and yet create a twist. The overall
protofilament creates a β-spiral conformation with dimensions and pitch that
agree with experimental data. I presented both right-handed and left-handed
twists as possibilities, but further endeavors must be made towards determining
exactly which type of twist hIAPP fibrils take. Overall, the discovery of this
structure is a promising leap towards the discovery of potential therapeutic
agents that can modify the structure and hence the disease outcome. Future
endeavors should include using this structure to create specific inhibitors for the
β-sheet regions and possibly inhibiting the formation of the hydrophobic wall that
can induce further misfolding. In the case of TTDM, it is important to note, that
current diabetes medications (i.e. insulin, metformin, glitazones) treat the
symptoms of the disease or reduce the burden of glucose that can cause further
damage. Our structure will give insight on how to directly inhibit the agent that
causes β-cell destruction in TTDM (hIAPP), thereby potentially sparing the total
loss of insulin that is seen in advanced diabetes. Our structural information can
also be used for other misfolded peptides that play a role in amyloid diseases,
especially Aβ (Alzheimer disease) and α-synuclein (Parkinson disease).
56

Chapter 3: Annexin A5 Reduces Human IAPP Induced Beta-Cell
Apoptosis by Directly Interacting With Human IAPP

The work done in this chapter was originally studied by Robert Ritzel and Sajith
Jayasinghe. I resumed the study and contributed data pertaining to the structural
features (EPR, ThT) and morphology of hIAPP (EM) as it relates to its interaction
with annexin A5.

3.1 Introduction
Protein misfolding and the deposition of amyloid fibrils are hallmarks of
many diseases including Alzheimer disease, Parkinson disease, Huntington
disease and type 2 diabetes (Lashuel 2002; Bates 2003; Kayed 2003; Janson,
Laedtke et al. 2004; Lansbury 2006; Jayasinghe and Langen 2007). Although
different polypeptides with little or no sequence similarity form fibrils in the
various diseases, there are a number of structural similarities with respect to the
mechanism of misfolding as well as the toxicity of the misfolded proteins. In vitro
studies have shown that amyloid fibril formation is a step wise process, in which
non-fibrillar oligomeric structures are formed before the first fibrils can be
detected (Anguiano, Nowak et al. 2002; Janson, Laedtke et al. 2004). Although
the pathology of amyloid diseases is dominated by an abundance of fibrillar
deposits, the non-fibrillar oligomers have begun to receive much attention, as
57
they are now thought to represent the primary toxic species (Janson, Ashley et
al. 1999; Kayed 2003; Glabe 2006; Ritzel, Meier et al. 2007). In fact, in some
disease which are characterized by proteinaceous deposits, non-fibrillar amyloid
oligomers rather than amyloid fibrils represent the primarily deposited material. In
the endocrine pancreas of humans with type 2 diabetes the beta-cell peptide,
hIAPP, misfolds and aggregates into islet amyloid (Maloy, Longnecker et al.
1981; Clark, Wells et al. 1988) and is thought to promote beta-cell apoptosis in
type 2 diabetes (Butler, Janson et al. 2003; Butler, Jang et al. 2004; Hull,
Westermark et al. 2004).The property of human hIAPP to induce apoptosis has
been described in different experimental models including beta-cell and non
beta-cell lines (Lorenzo, Razzaboni et al. 1994; O'Brien, Butler et al. 1995;
Zhang, Liu et al. 1999; Saafi 2001; Butler, Janson et al. 2003), human islets
(Lorenzo, Razzaboni et al. 1994; Butler, Janson et al. 2003; Ritzel, Meier et al.
2007) and transgenic animal models (Janson, Soeller et al. 1996; Verchere,
D'Alessio et al. 1996; Butler, Jang et al. 2004). As in the case of other
amyloidogenic peptides, membranes are likely to play an important role in the
toxicity of IAPP, because misfolded, oligomeric forms of hIAPP cause
pronounced membrane permeabilization (Mirzabekov 1996).
Annexins represent a highly conserved family of Ca
2+
-dependent
membrane-binding proteins. They are abundantly expressed in a wide range of
tissues (Reutelingsperger 2001; Rescher 2004) and found in intra as well as
extracellular locations (Flaherty 1990; Romisch 1992). Annexin A5 has been
shown to protect from cytotoxicity of Alzheimer’s β-protein (Aβ) (Lee G 2002),
58
however the mechanism of protection has remained unclear. Recently it was
reported that the aggregated extracellular deposits (drusen) of subjects with age-
related macular degeneration contain annexin A5 (Crabb 2002). Thus, annexin
A5 is co-deposited with misfolded proteins in disease. To shed further insight into
annexin A5’s potential role to prevent and/or reverse the toxic actions of
amyloidogenic protein oligomers as well as to extend the concept to hIAPP, we
hypothesized that annexin A5 could have a protective property against amyloid
toxicity from hIAPP. Therefore we posed the following questions: (1) Does
annexin A5 protect human pancreatic islets from hIAPP induced apoptosis? (2) If
so, does it also protect from hIAPP induced membrane perturbation? (3) Is the
mechanism of protection by annexin A5 due to a direct alteration of hIAPP
misfolding?

3.2 Materials and Methods
3.2.1 Human islet tissue
Human pancreatic islets were isolated from the pancreas retrieved from
three non-diabetic, heart-beating organ donors by the Diabetes Institute for
Immunology and Transplantation, University of Minnesota (Bernhard J. Hering)
and the Northwest Tissue Center Seattle (R. Paul Robertson). The islets were
maintained in RPMI culture medium at 5 mM glucose and 37° C in humidified air
containing 5 % CO
2
. After the islet isolation process the islets were cultured for
three to five days before experiments were performed. Aliquots of human islets
59
were incubated for 48 hours with vehicle (H
2
O + 0.5% acetic acid), 40 μM rat
IAPP, 40 μM hIAPP and 40 μM hIAPP + 1 μM annexin A5. After static incubation
the number of apoptotic cells in human islets was detected using the TUNEL
staining method (In Situ Cell Death Detection Kit, AP; Roche Diagnostics,
Indianapolis, IN). Tissue samples were analyzed on an inverted microscope
(Inverted System Microscope IX 70; Olympus, Melville, NY) as described
previously (Butler, Janson et al. 2003).

3.2.2 Vesicle leakage assay
This method is based on the concept that leakage through the vesicle
membranes can be detected with the fluorophore/quencher pair ANTS/DPX
(Molecular Probes, Eugene, OR), which are trapped inside the vesicles. Any
leakage through the vesicle membrane will result in an increase in the
fluorescence intensity of the sample. For the preparation of vesicles (~0.1μm
diameter) phosphatidylcholine (PC) and PS (Avanti Polar-Lipids, Birmingham,
AL) were suspended in buffer containing a solute of ANTS/DPX and then frozen
and thawed five times. Lipid solutions were prepared at 100 mM to maximize
entrapment. The vesicles were formed by extrusion through polycarbonate
membranes (pore size ~0.1μm) and purified from unencapsulated ANTS and
DPX using Sephadex G-100 packed in a 2.5 ml Pasteur pipette. A more detailed
description of the method has been given by Wimley et al (Wimley 1994). We
prepared vesicles with a molar PS concentration of 1 % to reproduce PS
60
concentrations of subcellular compartments (Daum 1985), since hIAPP induced
beta-cell toxicity is initiated intracellularly.
Leakage assays were performed in 10 mm Quartz Fluorometer Cell Micro
Cuvettes (Starna Cells, Atascadero, CA) with a JASCO FP-6500
spectrofluorometer at a wavelength of 350 nm for excitation and 520 nm for
emission. During a time-course measurement a small volume (5 μl) of vesicle
stock solution with entrapped ANTS/DPX was added to 500 μl buffer (10 mM
phosphate, 100 mM NaCl) and quickly mixed. Subsequently small volumes (10-
20 μl) of vehicle, hIAPP or annexin A5 were added. Triton X (0.1 %) was used as
a detergent to induce maximum fluorescence at the end of each experiment. To
allow a better comparison of individual experiments the fluorescence change was
normalized to the maximum observed after application of Triton X.

3.2.3 Thioflavin T fluorescence assay
The fibrillization process of hIAPP was monitored using the fluorescence
increase of thioflavin T (ThT), a dye known to preferentially bind amyloid fibrils.
Therefore ThT fluorescence increases in a solution of freshly reconstituted hIAPP
as amyloid fibrils grow. Each fibrillization reaction was performed at 5 μM ThT in
glycine buffer (50 mM) and real-time emission intensities were measured at 482
nm with excitation at 450 nm. Measurements were performed at room
temperature with excitation and emission slit widths of 1 and 10 nm, respectively.
Fluorescence measurements were taken using a Jasco FP-6500
61
spectrofluorometer. Plots of ThT emission intensity as a function of time were
fitted to a sigmoidal curve (non-linear regression analysis).

3.2.4 Western Blotting
Western blotting was performed to determine annexin A5 protein
expression in human islets. Total protein samples were prepared in Laemmli
sample buffer from human islets, and B6CBA mouse tissues: lung and skeletal
muscle. Twenty ug of protein was separated on a 12% SDS-PAGE and
transferred to PVDF membranes (Bio-Rad, Hercules, CA). Membranes were
blocked with 5% nonfat dry milk in 0.1% Tween/Tris-buffered saline (TBS-T), and
incubated overnight at 4
o
C in 2.5% milk in TBS-T containing polyclonal rabbit
anti-annexin A5 antibodies (1:500; Santa Cruz Biotechnology, Santa Cruz, CA).
Membranes were washed five times with TBS-T and incubated with horseradish
peroxidase-conjugated goat anti-rabbit IgG (1:3000; Zymed Laboratories, S. San
Francisco, CA) for 1 hr. Proteins were visualized using enhanced
chemiluminescence (ECL; Amersham, UK).

3.2.5 IAPP stock solutions
Human and rat IAPP were purchased from Bachem, Torrance, CA. For
preparation of stock solutions the lyophilized peptides were reconstituted with
deionized water containing 0.5 % acetic acid. Human IAPP was prepared as in
previous preparations that gave rise to toxic species (Janson 1999; Butler,
Janson et al. 2003). Briefly, cell culture experiments were carried out by dilution
62
of stock solutions with the culture medium to obtain a final concentration of 40
μM and then directly applied to the cells. For vesicle leakage assays the hIAPP
stock solutions were diluted with buffer to obtain final concentrations of 40 μM.
Acetic acid concentrations in the culture medium applied to cells were always
<0.003 %.
For Thioflavin T assays hIAPP was reconstituted in 6 M Guanidine-HCl
buffer (150 μl per 0.5 mg hIAPP) and loaded into a hydrated (with deionized
water) C-18 silica column (Macro Spin Column, Havard Biosciences, Holliston,
MA). Guanidine was washed out with 150 μl deionized water. Then IAPP was
recovered from the column by application of 150 μl Hexafluoro-2-isopropanol
(HFIP). The HFIP in the flow-through was allowed to evaporate and the
remaining peptide was reconstituted with deionized water containing 0.5 % acetic
acid and 2.5 % HFIP.

3.2.6 Electron paramagnetic resonance (EPR) and site directed spin
labeling
Spin labeled samples were prepared as described previously (Jayasinghe
and Langen 2004). Briefly, hIAPP containing single cysteines at either position
14 or 21 were obtained from Biomer Technology (Hayward, CA). Mutant hIAPP
was labeled with 3-fold excess MTSL label (1-oxyl-2,2,5,5-tetramethyl-Δ3-
pyrroline-3-methyl methanethiosulfonate; obtained from Toronto Research
Chemicals (Toronto, Canada)) for 1 hour at room temperature. Labeled peptide
was recovered with cation exchange chromatography and desalted with a C-18
63
silica column (Macro Spin Column, Havard Biosciences, Holliston, MA). The spin
labeled cys residue generated under these conditions is referred to as R1. Spin-
labeled hIAPP was collected and stored in hexafluoroisopropanol (HFIP). Prior
to usage, HFIP was evaporated and the sample was dissolved in 10 mM
phosphate, 100 mM NaCl (pH 7) buffer to a final concentration of 100 μM.
To discern the effect of annexin A5 on monomeric hIAPP, annexin A5
(100 μM) was added to freshly dissolved hIAPP (100 μM) and EPR spectra were
obtained immediately. To evaluate the effect of annexin A5 on hIAPP fibril
formation 100 μM spin labeled hIAPP was incubated at room temperature for a
minimum of 3 days in the absence or presence of annexin A5 (20 μM). Fibrils
and aggregates were harvested by centrifugation (13,000 rpm for 10 min using
Eppendorf table top centrifuge 5415D) and washed with buffer. For EPR
spectroscopy, samples were loaded into glass capillary tubes (0.6 mm inner
diameter and 0.84 outer diameter, VitroCom, Mt. Lakes, NJ). Spectra were
obtained on a Bruker EMX spectrometer (Bruker Instruments, Billerica, MA) at
room temperature at an incident microwave power of 12 mW with a scan range
of 150 gauss.

3.2.7 Electron Microscopy
5–10 μl of hIAPP

solution was adsorbed onto glow-discharged carbon-
coated copper

grids. Excess sample was removed by blotting on filter paper.

Grids were stained with 5 μl of 3% uranyl acetate for

3 min. After blotting excess
64
uranyl acetate and washing with

5 μl of distilled deionized water, staining was
repeated

for 1 min with 5 μl uranyl acetate.

3.2.8 Calculations and statistical analysis
The best-fit curves of hIAPP induced change in relative fluorescence
intensity in relationship to annexin A5 concentration in vesicle leakage
experiments and of thioflavin T time-dependent normalized fluorescence intensity
were derived by nonlinear regression analysis.
Differences in the number of apoptotic cells per islet and in normalized
fluorescence intensity were analyzed by the Student’s t-test. A probability of <5
% occurrence by chance alone denoted statistical significance.

3.3 Results
3.3.1 Annexin A5 protects human islet tissue from hIAPP toxicity
Static incubation of human islets with freshly reconstituted hIAPP for 48
hours resulted in a 4-fold increase in the number of apoptotic cells compared to
incubation with vehicle (control) or non-amyloidogenic rat IAPP (Fig. 3.1).
Annexin A5 reduced the number of apoptotic cells in human islets after 48 h
incubation by ~50% (Fig. 3.1) demonstrating that annexin A5 has a protective
effect from hIAPP toxicity. For such protection to be of physiological relevance, it
is essential to know that annexin A5 is present in human islet tissue in which
65
Control r-IAPP h-IAPP h-IAPP + Anx A5
0
5
10
15
20
*
*
#
n=97
n=69
n=61
n=52
Apoptosis
[cells/mean islet area]
Fig. 3.1: Annexin A5 protects islet tissue against hIAPP toxicity
Mean number of apoptotic cells in human islets (n = 3 donors)
incubated for 48 h with vehicle (control), rat IAPP (40 μM), h-IAPP
(40 μM) or h-IAPP (40 μM) + annexin A5 (1 μM). Data ± SEM. *: p <
0.05 versus control. #: p < 0.05 versus h-IAPP.

66
annexin A5
Fig. 3.2: Western blot of Islet cell lysates
Total protein extracts from human islet tissue were analyzed on a
Western blot using a polyclonal rabbit anti-annexin A5 antibody.
Mouse muscle and lung tissue are shown for comparison (low and
high annexin A5 expression).

67
hIAPP is expressed. In order to test this, we performed Western blot analysis of
islet lysates from two non-diabetic tissue donors. As shown in figure 3.2,
significant amounts of annexin A5 are present in human islet tissue. The intensity
of the respective bands are significantly stronger than those in mouse muscle
(control for low annexin A5 expression) and comparable to that of mouse lung
tissue (control for a high level of expression).

3.3.2 Annexin A5 reduces hIAPP-dependent membrane perturbation
To examine whether the mechanism for the protective effect of annexin A5
involves the prevention of membrane destabilization by hIAPP we performed
vesicle leakage assays. We generated liposomes that were filled with the
fluorophore/quencher pair ANTS/DPX. When trapped in liposomes, these dyes
give rise to low fluorescence intensity, but upon leakage (and subsequent
dilution) the fluorescence intensity of the dyes increases strongly. Figure 3.3 A
shows that administration of hIAPP induced membrane leakage. In contrast, the
presence of annexin A5 (no Ca
2+
added) strongly reduced membrane leakage.
The membrane interaction of annexin A5 has been well characterized and under
the present buffer conditions annexin A5 only interacts with membranes in the
presence of high (μmolar) concentrations of Ca
2+
. We were able to verify the lack
of membrane interaction under the current conditions using gel filtration as well
as Trp fluorescence assays (data not shown). Thus, the protective effect of
annexin A5 was not due to an interaction with membranes.
68
h-IAPP h-IAPP+Anx A5
0.0
0.1
0.2
0.3
0.4
0.5
0.6
*
Δ Δ Δ Δ normalized
fluorescence intensity
A
0 1 2 3 4
0.0
0.1
0.2
0.3
0.4
0.5
1% PS vesicles
1% PS vesicles 1mM Ca
2+
Annexin A5 [μ μ μ μM]
Δ Δ Δ Δ relative
fluorescenceintensity
B
Fig. 3.3: Membrane Leakage Assays of hIAPP in the presence of
annexin A5
Panel A: Change of normalized fluorescence intensity in vesicle leakage
assays after application of r-IAPP (40 μM) or h-IAPP (40 μM) with and
without annexin A5 (4 μM), n = 3 experiments per group. Experiments were
performed without addition of Ca
2+
, conditions under which annexin A5
does not interact with these vesicles.
Data ± SEM. *: p < 0.05 versus control. #: p < 0.05 versus h-IAPP.
Panel B: Dose-response relationship between annexin A5 and change of
normalized fluorescence intensity in vesicle leakage assays after
application of h-IAPP (40 μM) with and without calcium added to the buffer.
The solid and dashed lines are derived by non-linear regression analysis.
69
Next, we performed a dose-response analysis, in which the amount of leakage
was quantified in the presence of varying amounts of annexin A5 (Fig. 3.3 B). A
strong reduction in the leakage was observed as the annexin A5 concentration
was increased from 0 to 1uM. The protective effect then began to level off as the
annexin A5 concentration was increased further to 4uM.
Considering that the hIAPP concentration was 40uM, the annexin A5
effect plateaus at concentrations that are more than one order of magnitude
lower than those of hIAPP. These data show that annexin A5 does not need to
bind stoichiometrically to each and every hIAPP molecule to exert its effect;
rather the data are consistent with the binding of annexin to a (possibly
oligomeric) subset of hIAPP molecules. As an additional control, we also
performed the same experiments under conditions that induce annexin A5
binding to the PS molecules in the membrane (Swairjo 1995). Despite the
addition of Ca
2+
and the concomitant complexation of PS, the protective effect of
annexin A5 did not exhibit a noticeable change (Fig. 3.3B). Virtually identical
results were obtained in the presence of 1mM Ca
2+
, which is known to promote
membrane interaction of annexins (Fig. 3.3B). These data further supported the
notion that the protective effect of annexin A5 was not mediated by an interaction
with membranes. Therefore we used fluorescence and EPR spectroscopy to test
whether annexin A5 might have a direct influence on the misfolding of hIAPP.

70
0 25 50 75 100 125
0.00
0.25
0.50
0.75
1.00
1.25
h-IAPP + Anx A5
h-IAPP
Time [min]
Normalized fluorescence
intensity at 482 nm
Fig. 3.4: Thioflavin T and EM of hIAPP in the presence of annexin A5
Upper panel: Time-dependent normalized thioflavin T (ThT) emission
curves of h-IAPP (20 μM) with and without annexin A5 (4 μM). The ThT
curves are normalized to the maximal observed intensity at the end of each
aggregation reaction. The solid and dashed lines are derived by non-linear
regression analysis of n = 3 experiments per group (all data points shown).
Lower panels: Electron Micrographs of h-IAPP aggregates formed in the
absence (A) and presence of annexin A5 (B, C). Picture A shows 100 μM
IAPP fibrils grown with 20 mM Hepes 100 mM NaCl buffer. The presence
of long fibrils is seen at 40K magnification. Picture B shows h-IAPP with
annexin A5 in a 5 to 1 ratio. Protein aggregates are seen throughout the
sample (20K magnification). Large aggregates are also observed at this
ratio, but become more predominant at higher concentrations of annexin
A5 as shown in picture C (h-IAPP with annexin A5 in a 3 to 1 ratio, 20K
magnification). The scale bars in A-C represent 1.0 μm.
71
3.3.3 Annexin A5 reduces thioflavin T fluorescence of hIAPP and fibril
formation (EM)
First we performed a Thioflavin T fluorescence assay. This assay is based
upon the strong fluorescence enhancement that occurs when amyloidogenic
peptides or proteins undergo a conformational reorganization into fibrils and
other oligomers. Consistent with previous reports the time-dependent kinetics of
hIAPP fibrillization followed a sigmoidal pattern (Padrick 2002) (Fig. 3.4 A). With
annexin A5 the time-dependent increase in thioflavin T fluorescence is largely
prevented, indicating that annexin A5 directly interacts with hIAPP and prevents
the formation of thioflavin positive fibrils. These findings were further supported
by electron microscopy, which indicated that annexin A5 resulted in formation of
predominately amorphous aggregates rather than fibrils (Fig. 3.4 B).

3.3.4 Site-directed spin labeling and EPR spectroscopy reveal that IAPP
fibril formation is strongly altered in the presence of annexin A5
To further investigate the notion that annexin A5 has a direct impact on
hIAPP misfolding and fibril formation, I employed site-directed spin labeling and
EPR spectroscopy. Using this approach, our lab had previously shown that
hIAPP fibrils (Jayasinghe and Langen 2004), like those from many other
amyloidogenic proteins (Torok 2002; Der-Sarkissian A 2003; Margittai and
Langen 2006), form a parallel, in register structure in which same residues are
stacked on top of each other. Due to the close proximity of same sites in fibrils,
72
Fig. 3.5: EPR spectra of hIAPP in the presence and absence of annexin
A5
EPR Spectra of h-IAPP fibrils labeled with spin label R1 at position 14 or 21
(black line). The red spectra are obtained from aggregates formed from the
same spin labeled h-IAPP derivatives in the presence of annexin A5. Each
spectrum was normalized to represent the same number of spins. The
addition of annexin A5 causes increased signal amplitude due to a loss of
spin-spin interaction.

73
fibrils of spin labeled hIAPP derivatives give rise to highly characteristic EPR
spectra (32). As shown with two representative examples in figure 5, the EPR
spectra of fibrils grown from hIAPP harboring spin labels either at position 14 or
21 have strong components of exchange narrowing and dipolar broadening, both
of which result from the close proximity of same residues. In order to test,
whether annexin A5 can interfere with the formation of these specific structures,
we repeated the aforementioned experiments in the presence of sub
stoichiometric amounts of annexin A5. The EPR spectra of the aggregates that
formed in the presence of annexin A5 were strikingly altered and characterized
by a loss of the pronounced spin-spin interaction. Thus, these data further
illustrate that the structures of the hIAPP aggregates formed in the presence of
annexin A5 are different from those formed in the absence of annexin A5. Again,
the strong effect was observed using sub-stoichiometric amounts of annexin A5
(5-fold excess of hIAPP) further supporting the notion that annexin A5 does not
act by interacting stoichiometrically with monomeric hIAPP, but that it is likely to
interact with a sub population of hIAPP that is on the aggregation pathway and
that is present only at low concentrations. Indeed, we found no detectable EPR
spectral changes when equimolar amounts of annexin A5 were mixed with
freshly prepared, spin-labeled hIAPP (data not shown), conditions under which
hIAPP is predominantly monomeric.

74
3.4 Discussion
In the present study, we show that annexin A5 protects cultured human
islet tissue by attenuating hIAPP induced apoptosis. This protective effect
correlates well with the ability of annexin A5 to inhibit the membrane
permeabilizing effect of hIAPP. Several lines of evidence show that annexin A5
directly interferes with the aggregation and misfolding of hIAPP in solution while
the interaction of annexin A5 with membranes was less important in the present
study. First, annexin A5 protects membranes in the absence of Ca
2+
under
conditions in which annexin A5 does not interact with membranes. Second,
annexin A5 alters hIAPP misfolding in the absence of membranes as judged by
Thioflavin fluorescence, electron microscopy, and EPR spectroscopy. Third, only
sub-stoichiometric amounts of annexin A5 are required to attenuate the pro-
apoptotic and membrane permeabilizing effects of hIAPP suggesting that
interaction of annexin A5 with a subset of IAPP molecules is sufficient to exert
these effects. Furthermore, the EPR and electron microscopy data show that
annexin A5 interaction with this subset of hIAPP molecules alters the aggregation
pathway and inhibits the formation of the typical hIAPP fibrils. Thus, annexin A5
is likely to interact with a subset of hIAPP molecules that are in a pathogenic
conformation and on the pathway toward fibril formation.
Annexin A5 has been reported to be present primarily intra- but also
extracellularly (Flaherty 1990; Romisch 1992). Therefore an interaction of
annexin A5 with hIAPP might occur outside and/or inside beta-cells. In the
present experiments with isolated human islets it is likely that the extracellular
75
presence of annexin A5 directly prevented the formation of membrane and islet-
cell toxic protofibrillar species of applied hIAPP. The data on annexin A5
concentrations in human cells is sparse, but the concentrations applied to human
islets in culture and vesicles in the present experiments (1 μM and up to 4 μM)
corresponds to concentrations that have been found intracellularly (Flaherty
1990; Giambanco 1993). Lower concentrations might be expected in the
extracellular environment.
The endocrine pancreas of humans with type 2 diabetes is characterized
by islet amyloid deposits and beta-cell deficiency due to increased apoptosis
(Lohr, Bergstrome et al. 1992; Butler, Janson et al. 2003; Ritzel 2006). The
specific mechanism for the induction of beta-cell apoptosis is still unclear (e.g.
hIAPP, hyperglycemia, free fatty acids, cytokines), especially in prediabetic
conditions before decompensation of glucose metabolism. Animal models with
transgenic expression of hIAPP develop a phenotype with islet amyloid, beta-cell
loss due to increased apoptosis and impaired beta-cell function similar to human
type 2 diabetes (Butler, Janson et al. 2003; Butler, Jang et al. 2004). Therefore
hIAPP is one candidate causing beta-cell failure in type 2 diabetes and the
present results showing a cytoprotective effect of annexin A5 could provide a
new target to develop strategies for the preservation and/or restoration of beta-
cell mass in diabetes mellitus.
In conclusion, the present data support the concept of hIAPP induced
beta-cell apoptosis by the mechanism of membrane permeabilization. Annexin
A5 protects human islet tissue from hIAPP toxicity and directly reduces the
76
misfolding and membrane permeabilization caused by hIAPP. Considering the
previously reported effect of annexin A5 on Aß-dependent toxicity, it is interesting
to speculate that at least some members of the annexin family of proteins might
play a physiological role in protecting from amyloid protein toxicity.
77
Chapter 4: The effect of annexin A5 on Aβ and α-synuclein
toxicity and fibril morphology
4.1 Introduction
The previous chapter shows that annexin A5 has the ability to directly
interact with hIAPP, changing its fibrillar structure and inhibiting its toxicity in islet
cells. We broadened our speculation that annexin A5 protects against amyloid
toxicity by studying how annexin A5 affects the toxicity and misfolding associated
with two other amyloid proteins, Aβ and α-synuclein. As previously stated Lee et
al have shown that annexin A5 has a protective effect on Aβ toxicity (Lee, Pollard
et al. 2002), but the mechanism by which this happens has not been elucidated.
We investigated the ability of annexin A5 to alter the misfolding of Aβ and α-
synuclein by directly interacting with the amyloid protein. With the use of
biophysical methods, such as electron paramagnetic resonance (EPR), electron
microscopy (EM) and Thioflavin T (ThT) assays and transgenic expression of
annexin A5 in C. elegans we show that annexin A5 has the ability to interact with
and alter the misfolding of amyloid proteins as well as inhibit its aggregation and
toxicity.

4.2 Methods
4.2.1 Chemicals and Peptides
Synthetic wild-type human IAPP was obtained from Bachem Bioscience
Inc. (King of Prussia, PA). The spin label, 1-oxyl-2,2,5,5-tetramethyl- 3-pyrroline-
78
3-methyl methanethiosulfonate (MTSL), was obtained from Toronto Research
Chemicals (Toronto, Canada). Hexafluoroisopropanol (HFIP) was obtained from
Sigma-Aldrich. Single cysteine mutants of full-length IAPP containing a single
cysteine at position 21 and Aβ(1-40) containing a cysteine at position 35 were
purchased from Biomer Technology (Pleasanton, CA). The native cysteine
residues at position 2 and 7 are substituted with an alanine (Jayasinghe and
Langen 2004). Peptides arrived 90% purified and lyophilized and were stored at -
80 ° C until use.

4.2.2 Generation of α-Synuclein Single Cysteine Mutants
A truncation mutant (115ter) of human α-synuclein, containing residues 1–
115, was generated by adding two stop codons between residues 115 and 116.
Single cysteines were introduced by site-directed mutagenesis and verified by
DNA sequencing.

4.2.3 Spin Labeling and Fibril Assembly of 115ter α-Synuclein
Recombinant α-synuclein was filtered through an Amicon YM-100 spin
filter (MWCO 1 x 105, Millipore) to remove any pre-aggregates. Immediately
before spin labeling, dithiothreitol was removed by loading the protein solution
onto a PD-10 column (Amersham Biosciences, GE Healthcare) equilibrated with
buffer containing 20 mM Hepes, pH 7.4, 100 mM NaCl, and 1 mM EDTA, then
eluted using the same buffer. α-Synuclein was labeled with 10-fold molar excess
of R1 MTSL spin label [1-oxyl-2,2,5,5-tetramethyl-D-pyrroline-3-
79
methyl]methanethiosulfonate (Toronto Research Chemicals, Toronto, Ontario,
Canada) for 1 h at room temperature. Excess label was removed using PD-10
columns with the aforementioned elution buffer. Spin-labeled proteins were
washed twice with elution buffer, concentrated using Amicon Ultra-4 centrifugal
filter units (MWCO 5 x 103, Millipore), and assembled into fibrils as described
previously (Der-Sarkissian A 2003). To evaluate the effect of annexin A5 on α-
synuclein fibril formation, 100 μM of spin-labeled α-synuclein was incubated in
the presence and absence of annexin A5 at 37 ° C under constant agitation for 3–
5 days. Fibrils were harvested by ultracentrifugation and washed twice with the
same elution buffer.

4.2.4 Spin Labeling and Fibril Assembly of hIAPP and Aβ
Spin labeled samples were prepared as described previously (Jayasinghe
and Langen, 2004). Mutant h-IAPP and Aβ was labeled with 3-fold excess MTSL
for 1 hour at room temperature. Labeled peptide was recovered with cation
exchange chromatography and desalted with a C-18 silica column (Macro Spin
Column, Harvard Biosciences, Holliston, MA). Spin-labeled peptide was
collected and stored in hexafluoroisopropanol (HFIP). Prior to usage, HFIP was
evaporated and the sample was dissolved in 10 mM phosphate, 100 mM NaCl
(pH 7) buffer to a final concentration of 100 μM. To evaluate the effect of
annexin A5 on h-IAPP and Aβ fibril formation 100 μM spin labeled h-IAPP or Aβ
was incubated at room temperature for a minimum of 3 days in the absence or
presence of annexin A5 (20 μM). Fibrils and aggregates were harvested by
80
centrifugation (13,000 rpm for 10 min using Eppendorf table top centrifuge
5415D) and washed with buffer.

4.2.5 EPR spectroscopy
Sample containing amyloid peptide incubated with annexin A5 were
loaded into glass capillary tubes (0.6 mm inner diameter and 0.84 outer diameter,
VitroCom, Mt. Lakes, NJ) at the end of the 3-4 day incubation period. The
sample was centrifuged and all supernatant was removed to eliminate smaller
aggregates and free spin label. Spectra were obtained on a Bruker EMX
spectrometer (Bruker Instruments, Billerica, MA) at room temperature at an
incident microwave power of 12 mW with a scan range of 150 gauss.

4.2.6 Electron Microscopy
To examine the effect of annexin A5 on fibril growth, 10uL of amyloid
incubated with annexin A5 were adsorbed onto carbon and formavar-coated
copper grids and negatively stained with 2% (w/v) uranyl acetate solution for 5
min. The stained grids were examined and photographed using a JEOL JEM-
1400 electron microscope at 80-100 kV.

4.2.7 Thioflavin T Assays
The fibrillization process of hIAPP and α-synuclein was monitored using
the fluorescence increase of thioflavin T (ThT), a dye known to preferentially bind
amyloid fibrils. ThT fluorescence increases in a solution of freshly reconstituted
81
peptide as amyloid fibrils grow. Each fibrillization reaction was performed with
100uM amyloid peptide in the presence of 20uM annexin A5. A 5 μM ThT aliquot
in glycine buffer (50 mM) was added and real-time emission intensities were
measured at 482 nm with excitation at 450 nm. Measurements were performed
at room temperature with excitation and emission slit widths of 1 and 10 nm,
respectively. Fluorescence measurements were taken using a Jasco FP-6500
spectrofluorometer. Plots of ThT emission intensity as a function of time were
fitted to a sigmoidal curve (non-linear regression analysis).

4.2.8 Western blotting for annexin A5 expression in human islets
Total protein samples were prepared in Laemmli sample buffer from
human islets, and B6CBA mouse tissues: lung and skeletal muscle. Twenty ug of
protein was separated on a 12% SDS-PAGE and transferred to PVDF
membranes (Bio-Rad, Hercules, CA). Membranes were blocked with 5% nonfat
dry milk in 0.1% Tween/Tris-buffered saline (TBS-T), and incubated overnight at
4° C in 2.5% milk in TBS-T containing polyclonal rabbit anti-annexin A5 antibodies
(1:500; Santa Cruz Biotechnology, Santa Cruz, CA). Membranes were washed
five times with TBS-T and incubated with horseradish peroxidase-conjugated
goat anti-rabbit IgG (1:3000; Zymed Laboratories, S. San Francisco, CA) for 1 hr.
Proteins were visualized using enhanced chemiluminescence (ECL; Amersham,
UK).

82
4.2.9 Transgenic C. elegans strains
To make the annexin A5 expression construct, a PCR amplified full length
human annexin A5 cDNA fragment was subcloned into NheI/KpnI sites of the
expression vector pPD40.26 (gift of Andrew Fire) containing the myo-3 promoter.
The arrays vjEx171 and vjEx172 were generated by co-injecting this annexin A5
expression construct (25ng/ul) together with the injection marker KP# (40ng/ul)
into the a Punc-54::synuclein::YFP-expressing strain NL5901 (a gift of Ellen
Nollen). These arrays were subsequently crossed into a punc-54::YFP UA52
(gift of Guy Caldwell ). Young adult animals were paralyzed using sodium azide
and mounted on 2% agarose pads for imaging. Images were acquired on a Nikon
90i microscope using a Nikon Plan Apo 60× objective (NA = 1.4) and an
Coolsnap ES2 (Photometrics) camera controlled by Metamorph software
(Universal Imaging/Molecular Devices). Body wall muscles in the pharynx were
imaged from ~30 laterally oriented animals per genoptype. A maximum intensity
projection was obtained from image stacks of muscles. Images were
thresholded in Metamorph and the number of puncta above threshold
(determined by background muscle fluorescence) were counted.

4.2.10 Western Blotting for synuclein expression in annexin A5 transgenic
C. elegans
Western blotting was performed to determine if annexin A5 protein
expression in C. elegans alters the expression of α-synuclein. Whole-worm
lysates were treated with a radio immuno-precipitation assay to extract total
83
protein. One ug/uL of protein was separated on a 12% SDS-PAGE and
transferred to PVDF membranes (Bio-Rad, Hercules, CA). Membranes were
blocked with 10% nonfat dry milk in 0.1% Tween/Tris-buffered saline (TBS-T),
and incubated with TBS-T containing polyclonal rabbit anti- α-synuclein antibody
(1:1,000; Chemicon International, Temecula, California) and monoclonal mouse
anti-actin antibody (Calbiochem, Darmstadt, Germany). Membranes were
washed five times with TBS-T and incubated with horseradish peroxidase-
conjugated goat anti-rabbit IgG (1:3000; Zymed Laboratories, S. San Francisco,
CA) and sheep anti-mouse IgG (Amersham, UK) for 1 hr. Proteins were
visualized using enhanced chemiluminescence (ECL; Amersham, UK).

4.2.11 Western blotting for GFP expression in annexin A5 transgenic C.
elegans
To assess if the presence of annexin A5 transgene alters the expression
of GFP, whole-worm lysates of vjEx171 arrays in UA52 animals were treated with
a radio immuno-precipitation assay to extract total protein. Total protein lysate
was separated on a 12% SDS-PAGE and transferred to PVDF membranes (Bio-
Rad, Hercules, CA). Membranes were blocked with 10% nonfat dry milk in 0.1%
Tween/Tris-buffered saline (TBS-T), and incubated with TBS-T containing
monoclonal anti-GFP antibody (1:1000 Invitrogen, Inc) or actin antibody.
Membranes were washed five times with TBS-T and incubated with horseradish
peroxidase-conjugated sheep anti-mouse IgG (Amersham, UK) for 1 hr. Proteins
were visualized using enhanced chemiluminescence (ECL; Amersham, UK).
84

4.3 Results
4.3.1 Annexin A5 alters amyloid protein misfolding (EPR)
In previous studies (Chapter 3), we showed that annexin A5 directly binds
to hIAPP, alters its misfolding and inhibits hIAPP toxicity in tissue culture. From
these results, I asked whether annexin A5 can be a more general inhibitor of
amyloid misfolding by extending our investigation to include Aβ and α-synuclein.
We sought to determine whether annexin A5 can alter amyloid misfolding in vitro
using a combination of biophysical methods including, electron paramagnetic
resonance (EPR), electron microscopy (EM) and Thioflavin T (ThT) fluorescence.
To investigate the notion that annexin A5 has a direct impact on amyloid
misfolding and fibril formation, I employed site-directed spin labeling and EPR
spectroscopy. The EPR experiments were based on the introduction of a single
spin label (R1) into these peptides. Previous spin labeling studies on hIAPP, Aβ,
α-synuclein, tau and the human prion protein have shown that fibrils from these
proteins give rise to a parallel, in-register structure in which same residues stack
on top of each other (Török M 2002; Jayasinghe and Langen 2004; Margittai and
Langen 2006). As illustrated with the examples of hIAPP (21R1), Aβ (35R1) and
α-synuclein (52R1), the stacking of the same residues from different molecules
gives rise to strong spin-spin interaction (Fig. 4.1 A-C black spectra). This
interaction is reflected by a predominantly single line EPR spectrum (rather than
the typically observed 3 line spectrum) that is caused by the rapid exchange
between electrons from stacked spin labels.
85
Fig. 4.1: EPR spectra of R1-labeled h-IAPP, Aß and a-synuclein
in the presence of annexin A5
To assess the effect of annexin A5 on amyloid misfolding, 20uM of
annexin A5 was added to 100uM concentration of spin-labeled (A)
h-IAPP (21R1), (B)Aß (35R1) and (C) a-synuclein (52R1) peptide
(red spectra) and allowed to form fibrils. A sample of a-synuclein
was also incubated with lysozyme in order to control for the effect of
crowding. (D) EPR spectra were compared to control (black
spectra). All spectra were normalized to the same number of spins.
86
In order to assess whether annexin A5 can interfere with the formation of these
specific structures, I incubated 100uM of spin-labeled Aβ (35R1) and α-synuclein
(52R1) with a sub-stoichiometric amount (20uM) of annexin A5. For the purpose
of illustrating the general effect of annexin A5 on different amyloid proteins, the
EPR spectrum of hIAPP (21R1) with and without annexin A5 has been
reproduced from Chapter 3 (Fig. 3.5) as a part of Fig. 4.1A. The EPR spectra of
the aggregates that formed in the presence of annexin A5 after 4 days of
incubation at room temperature were strikingly altered and characterized by a
loss of the pronounced spin-spin interaction (Fig. 4.1B-C, red spectra). Thus,
these data illustrate that the structures of the hIAPP, Aβ and α-synuclein
aggregates formed in the presence of annexin A5 are different from those formed
in the absence of annexin A5. Again, the strong effect was observed using sub-
stoichiometric amounts of annexin A5 (5-fold excess of amyloid protein) further
supporting the notion that annexin A5 does not act by interacting
stoichiometrically with monomeric amyloidogenic proteins, but that it is likely to
interact with a sub population that is on the aggregation pathway and that is
present only at low concentrations. As a negative control, I incubated α-
synuclein (52R1) with a protein, lysozyme (chick egg), that should not have any
direct interaction with amyloid proteins and assessed whether there were any
structural changes. The resulting EPR spectrum (Fig. 4.1D) shows that the
addition of lysozyme to amyloid protein does not alter the misfolding.
Overall, the EPR data indicate that annexin A5 causes structural
alterations during the hIAPP, Aβ and α-synuclein misfolding process. We wanted
87
to understand if this interaction is specific for annexin A5 protein or whether it is a
more broad interaction between all annexin and amyloid proteins. Thus, I
incubated hIAPP (21R1), Aβ (35R1) and α-synuclein (52R1) in the presence of
annexin A2, annexin A5, annexin A6 and annexin B12 in a 1:5 (annexin:amyloid)
ratio. In the case of hIAPP (21R1), the addition of annexin A2, annexin A6 and
annexin B12 show an insignificant change from control fibrils (Fig. 4.2 green
spectra). The only change seen in the hIAPP samples is when hIAPP is
incubated with annexin A5. The Aβ fibrils show insignificant changes in the
presence of annexin A6 and annexin B12 (Fig. 4.2 black spectra). There is an
expected change with annexin A5 (as seen in Fig. 4.1) and also a nominal
change when Aβ fibrils are grown in the presence of annexin A2. The α-synuclein
fibrils show a significant change in EPR spectra in the presence of annexin A5
and annexin A6 and no change when in the presence of annexin A2 (Fig.4.2 red
spectra). While it appears that other annexins, specifically annexin A2 and
88
Fig. 4.2: EPR Spectra of amyloids with various annexin proteins
IAPP(21R1), Aß(35R1) and a-synuclein(52R1) were incubated in
the presence and absence (control spectra) of annexin A2, annexin
A5, annexin A6 and annexin B12.
89
annexin A6, may have the potential to interact with and change the misfolding of
amyloid proteins, it is clear that there is a stronger predilection for annexin A5 to
bind to all amyloid proteins tested. Overall, the EPR data indicate that annexin
A5 causes structural alterations during the hIAPP, Aβ and α-synuclein misfolding
process.

4.3.2 Annexin A5 reduces fibril formation (EM) and thioflavin T
fluorescence of α-synuclein
In addition to the structural changes seen by EPR, our findings were
further supported by electron microscopy (EM) images taken of amyloid samples
incubated with and without annexin A5. Wildtype, full-length unlabeled α-
synuclein was incubated with annexin A5 in a 3:1 (amyloid: annexin A5) ratio and
allowed to undergo fibrillization. After 4 days of incubation, EM images show that
the addition of annexin A5 dramatically alters the fibril morphology (Fig. 4.3D). In
Fig. 4.3 A and B we show the same hIAPP images discussed in Chapter 3 (Fig.
3.4) involving hIAPP and annexin A5 in order to illustrate the reproducibility of
adding annexin A5 to different amyloid proteins. The control samples of both
hIAPP and α-synuclein show typical long, unbranched fibrils (Fig. 4.3A,C), but
the samples with annexin A5 show a complete loss of fibrillar structure (Fig.
4.3B,D) and an increase in amorphous material. The striking similarities that are
seen when annexin A5 is added to different amyloid proteins, such as hIAPP and
α-synuclein suggests that annexin A5 may play a more general role in changing
amyloid protein structure and morphology.
90
Fig. 4.3: Electron Micrographs of amyloid fibrils with and
without annexin A5
Electron Micrographs of hIAPP in the absence (A) and presence of
annexin A5 (B). Panel A and C show 100uM hIAPP and α-
synuclein fibrils, respectively, grown in the absence of annexin A5
with 20 mM Hepes 100 mM NaCl buffer. The presence of
long,unbranched fibrils are seen. Panel B and D show hIAPP and
a-synuclein with annexin A5 in a 3:1 ratio. Large protein
aggregates are seen throughout the sample when amyloid is in the
presence of annexin A5.
91
Fig. 4.4: Thioflavin T curves of amyloid fibrils with and without
annexin A5
Time-dependent normalized ThT emission curves of a-synuclein (100uM)
with (red curve) and without annexin A5 (20uM) (black curve). The ThT
curves are normalized to the maximal observed intensity at the end of
each aggregation reaction.
92
To further our study on how annexin A5 directly interacts with and changes
amyloid proteins, I used Thioflavin T (ThT) assays to monitor the effect of
annexin A5 on α-synuclein fibril formation. This commonly used assay is based
upon the fluorescence enhancement that occurs as ThT binds to amyloid fibrils
such as those formed by α-synuclein (Naiki H 1989; Kudva, Mueske et al. 1998).
Consistent with previous reports (Padrick and Miranker 2002) the time-
dependent kinetics of amyloid fibrillization follows a sigmoidal pattern (Fig. 4.4
black spectra). We added annexin A5 and α-synuclein in a 1:5 ratio and
monitored the ThT fluorescence over several hours. With the addition of annexin
A5, the time-dependent increase in thioflavin T fluorescence is largely prevented.
This reduction indicates that annexin A5 alters the misfolding pathway and
inhibits the formation of thioflavin positive α-synuclein fibrils.

4.3.3 Annexin A5 expression reduces α-synuclein inclusions in vivo
We have shown that annexin A5 has the ability to alter the misfolding of
several amyloid proteins and reduce the islet cell cytotoxicity associated with
hIAPP. Considering that annexin A5 reduces hIAPP cytoxicity in islet cells
(chapter 3), we employed another intracellular amyloid protein, α-synuclein, in
our in vivo model involving C. elegans. We determined the effect of annexin A5
expression on the formation of α-synuclein inclusions in C. elegans. α-synuclein
inclusions can be visualized in C. elegans body wall muscles as fluorescent
puncta in animals expressing α-synuclein-YFP fusion proteins (α-synuclein::YFP)
(Hamamichi S 2008; van Ham TJ 2008). We expressed full length human
93
Fig. 4.5: Annexin A5 expression decreases a-synuclein
inclusions in vivo
(A) Top panel: representative fluorescent images of head
muscles of transgenic young adult animals either expressing a-
synuclein::YFP alone (the pkIs2386 transgene) or a-
synuclein::YFP with human annexin A5 (pkIs2386; vjEx137).
(B) Middle panel: False colored images from (A) with red and
purple representing the highest and lowest pixel intensities
respectively. (C) Bottom panel: thresholded images from (A)
showing pixel intensities above background in green. (D)
Quantification of the number of inclusions above threshold per
pharynx averaged over 20 animals. Data from two
independently generated transgenic lines expressing annexin
A5 is shown. Standard error is shown (p < xxx)
94
Fig. 4.6: Abundance of a-synuclein in annexin A5 transgenic C.
elegans
A) Immunoblot of a-synuclein from lysates of: lane 1: control wild type
animals (N2), lane 2: animals expressing an a-synuclein transgene
(pkIs2386) and lane 3: animals co-expressing a-synuclein and annexin A5
transgenes (vjEx137; pkIs2386). The samples were subjected to SDS-
PAGE and western blotting with anti-a-synuclein and anti-actin antibodies.
B) Average abundance of a-synuclein normalized to actin. C) Immunoblot
of GFP from total protein extracts of animals expressing the following
transgenes: lane 1: GFP driven by the unc-54 promoter in muscles
(baInl12), lane 2: GFP and annexin A5 expressed in muscles bt he unc-54
and myo-3 promoters, respectively (baInl12; vjEx137). The samples were
subjected to SDS PAGE and western blotting with anti-GFP and anti-actin
antibodies to detect whether the annexin A5 transgene alters the
95
annexin A5 cDNA in body wall muscles (using the myo-3 promoter) in animals
overexpressing synuclein::YFP (under the unc-54 promoter) (van Ham TJ 2008)
and compared the number of inclusions to control animals that were not
expressing annexin A5. We quantified the number of fluorescent puncta in
muscles in the head region that were brighter than background muscle
fluorescence. We found that muscles of animals expressing annexin A5 had a
reduced number of fluorescent puncta compared to non transgenic controls (Fig.
4.5). This effect is consistent with a reduction in the formation of inclusions by
annexin A5, but could potentially be due to lower α-synuclein protein levels in
muscles. Thus, as a control for this, we examined the abundance of
synuclein::YFP in animals expressing annexin A5 compared to controls. We
found no change in normalized synuclein::YFP protein levels in whole animal
lysates (Fig. 4.6 A, B). To test whether annexin A5 alters α-synuclein
transcription, we examined if annexin A5 expression could alter the abundance of
another protein, GFP, expressed under the same promoter used to drive α-
synuclein expression. We found no change in GFP abundance in whole animal
lysates from annexin A5 expressing animals compared to controls (Fig. 4.6C, D),
confirming that annexin A5 does not alter the expression of α-synuclein from this
promoter in these animals. Overall, our results show that annexin A5 alters
amyloid aggregation in vivo by decreasing the fluorescently-labeled α-synuclein
aggregates without changing the transcription of this protein.

96

4.4 Discussion
We use in vitro biophysical, tissue culture and an animal model to show
that annexin A5 interacts with and inhibits the misfolding of amyloid proteins. In
this study I first use EPR spectroscopy to show that annexin A5 alters the
misfolding from the typical single-line EPR spectrum found in amyloid fibrils to an
unordered (non-fibrillar) species. The corresponding EM images confirm that the
addition of annexin A5 alters α-synuclein morphology into a non-fibrillar structure.
The last of the biophysical assays show that the addition of annexin A5
significantly decreases the ThT fluorescence intensity of α-synuclein, suggesting
an inhibition of fibril formation. The transgenic expression of annexin A5 in C.
elegans overexpressing synuclein::YFP results in a decrease of the α-synuclein
aggregated puncta without seeing a decrease in expressed protein. This
reduction of puncta can either be caused by the reduction of misfolded and
subsequent aggregation of amyloid protein and/or an increase in the degradation
of misfolded proteins.
Interestingly, annexins are known to bind to vesicles as well as detergent
micelles (Meers P 1993; Schmidt 2008). Amyloid oligomers have also been
referred to as micelle-like intermediates of misfolding (Yong W 2002). This
suggests that annexin A5 may have an inherent propensity to bind to amyloid
oligomers and change its misfolding pathway.
The structural changes and protective effect we observed both in this
study and in previous studies (chapter 3) suggests that annexin A5 may play a
97
significant physiological role in amyloid diseases. Thus, future animal model
systems as well as proteomic screens of amyloid deposits could help pave way
towards therapeutic interventions.

98

Chapter 5: The effect of Curcumin on hIAPP
toxicity and aggregation
5.1 Introduction
The clear link between amyloid misfolding and toxicity prompts the
investigation of inhibitors of the misfolding pathway. Inhibitors of these misfolded
intermediates can reduce toxicity associated with amyloid diseases by blocking
the toxic species from forming. Thus, it is important to identify modulators and
inhibitors of the misfolding pathway as they may give rise to potential therapy for
patients with amyloid diseases. In previous chapters, I discussed the role
annexin A5 plays in reducing endogenous amyloid misfolding and toxicity. In this
chapter, I focus on the effect of curcumin, a small molecule, on hIAPP toxicity
and aggregation.
Curcumin (diferulomethane) is a biphenolic small molecule and the main
constituent of the rhizome C. longa (turmeric). It is known to have many
therapeutic effects including anti-inflammatory, anti-oxidant and anti-HIV effects
(Ammon HP 1991; Mazumder A 1995; Aggarwal BB 2003). Recent studies have
also shown that curcumin has anti-amyloidogenic effects on Aβeta (Yang, Lim et
al. 2005; Garcia-Alloza, Borrelli et al. 2007), α-synuclein (Pandey, Strider et al.
2008) and prion proteins (Hafner-Bratkovic, Gaspersic et al. 2008) as it relates
to Alzheimer disease, Parkinson disease and transmissible spongiform
encephalopathies, respectively. While these studies have shown the anti-
99
amyloidogenic effects of curcumin, it is still poorly understood how curcumin
interacts with amyloid proteins and what mechanism it uses to inhibit misfolding.
Additionally, it is important to know at what physiological concentrations curcumin
is protective and at what concentrations it becomes toxic to cells. This prompted
us to investigate the molecular interaction of curcumin with hIAPP and test its
effect on INS cells and HIP rat islets. First I sought to understand if curcumin
inhibits or changes the misfolding of hIAPP in vitro using site-directed spin
labeling and electron paramagnetic resonance (EPR). In order to further monitor
the molecular interaction of curcumin and hIAPP and its effect on IAPP
misfolding Thioflavin T (ThT) fluorescence assays were used. We also
demonstrated the in vivo effects of curcumin and hIAPP using INS cells and HIP
rat islets.

5.2 Materials and Methods
5.2.1 Chemicals and Peptides
Hexafluoroisopropanol (HFIP) and dimethyl sulfoxide (DMSO) were
obtained from Sigma-Aldrich. Curcumin was obtained in powdered form from
Sigma-Aldrich. Synthetic wild-type human IAPP (lot # E-0509) was purchased
from Polypeptide laboratories (Wolfenbuettel, Germany). The spin label, 1-oxyl-
2,2,5,5-tetramethyl- 3-pyrroline-3-methyl methanethiosulfonate (MTSL), was
obtained from Toronto Research Chemicals (Toronto, Canada). Single cysteine
mutants of full-length IAPP were purchased from Biomer Technology
100
(Pleasanton, CA). The native cysteine residues at position 2 and 7 are
substituted with an alanine (Jayasinghe and Langen 2004)

5.2.2 Spin Labeling and Fibril Formation for EPR
For spin labeling, single cysteine containing peptides were reacted with 3-
fold molar excess of MTSL for 1 h at room temperature. Unreacted MTSL was
removed using a Toyopearl cation exchange column. The spin-labeled peptide
was then desalted using a C18 reverse phase SpinColumn from Harvard
Apparatus (Holliston, Massachusetts) and eluted with 100% HFIP. Peptide
concentrations were calculated by UV absorbance at 280 nm in 6 M guanidine
HCl using an extinction coefficient of 1400 M-1 cm-1. Labeled peptides were
stored at -80 ° C until use. Stock solutions of spin-label ed IAPP and wildtype
IAPP were lyophilized. For preparation of stock solutions the lyophilized peptides
were reconstituted with deionized water containing 0.5 % acetic acid. For fibril
formation, aliquots of stock IAPP was added to 10mM Phosphate buffer with
100mM NaCl (pH 7.4) for a total volume of 100uL and 100uM concentration.
Curcumin was freshly dissolved in DMSO prior to use. The samples were
incubated with or without curcumin (1-100uM) at room temperature for 4 days to
allow for fibril formation. To test the effect of curcumin on preformed fibrils, 16R1
labeled IAPP was prepared to form fibrils for 3-7 days at a concentration of
100uM. Curcumin was added to the preformed fibrils, allowed to incubate for
either 1 hr or 24 hrs. The samples were then harvested for EPR spectroscopy.

101

5.2.3 EPR spectroscopy
For EPR Spectroscopy, fibril samples were pippetted into a glass capillary
(0.6 mm inner diameter, 0.84 outer diameter, VitroCom, mt. Lks., New Jersey)
that had been sealed at one end. The sample was centrifuged and all
supernatant was removed to eliminate smaller aggregates and monomeric
protein. EPR spectra of fibrils were recorded on a Bruker (Billerica, MA) EMX
spectrometer with a HS resonator at 12 mW incident microwave power with a
scan range of 150 Gauss.

5.2.4 Fibril Formation for Electron Microscopy:
To examine the effect of curcumin on fibril growth, 10uL of samples were
adsorbed onto carbon and formavar-coated copper grids and negatively stained
with 2% (w/v) uranyl acetate solution for 5 min. The stained grids were examined
and photographed using a JEOL JEM-1400 electron microscope at 100 kV.

5.2.5 Thioflavin Assay
To test the effect of curcumin on ThT fluorescence intensity of IAPP fibrils,
20uL aliquots of wildtype IAPP peptide solutions in 0.5% acetic acid were mixed
with 500uL of 10mM Phosphate buffer (pH 7.4) to yield 50uM IAPP
concentrations in the presence and absence of 10uM and 50uM curcumin
concentrations. To each fibrillization reaction, 25uM ThT was added and
102
emission intensities were measured at 482nm with excitation at 450nm for 16
hours. Measurements were performed at room temperature with excitation and
emission slit widths of 1 and 10nm, respectively. To assess whether curcumin
can bind to preformed IAPP fibrils and block ThT from binding to the fibril,
wildtype hIAPP fibrils were allowed to undergo complete fibrilization for 4 days in
10mM Phosphate buffer (pH 7.4). This preformed sample of hIAPP fibrils had a
total volume of 500uL at 100uM. After they reacted with ThT for 30 minutes the
samples were incubated with 0uM, 1uM, 10uM or 100uM concentrations of
curcumin and monitored for fluorescence changes for 8 hours. All fluorescence
measurements were carried out using a Jasco FP-6500 spectrofluorometer.

5.2.6 Cell culture
Rat insulinoma cell line INS 832/13 was kindly provided by Dr. C.
Newgard, Durham, NC. INS cells were grown in RPMI 1640 medium
supplemented with 10 mM HEPES, 1 mM sodium pyruvate, 100 IU/ml penicillin,
and 100 μg/ml streptomycin (Invitrogen, Carlsbad, CA), 10% heat-inactivated
FBS (Gemini, West Sacramento, CA), 50 μM ß-mercaptoethanol (Sigma, St.
Louis, MO), at 37° C in a humidified 5% CO2 atmosphere .

5.2.7 Adenovirus generation and transduction experiments
The complementary cDNA encoding the full-length human and rat prepro-
IAPP were used as templates to generate the human and rat IAPP adenovirus.
KpnI and XhoI or Ecor RI and EcoRV restriction sites were respectively
103
introduced in front of the ATG and after the stop codon of the human and rat
cDNA using PCR. A 290 bp human preproIAPP PCR fragment was digested with
KpnI and XhoI and a 300 bp rat preproIAPP PCR fragment was digested with
EcoRI and EcoRV. The fragments were ligated into the pENTR2B vector
(Invitrogen, CA) and the entry clones were subsequently subcloned into the
pAd/cytomegalovirus/DEST adenovirus vector (Invitrogen). Recombinant
adenovirus expressing human and rat preproIAPP (Ad-hIAPP and Ad-rIAPP,
respectively) were produced and purified according to the manufacturer's
instructions (Invitrogen). For the transduction experiments, INS cells were plated
on 96-well or 6-well tissue culture plates at a density of 2.5 x 104 or 1.0 x 106
cells/well respectively, and cultured overnight. Cells were transduced with r-IAPP
or hIAPP adenoviruses at 400 moi for 48 h in complete RPMI medium. Two
hours after cell transduction, curcumin or DMSO (vehicle) was added to cells at
desired concentrations. At the end of the transduction, cells plated on 96-well
plates were used to assess cell viability by MTT assay (see below) and cells
plated on 6-well plates were washed in PBS and lysed in NP40 buffer (50 mM
Tris-Hcl, pH 7.4, 250 mM NaCl, 1 mM EDTA, 0.5% NP40, 1 mM DTT and
protease inhibitors (Sigma)).

5.2.8 hIAPP preparation and toxicity assays
Lyophilized hIAPP was dissolved in 0.5% acetic acid to prepare a 2 mM
stock solution and vortexed for 30 sec. To assess the effect of curcumin on cell
death induced by exogenous hIAPP, INS cells were seeded on a 96 well plate at
104
5 x 104 cells/well and cultured for 24 h. The day after, the medium was replaced
by DME medium (Irvine scientific, Santa Ana, CA) containing either 15 μM of
freshly dissolved hIAPP or hIAPP solution previously incubated for 10 days at
4° C, vehicle or curcumin. After 20 h incubation, cell v iability was assessed.

5.2.9 MTT assay
The reduction of MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium
bromide) was used to assess cell viability. At the end of the incubation or
transduction time, cell medium was changed to 1 mg/ml MTT (Sigma) containing
medium according to the manufacturer's instructions. Colorimetric measurements
were made 2 h after addition of the MTT reagent at 570 nm with an ELISA plate
reader (Spectramax 250; Molecular Devices, Sunnyvale, CA). The background
wavelength at 690 nm was subtracted from the 570 nm measurement.

5.2.10 Rat islet Isolation
Pancreatic islets were prepared by collagenase digestion. Rats were bred
and housed at the animal housing facility of the University of California Los
Angeles. Five month-old WT (n=5) and HIP (n=6) rats were euthanized using
isoflurane. The bile duct was cannulated, and the pancreas was perfused with a
collagenase solution (HBSS supplemented with 25 mM HEPES (Invitrogen), 0.23
mg/ml liberase (Roche, Penzberg, Germany), 0.1 mg/ml DNAse (Roche)). The
pancreas was then removed, incubated for 20 min in collagenase solution at
37° C and dispersed by shaking for 30 sec. The suspension w as transferred into
105
30 ml of ice cold HBSS. After 5 min incubation on ice, ¾ of supernatant was
removed by aspiration, and fresh ice-cold HBSS added to the pellet. Islets were
washed 3 times, then manually picked and incubated in RPMI 1640 medium
supplemented with 100 μg/ml streptomycin and 10% heat-inactivated FBS at
37° C in a humidified 5% CO2 atmosphere. Rat islets wer e cultured for 48 h in the
absence or presence of curcumin and either lysed in NP40 buffer or used for
propidium iodide staining.

5.2.11 Propidium iodide staining (PI)
Aliquots of 30 islets were incubated with or without curcumin (0.5, 1, 2 and
5 μM). Subsequently, islets were stained with 50 μg/ml propidium iodide
(Molecular Probes, Eugene, OR) for 30 min at 37° C. I slets were then washed in
PBS and fixed in 4% paraformaldehyde for 30 min at room temperature. Islets
were imaged using a Leica DM600 microscope (Leica Microsystems, Wetzlar,
Germany) and images of 25 islets per condition were acquired using OpenLab
software (Improvision). The fractional area of the islets positive for PI was
digitally quantified using Image-Pro Plus software (Media Cybernetics, Silver
Spring, MD).

5.2.12 Western Blot analysis
Proteins were separated on a 4-12% Bis-Tris NuPAGE gel and blotted
onto a nitrocellulose membrane (Whatman, Germany). Membranes were probed
overnight at 4° C with pro and cleaved caspase 3, PARP , α-actin and GAPDH
106
antibodies (Cell Signaling Technology, Beverly, MA), followed by washes and 1 h
incubation with horseradish peroxidase-conjugated secondary antibodies (Zymed
Laboratories, South San Francisco, CA). Proteins were visualized by enhanced
chemiluminescence (Millipore) and protein expression levels were quantified
using the Labworks software (UVP, Upland, CA).

5.2.13 Statistical analysis
Data are presented as the means ± SE. Statistical analyses were carried
out by ANOVA followed by Duncan's post hoc test using Statistica (Statsoft,
Tulsa, OK). A P value of <0.05 was taken as evidence of statistical significance.

5.3 Results
5.3.1 Curcumin alters IAPP misfolding
We sought to test whether curcumin can modulate IAPP misfolding in vitro
using a combination of biophysical methods including electron paramagnetic
resonance (EPR), electron microscopy (EM) and ThT fluorescence. The EPR
experiments were based on the introduction of a single spin label into the
peptide. Previous spin labeling studies on IAPP, Aβeta, α-synuclein, tau and the
human prion protein have shown that fibrils from these proteins give rise to a
parallel, in-register structure in which same residues stack on top of each other.
As illustrated with the example of IAPP spin labeled at position 16 (Fig. 5.1D), the
stacking of same residues from different molecules gives rise to strong spin-spin
interaction (Fig. 5.1A). The strong spin-spin interaction is reflected by a
107
predominantly single line EPR spectrum (rather than the typically observed 3 line
spectrum) that is caused by the rapid exchange between electrons from stacked
spin labels. In order to test whether curcumin can affect the misfolding and fibril
formation of IAPP, spin-labeled IAPP was incubated with curcumin (1-100uM) for
4 days and the EPR spectra of the resulting IAPP aggregates were recorded.
The addition of 1uM and 10uM concentrations of curcumin to hIAPP results in
EPR spectra (Fig. 5.1B and 1C respectively) that show a subtle but detectable
change as indicated by the formation of additional peaks in the EPR spectra. At
100uM curcumin concentration, this change in the EPR is more pronounced,
(Fig. 5.1D, black arrow), and reflects a gradual loss of spin-spin interaction
indicating an alteration in the misfolding process. Moreover, the signal to noise is
significantly weaker in the samples containing curcumin. The loss of signal is
caused by the reduced amount of aggregated and misfolded hIAPP that could be
harvested under these conditions. Thus, the EPR data indicate that curcumin
causes structural alterations during the IAPP misfolding process and appears to
reduce the overall amount of deposits that are formed.
These findings are further supported by EM analysis of the same samples.
While the fibrils formed in the absence of curcumin have characteristic hIAPP
fibril morphology (Fig. 5.1E), the addition of curcumin in high concentrations
decreases fibril formation and morphology. The most significant changes are
observed at 100uM curcumin concentrations (Fig. 5.1H) resulting in a significant
decrease in the amount of fibrils as well as a change to shorter and more
108

Figure 5.1: EPR Spectra and EM images of 16R1 hIAPP in the presence
of curcumin
EPR spectra of 16R1 hIAPP incubated with 0uM (A), 1uM (B), 10uM (C), and
100uM (D) curcumin. All spectra were rescaled and are shown at the same
amplitude. The corresponding EM images of 16R1 IAPP incubated with 0uM
(E), 1uM (F), 10uM (G), 100uM (H) curcumin are also shown.
109
Figure 5.2: Thioflavin T Assay for hIAPP in the presence and
absence of curcumin
hIAPP was incubated with curcumin at 10uM (red line) and 50uM
(blue line) during fibril formation. The addition of curcumin
decreases fibril formation.
110
dysmorphic structures. Thus, the EPR and the EM data demonstrate that
micromolar concentrations of curcumin have a strong effect on hIAPP misfolding.
In addition, I also used ThT fluorescence to monitor the effect of curcumin
on IAPP misfolding and fibril formation. This commonly used assay is based
upon the fluorescence enhancement that occurs as ThT binds to amyloid fibrils
such as those formed by hIAPP (Naiki H 1989; Kudva, Mueske et al. 1998). As
expected from the aforementioned curcumin-dependent reduction of hIAPP fibril
formation, the ThT fluorescence of hIAPP is strongly reduced in the presence of
10uM and 50uM, of curcumin (Fig. 5.2). While these data are consistent with the
EPR and EM data, it is important to note that ThT fluorescence, in contrast to EM
and EPR, is an indirect readout that cannot readily distinguish between a
reduction in fibril formation or competitive inhibition of ThT binding to the fibrils
(see below).

5.3.2 The effect of curcumin on preformed fibrils
Having shown that curcumin has pronounced effects on hIAPP misfolding
when added during the misfolding process, I next tested whether curcumin might
also be similarly efficient at reversing fibril formation when added to preformed
fibrils. In order to address this question, we again turned to the EPR based assay
using IAPP spin labeled at position 16. As shown in Fig. 5.3, the addition of
curcumin to pre-formed, spin-labeled hIAPP fibrils did not significantly change
their EPR spectra after 1hr (Fig. 5.3B) or 24hr (Fig. 5.3C). Thus, curcumin did not
rapidly dissolve fibrils under the present conditions. This finding was further
111
Figure 5.3: Curcumin does not change the fibrillar structure of preformed
hIAPP
A) hIAPP control (B) hIAPP incubated with 100uM curcumin for 1 hour (C)
hIAPP incubated with 100uM curcumin for 24 hours (D) overlay of hIAPP fibrils
(16R1) incubated with 100uM curcumin for 24 hours (red line) and compared to
control fibrils (black line). Each spectrum was normalized to represent the same
number of spins. E) EM of hIAPP control (F) EM of hIAPP incubated with 100uM
curcumin for 24 hours.
112
Figure 5.4: Thioflavin assay for hIAPP preformed fibrils in the
presence and absence of curcumin
hIAPP was incubated with curcumin at 1uM (blue line), 10uM (green
line), and 100uM (red line) and compared to control (black line).
Time =0 indicates the point at which curcumin was added to the
sample.
113
supported by negative staining EM, which revealed largely unchanged fibril
morphology and density (Fig. 5.3 E and 5.3F).In contrast to the negligible
changes seen by EPR and EM, the ThT assay resulted in more pronounced
changes. In these experiments curcumin was added to a mixture of hIAPP fibrils
that had been preincubated with ThT for 30 minutes. While control fibrils (Fig. 5.4
black line) exhibited a steady level of fluorescence intensity, the addition of 1uM
curcumin (Fig. 5.4 blue line) showed a detectable decrease in fluorescence
intensity over time. Even more significant changes in fluorescence were
observed upon addition of 10uM (Fig. 5.4 green line) and 100uM concentrations
of curcumin (Fig. 5.4 red line). Interestingly, the strong reduction in ThT
fluorescence occurred within only a few hours, at a time when no significant
changes in fibril structure and overall amount could be observed by EM or EPR.
The rather rapid change in ThT fluorescence in the absence of detectable
structural changes in the fibrils suggests that curcumin directly interferes with
ThT fluorescence. Such interference could be brought about by a direct
competition between curcumin and ThT for the same binding sites on the fibril. In
fact, this mechanism has previously been reported in the case of ThT binding to
Aß fibrils (Yang, Lim et al. 2005) and this notion is supported by our findings that
fibrils which are incubated with curcumin and harvested by centrifugation have a
yellow appearance due to the binding of curcumin (data not shown).

114
Figure 5.5: Curcumin fails to prevent apoptosis induced by h-IAPP
overexpression in INS cells
A) Effect of 48 h curcumin treatment on INS cell viability (MTT assay). B, C)
Curcumin does not protect INS cell against apoptosis induced by h-IAPP
adenoviral overexpression (400 moi, 48h) as assessed by MTT assay (B),
cleaved caspase 3 and cleaved PARP immunoblotting (C). D) Western blot
quantification. Data are expressed as mean ± SE (n=4). *P<0.05, **P<0.01,
**P<0.0001, significant differences vs. control.
115

5.3.3 Curcumin partially protects INS cells against exogenous hIAPP
toxicity
Having demonstrated that curcumin affects hIAPP misfolding and
morphology; we sought to establish whether it could protect β-cells against
exogenous hIAPP toxicity. At first, we determined if curcumin by itself could
influence β-cell survival and therefore assessed cell viability in response to
increasing concentrations of curcumin. Curcumin concentrations between 0 to 10
μM did not affect cell viability, however after 20 h, the higher concentrations of
curcumin decreased cell survival by 25% (Fig. 5.5A). We then assessed the
potential protective effect of curcumin on exogenous hIAPP-induced apoptosis. A
2 mM stock solution of hIAPP was prepared in 0.5% acetic acid. This solution
was applied to the cells alone or in combination with curcumin. After a 20 h
incubation with hIAPP (15 μM), INS cell viability decreased by 75% (Fig. 5.5B).
Low concentrations of curcumin (0 to 5 μM) did not influence the deleterious
effect of hIAPP, whereas higher concentrations, 10 to 25 μM significantly
reduced hIAPP toxicity (Fig. 5.5B). Micromolar concentrations of curcumin are
thus able to interfere with hIAPP misfolding and toxicity. However, the most
significant protective effect of curcumin is reached at a concentration of 25 μM,
which is in itself cytotoxic (Fig. 5.5A) to INS cells.

116
Figure 5.6: Curcumin does not protect HIP rat islets overexpressing
h-IAPP from apoptosis
Wt and h-IAPP transgenic (HIP) rat islet viability was assessed after 48 h
treatment with or without curcumin (0.5, 1, 2 and 5 μM) by PI staining (A),
cleaved caspase 3 and cleaved PARP immunoblotting (B). C) Western blot
quantification. Data are expressed as mean ± SE (n=4). *P<0.05,
**P<0.01, **P<0.0001, significant differences vs. Wt + vehicle. # P<0.05
significant differences vs. vehicle treatment.
117
5.3.4 Curcumin does not prevent endogenous hIAPP-induced apoptosis
We next investigated a potential protective effect of curcumin against
endogenous hIAPP-induced apoptosis (perhaps need to explain the rationale
here-after all IAPP is secreted and extracellular). A new toxicity assay was
performed on INS cells treated with curcumin alone in the time frame required to
detect β-cell death induced by endogenous hIAPP. Treatment of INS cells with
curcumin for 48 h at concentrations from 0 to 1 μM did not affect cell viability,
whereas a dose-dependant decrease in cell viability was found at concentrations
from 2 to 100 μM (Fig. 5.6A). Since curcumin treatment is independently
cytotoxic at concentrations higher than 2 μM, the highest nontoxic concentrations
were used to determine a potential protective effect of curcumin on cell death
induced by overexpression of hIAPP. The potential protective effect of curcumin
was assessed in INS cells transduced with an adenovirus expressing hIAPP or r-
IAPP as control (400 moi). After 48 h transduction, hIAPP overexpression led to
40% decrease in cell viability when compared to cells transduced with rIAPP
(Fig. 5.6B). Addition of curcumin at 0.1, 0.5 and 1 μM for the last 46 h of
transduction did not exert any protective effect on hIAPP-induced cell death (Fig.
5.6B). These observations were confirmed with the analysis of the apoptotic
markers, caspase 3 and PARP by western blotting. Endogenous hIAPP
expression in INS cells induced apoptosis as illustrated by an increase in the
cleavage of both caspase 3 and PARP (Fig. 5. 6C and 6D). Consistent with the
cell viability assay, curcumin treatment did not prevent caspase 3 and PARP
118
cleavage induced by the endogenous expression of hIAPP (Fig. 5. 6C and 6D).
Thus, curcumin failed to protect INS cells against hIAPP-induced apoptosis .

5.3.5 Curcumin does not protect HIP rat islets against apoptosis
We then examined the effect of curcumin on cell death in pancreatic islets
isolated from hIAPP transgenic (HIP) rats, a rat model for type 2 diabetes. Islets
were isolated from 5 month-old control rats (Wt) or HIP rats, which, at this age,
display an increased ß-cell apoptosis but do not yet have diabetes. In
accordance with previously published data (ref: Alexandra diabetes 2004, Juris
rifampicin, Andrew diabetes 2007), the extent of cell death estimated here by the
fractional area positive for PI staining was higher in HIP rat islets (10.3% ± 1.9)
than in Wt islets (4.8 ± 0.5, Fig. 5. 7A). No protection from cell death induced by
the hIAPP transgene was detected after curcumin treatment (Fig. 5. 7A). To the
contrary, at 1 μM and higher concentrations, curcumin tends to exert cytotoxic
effects in both Wt and HIP islets (Fig. 5. 7A). Moreover, no protective effect of
curcumin against apoptosis has been observed in HIP islets as assessed by
caspase 3 and PARP cleavage (Fig. 5. 7B). Indeed, a 3-fold increase in cleaved
caspase 3 and cleaved PARP was detected in HIP vs. Wt islets. However,
neither caspase 3 nor PARP cleavage was affected by curcumin, at any
concentration (Fig. 5.7B).
119
Figure 5.7: Curcumin prevents INS cell cytotoxicity
At high concentration, curcumin partially prevents INS cells against
apoptosis when h-IAPP, previously incubated for 10 days at 4° C is
exogenously applied to the cells. A) Effect of 24 h curcumin treatment
on INS cell viability (MTT assay). B) Viability of INS cells
was assessed by MTT assay after 20 h treatment with a solution of h-
IAPP freshly dissolved (B) or incubated 10 days at 4° C (C), at a final
concentration of 15 μM, alone or in combination with curcumin. Data
are expressed as mean ± SE (n=3). *P<0.05, **P<0.01, ***P<0.0001,
significant differences, curcumin vs. vehicle in h-IAPP treated cells.
120
5.4 Discussion
Curcumin has been suggested as a promising therapeutic agent for a
number of diseases, including a few that involve amyloid misfolding, such as
Alzheimer disease and Parkinson disease. In this study, we sought to determine
if curcumin alters the misfolding and toxicity of hIAPP, as it relates to T2DM, and
whether it does so at concentrations that are non-toxic to pancreatic β-cells.
With the use of EPR, EM and islet cell tissue culture we found that 10-25uM
concentrations of curcumin are required to significantly alter misfolding and
prevent hIAPP toxicity. However, we also found that these same concentrations
of curcumin were in itself toxic to the both INS cells and pancreatic islets isolated
from HIP rats. This suggests that the effective dose of curcumin is physiologically
harmful to β-cells. In the HIP islet cells, this is a physiologically relevant cell
culture because it expresses hIAPP intracellularly and mimics the cellular
environment of diabetes. It also appears that the benefit of curcumin does not
outweigh its harm. In fact, curcumin offered no protection from cell death and
apoptosis and instead exerted cytotoxic effects at 1uM concentrations.
In addition to finding the effective concentration required for curcumin to
alter misfolding and toxicity, we also discovered an important caveat when using
ThT assays to monitor fibril formation. In our studies, we found a reduced ThT
fluorescence when hIAPP fibrils are in the presence of curcumin. Knowing that
ThT binds to channels located in between β-sheets of the protofilaments (Krebs,
Bromley et al. 2005; Groenning M 2007), any subtle changes to these channels,
such as the binding of curcumin, could result in reduced ThT binding. An
121
important point to consider is that this reduction could either be due to actual fibril
reduction or by ThT displacement. In our studies, we used EPR and EM to
confirm the presence of fibrils (Fig. 5. 3) even when the ThT fluorescence was
significantly reduced (Fig. 5. 4), thus implying a ThT displacement by curcumin.
Taking both the biophysical and biological aspects of this study into
consideration we have found that curcumin has the ability to alter hIAPP
misfolding and fibril formation, while also showing protective effects against
hIAPP cytoxicity at ranges that in itself seem harmful to the cells. One possible
mechanism of binding might involve curcumin interacting with the hydrophobic
wall of hIAPP fibrils that was discussed in Chapter 2. Perhaps, curcumin can bind
to these “sticky-ends” and block the propagation of fibril formation. Nonetheless,
curcumin clearly has the ability to interact with and inhibit the formation of
amyloid proteins, such as Aβ (Yang, Lim et al. 2005) and hIAPP, but it does not
seem to have the therapeutic characteristics that would make it an ideal drug.
This finding should not necessarily discount curcumin as a treatment for amyloid
diseases, but should instead prompt the development of an analog for curcumin.
An ideal analog for curcumin would have the ability to bind amyloid proteins and
inhibit their toxicity while being delivered in a physiologically efficient manner and
dosage.

122

References
Aaron K. Chamberlain*, Cait E. MacPhee*, 1, Jesús Zurdo*, 1, Ludmilla A.
Morozova-Roche*, 1, H. Allen O. Hill†, Christopher M. Dobson* and Jason J.
Davis (2000). "Ultrastructural Organization of Amyloid Fibrils byAtomic Force
Microscopy." Biophys J 79(6): 3282-3293.
Aggarwal BB, K. A., Bharti AC (2003). "Anticancer potential of curcumin:
preclinical and clinical studies." Anticancer Res 23(1A): 363-98.
Ahren, B., C. Oosterwijk, et al. (1998). "Transgenic overexpression of human islet
amyloid polypeptide inhibits insulin secretion and glucose elimination after gastric
glucose gavage in mice." Diabetologia 41(11): 1374-80.
Ammon HP, W. M. (1991). "Pharmacology of Curcuma longa." Planta Med 57(1):
1-7.
Anant K. Paravastua, Isam Qahwashb, Richard D. Leapmanc, Stephen C.
Meredithb and Robert Tyckoa,2 (2009). "Seeded growth of ß-amyloid fibrils from
Alzheimer's brain–derived fibrils produces a distinct fibril structure." Proc Natl
Acad Sci U S A.
Anguiano, M., R. J. Nowak, et al. (2002). "Protofibrillar islet amyloid polypeptide
permeabilizes synthetic vesicles by a pore-like mechanism that may be relevant
to type II diabetes." Biochemistry 41(38): 11338-43.
Bates, G. (2003). "Huntingtin aggregation and toxicity in Huntington's disease."
Lancet 361: 1642-4.
Borbat P, R. T., Freed JH, Elizer D (2006). "Inter-helix distances in
lysophospholipid micelle-bound alpha-synuclein from pulsed ESR
measurements." J Am Chem Soc 128: 10004-10005.
Brian O'Nuallain, A. D. W., Per Westermark, and Ronald Wetzel (2004).
"Seeding Specificity in Amyloid Growth Induced by Heterologous Fibrils." J Biol
Chem 279(17): 17490-17499.
Brunger AT, A. P. (2002). "Molecular dynamics applied to X-ray structure
refinement." Acc Chem Res. 35(6): 404-12.
Butler, A. E., J. Jang, et al. (2004). "Diabetes due to a progressive defect in beta-
cell mass in rats transgenic for human islet amyloid polypeptide (HIP Rat): a new
model for type 2 diabetes." Diabetes 53(6): 1509-16.
123
Butler, A. E., J. Janson, et al. (2003). "Beta-cell deficit and increased beta-cell
apoptosis in humans with type 2 diabetes." Diabetes 52(1): 102-10.
Butler, A. E., J. Janson, et al. (2003). "Increased beta-cell apoptosis prevents
adaptive increase in beta-cell mass in mouse model of type 2 diabetes: evidence
for role of islet amyloid formation rather than direct action of amyloid." Diabetes
52(9): 2304-14.
Butler, P. C., Chou, J., Carter, W. B., Wang, Y. N., Bu, B. H., Chang, D., Chang,
J. K., and Rizza, R. A. (1990). "Effects of meal ingestion on plasma amylin
concentration in NIDDM and nondiabetic humans." Diabetes 39: 752-756.
C Altenbach, D. A. G., H G Khorana, and W L Hubbell (1994). "A collision
gradient method to determine the immersion depth of nitroxides in lipid bilayers:
application to spin-labeled mutants of bacteriorhodopsin." Proc Natl Acad Sci
USA 91: 1667-1671.
Casas, S., R. Gomis, et al. (2007). "Impairment of the ubiquitin-proteasome
pathway is a downstream endoplasmic reticulum stress response induced by
extracellular human islet amyloid polypeptide and contributes to pancreatic beta-
cell apoptosis." Diabetes 56(9): 2284-94.
CD, L. (1995). "Expression of human beta-amyloid peptide in transgenic
Caenorhabditis elegans." Proc Natl Acad Sci U S A 92(20): 9368-72.
Chiang YW, B. P., Freed JH (2005). "The determination of pair distance
distributions by pulsed ESR using Tikhonov regularization." J Magn Reson 172:
279-295.
Claire S. GoldsburyaGarth J. S. CooperbKenneth N. GoldiecShirley A.
MüllerdEtuate L. Saafie, W. T. M. G., Martin P. MisureAndreas Engelf, Ueli
AebifJoerg Kistlerg (1997). "Polymorphic Fibrillar Assembly of Human Amylin." J
Struct Biol 119(1): 17-27.
Clark, A., G. J. Cooper, et al. (1987). "Islet amyloid formed from diabetes-
associated peptide may be pathogenic in type-2 diabetes." Lancet 2(8553): 231-
4.
Clark, A., E. J. de Koning, et al. (1995). "Pancreatic pathology in non-insulin
dependent diabetes (NIDDM)." Diabetes Res Clin Pract 28 Suppl: S39-47.
Clark, A., C. A. Wells, et al. (1988). "Islet amyloid, increased A-cells, reduced B-
cells and exocrine fibrosis: quantitative changes in the pancreas in type 2
diabetes." Diabetes Res 9(4): 151-9.
124
Cooper, G. J., A. C. Willis, et al. (1987). "Purification and characterization of a
peptide from amyloid-rich pancreases of type 2 diabetic patients." Proc Natl Acad
Sci U S A 84(23): 8628-32.
Crabb, J. W., Miyagi, M., Gu, X., Shadrach, K., West, K. A., Sakaguchi, H.,
Kamei, M., Hasan, A., Yan, L., Rayborn, M. E., Salomon, R. G., and Hollyfield, J.
G. (2002). "Drusen proteome analysis: an approach to the etiology of age-related
macular degeneration." Proc Natl Acad Sci U S A 99: 14682-7.
Daum, G. (1985). "Lipids of mitochondria." Biochim Biophys Acta 822: 1-42.
de Koning, E. J., J. W. Hoppener, et al. (1994). "Human islet amyloid polypeptide
accumulates at similar sites in islets of transgenic mice and humans." Diabetes
43(5): 640-4.
DeMager PP, P. B., Walter R, Harkany T, Härtignny W. (2002). "Pathological
peptide folding in Alzheimer's disease and other conformational disorders." Curr
Med Chem 9(19): 1763-80.
Der-Sarkissian A, J. C., Chen J, Langen R. (2003). "Structural organization of
alpha-synuclein fibrils studied by site-directed spin labeling." J Biol Chem
278(39): 37530-5.
Eanes, E. D., and Glenner, G.G. (1968). "X-ray diffraction studies on amyloid
filaments." J. Histochem Cytochem 16: 673-677.
Flaherty, M. J., West, S., Heimark, R. L., Fujikawa, K., and Tait, J. F. (1990).
"Placental anticoagulant protein-I: measurement in extracellular fluids and cells of
the hemostatic system." J Lab Clin Med 115: 174-81.
Garcia-Alloza, M., L. A. Borrelli, et al. (2007). "Curcumin labels amyloid pathology
in vivo, disrupts existing plaques, and partially restores distorted neurites in an
Alzheimer mouse model." J Neurochem 102(4): 1095-104.
Gavins FN, S. P., Chatterjee BE, Perretti M. (2005). "A twist in anti-inflammation:
annexin 1 acts via the lipoxin A4 receptor." Prostaglandins Leukot Essent Fatty
Acids 73(3-4): 211-9.
Gebre-Medhin, S., H. Mulder, et al. (1998). "Increased insulin secretion and
glucose tolerance in mice lacking islet amyloid polypeptide (amylin)." Biochem
Biophys Res Commun 250(2): 271-7.
Gerke V, C. C., Moss SE. (2005). "Annexins: linking Ca2+ signalling to
membrane dynamics." Nat Rev Mol Cell Biol 6(6): 449-61.
Giambanco, I., Sorci, G., Antonioli, S., Rambotti, M. G., Spreca, A., Bocchini, V.,
and Donato, (1993). "Immunocytochemical analyses of annexin V (CaBP33) in a
125
human-derived glioma cell line. Expression of annexin V depends on cellular
growth state." FEBS Lett 323: 45-50.
Glabe, C. G., and Kayed, R. (2006). "Common structure and toxic function of
amyloid oligomers implies a common mechanism of pathogenesis." Neurology
66: S74-8.
Goldsbury C, G. K., Pellaud J, Seelig J, Frey P, Müller SA, Kistler J, Cooper GJ,
Aebi U. (2000). "Amyloid fibril formation from full-length and fragments of amylin."
J Struct Biol 130(2-3): 352-62.
Goldsbury CS, C. G., Goldie KN, Müller SA, Saafi EL, Gruijters WT, Misur MP,
Engel A, Aebi U, Kistler J. (1997). "Polymorphic fibrillar assembly of human
amylin." J Struct Biol 119(1): 17-27.
Goldsbury, C. S., S. Wirtz, S. A. Mu¨ ller, S. Sunderji, P. Wicki, U. and a. P. F.
Aebi (2000). "Studies on the in vitro assembly of Ab1–40:
implications for the search for Ab fibril formation inhibitors." J Struct Biol 130:
217-231.
Groenning M, N. M., Flink JM, van de Weert M, Bukrinsky JT, Schluckebier G,
Frokjaer S (2007). "Binding mode of Thioflavin T in insulin amyloid fibrils." J
Struct Biol 159(3): 483-97.
Haataja, L., T. Gurlo, et al. (2008). "Islet amyloid in type 2 diabetes, and the toxic
oligomer hypothesis." Endocr Rev 29(3): 303-16.
Hafner-Bratkovic, I., J. Gaspersic, et al. (2008). "Curcumin binds to the alpha-
helical intermediate and to the amyloid form of prion protein - a new mechanism
for the inhibition of PrP(Sc) accumulation." J Neurochem 104(6): 1553-64.
Hamamichi S, R. R., Knight AL, Cao S, Caldwell KA, Caldwell GA. (2008).
"Hypothesis-based RNAi screening identifies neuroprotective genes in a
Parkinson's disease model." Proc Natl Acad Sci U S A 105(2): 728-33.
Higham, C. E., E. T. Jaikaran, et al. (2000). "Preparation of synthetic human islet
amyloid polypeptide (IAPP) in a stable conformation to enable study of
conversion to amyloid-like fibrils." FEBS Lett 470(1): 55-60.
Howard, C. F., Jr. (1986). "Longitudinal studies on the development of diabetes
in individual Macaca nigra." Diabetologia 29(5): 301-6.
Hubbell WL, G. A., Langen R, Lietzow MA. (1998). "Recent advances in site-
directed spin labeling of proteins." Curr Opin Struct Biol 8(5): 649-56.
Hull, R. L., G. T. Westermark, et al. (2004). "Islet amyloid: a critical entity in the
pathogenesis of type 2 diabetes." J Clin Endocrinol Metab 89(8): 3629-43.
126
Jack, E., M. Newsome, et al. (2006). "The organization of aromatic side groups in
an amyloid fibril probed by solid-state 2H and 19F NMR spectroscopy." J Am
Chem Soc 128(25): 8098-9.
Janson, J., R. H. Ashley, et al. (1999). "The mechanism of islet amyloid
polypeptide toxicity is membrane disruption by intermediate-sized toxic amyloid
particles." Diabetes 48(3): 491-8.
Janson, J., Ashley, R.H., Harrison, D., McIntyre, s., and Butler, P.C. (1999). "The
mechanism of islet amyloid polypeptide toxicity is membrane disruption by
intermediate-sized toxic amyloid particles." Diabetes 48: 491-8.
Janson, J., T. Laedtke, et al. (2004). "Increased risk of type 2 diabetes in
Alzheimer disease." Diabetes 53(2): 474-81.
Janson, J., W. C. Soeller, et al. (1996). "Spontaneous diabetes mellitus in
transgenic mice expressing human islet amyloid polypeptide." Proc Natl Acad Sci
U S A 93(14): 7283-8.
Jayasinghe, S. A. and R. Langen (2004). "Identifying structural features of fibrillar
islet amyloid polypeptide using site-directed spin labeling." J Biol Chem 279(46):
48420-5.
Jayasinghe, S. A. and R. Langen (2007). "Membrane interaction of islet amyloid
polypeptide." Biochim Biophys Acta 1768(8): 2002-9.
Jeschke G, e. a. (2006). "DeerAnalysis2006- a comprehensive Software
Package for Analyzing Pulsed ELDOR Data." Appl Magn Reson 30: 473-498.
Johnson, K. H., Stevens, J.B. (1973). "Light and electron microscopic studies of
islet amyloid in diabetic cats." Diabetes 22: 341-344.
Kahn, S. E., S. Andrikopoulos, et al. (1999). "Islet amyloid: a long-recognized but
underappreciated pathological feature of type 2 diabetes." Diabetes 48(2): 241-
53.
Kahn, S. E., D. A. D'Alessio, et al. (1990). "Evidence of cosecretion of islet
amyloid polypeptide and insulin by beta-cells." Diabetes 39(5): 634-8.
Kajava AV, A. U., Steven AC. (2005). "The parallel superpleated beta-structure
as a model for amyloid fibrils of human amylin." J Mol Biol 348(2): 247-52.
Kayed, R., J. Bernhagen, et al. (1999). "Conformational transitions of islet
amyloid polypeptide (IAPP) in amyloid formation in vitro." J Mol Biol 287(4): 781-
96.
127
Kayed, R., Head, E., Thompson, J. L., McIntire, T. M., Milton, S. C., Cotman, C.
W., and Glabe, C. G. (2003). "Common structure of soluble amyloid oligomers
implies common mechanism of pathogenesis." Science 300(486-9).
Khurana R, C. C., Ionescu-Zanetti C, Carter SA, Krishna V, Grover RK, Roy R,
Singh S (2005). "Mechanism of thioflavin T binding to amyloid fibrils." J Struct
Biol 151(3): 229-38.
Koopman G, R. C., Kuijten GA, Keehnen RM, Pals ST, van Oers MH. (1994).
"Annexin V for flow cytometric detection of phosphatidylserine expression on B
cells undergoing apoptosis." Blood 84(5): 1415-20.
Krebs, M. R. H., E. H. C. Bromley, et al. (2005). "The binding of thioflavin-T to
amyloid fibrils: localisation and implications." Journal of Structural Biology 149(1):
30-37.
Kudva, Y. C., C. Mueske, et al. (1998). "A novel assay in vitro of human islet
amyloid polypeptide amyloidogenesis and effects of insulin secretory vesicle
peptides on amyloid formation." Biochem J 331 ( Pt 3): 809-13.
Lakso M, V. S., Moilanen AM, Sirviö J, Thomas JH, Nass R, Blakely RD, Wong G
(2003). "Dopaminergic neuronal loss and motor deficits in Caenorhabditis
elegans overexpressing human alpha-synuclein." J Neirochem 86(1): 165-72.
Lansbury, P. T., and Lashuel, H. A. (2006). "A century-old debate on protein
aggregation and neurodegeneration enters the clinic." Nature 443(774-9).
Lashuel, H. A., Hartley, D., Petre, B. M., Walz, T., and Lansbury, P. T., Jr.
(2002). "Neurodegenerative disease: amyloid pores from pathogenic mutations."
Nature 418: 291.
Lee G, P. H., Arispe N. (2002). "Annexin 5 and apolipoprotein E2 protect against
Alzheimer's amyloid-beta-peptide cytotoxicity by competitive inhibition at a
common phosphatidylserine interaction site." Peptides 23(7): 1249-63.
Lee, G., H. B. Pollard, et al. (2002). "Annexin 5 and apolipoprotein E2 protect
against Alzheimer's amyloid-[beta]-peptide cytotoxicity by competitive inhibition at
a common phosphatidylserine interaction site." Peptides 23(7): 1249-1263.
Lim, G. P., T. Chu, et al. (2001). "The curry spice curcumin reduces oxidative
damage and amyloid pathology in an Alzheimer transgenic mouse." J Neurosci
21(21): 8370-7.
Link, C. (1995). "Expression of human beta-amyloid peptide in transgenic
Caenorhabditis elegans." Proc Natl Acad Sci U S A. 92: 9368-9372.
128
Lohr, M., B. Bergstrome, et al. (1992). "Human cytomegalovirus in the pancreas
of patients with type 2 diabetes: is there a relation to clinical features, mRNA and
protein expression of insulin, somatostatin, and MHC class II?" Virchows Arch A
Pathol Anat Histopathol 421(5): 371-8.
Lorenzo, A., B. Razzaboni, et al. (1994). "Pancreatic islet cell toxicity of amylin
associated with type-2 diabetes mellitus." Nature 368(6473): 756-60.
Luca, S., W. M. Yau, et al. (2007). "Peptide conformation and supramolecular
organization in amylin fibrils: constraints from solid-state NMR." Biochemistry
46(47): 13505-22.
Lührs T, R. C., Adrian M, Riek-Loher D, Bohrmann B, Döbeli H, Schubert D, Riek
R. (2005). "3D structure of Alzheimer's amyloid-beta(1-42) fibrils." Proc Natl Acad
Sci U S A 102(48): 17342-7.
Luo W, R. A., Chiosis G. (2008). "Heat shock protein 90: translation from cancer
to Alzheimer's disease treatment?" BMC Neurosci 9(Suppl 2:S7).
MacArthur, D. L., E. J. de Koning, et al. (1999). "Amyloid fibril formation is
progressive and correlates with beta-cell secretion in transgenic mouse isolated
islets." Diabetologia 42(10): 1219-27.
Malinchik, S. B., H. Inouye, K. E. Szumowski, and D. A. Kirschner (1998).
"Structural analysis of Alzheimer’s b (1–40) amyloid: protofilament
assembly of tubular fibrils." Biophys J 74(537-545).
Maloy, A. L., D. S. Longnecker, et al. (1981). "The relation of islet amyloid to the
clinical type of diabetes." Hum Pathol 12(10): 917-22.
Margittai, M. and R. Langen (2006). "Spin labeling analysis of amyloids and other
protein aggregates." Methods Enzymol 413: 122-39.
Massimo Stefani, C. M. D. (2003). "Protein aggregation and aggregate toxicity:
new insights into protein folding, misfolding diseases and biological evolution."
Journal of Molecular Medicine 81(11): 678-699.
Mazumder A, R. K., Weinstein J, Kohn KW, Pommier Y (1995). "Inhibition of
human immunodeficiency virus type-1 integrase by curcumin." Biochem
Pharmacol 49(8): 1165-70.
Mchaourab HS, L. M., Hideg K, Hubbell WL. (1996). "Motion of spin-labeled side
chains in T4 lysozyme. Correlation with protein structure and dynamics."
Biochemistry 35(24): 7692-704.
129
Meers P, M. T. (1993). "Calcium-dependent annexin V binding to phospholipids:
stoichiometry, specificity, and the role of negative charge." Biochemistry 32(43):
11711-21.
Mirzabekov, T. A., Lin, M.C., Kagan, B.L. (1996). "Pore Formation by the
cytotoxic islet amyloid peptide amylin." J. Biol. Chem 271: 1988-1992.
Morley, J. B., HR; Weyers, JJ; Morimoto, RI. (2002). "The threshold for
polyglutamine-expansion protein aggregation and cellular toxicity is dynamic and
influenced by aging in Caenorhabditis elegans." Proc Natl Acad Sci U S A. 99:
10417-10422.
Moss SE, M. R. (2004). "The annexins." Genome Biol 5(4): 219.
Naiki H, H. K., Hosokawa M, Takeda T. (1989). "Fluorometric determination of
amyloid fibrils in vitro using the fluorescent dye, thioflavin T1." Anal Biochem
177(2): 244-9.
Nollen, E. G., SM; van, HG; Kim, S; Chavez, A; Morimoto, RI, (2004). "Genome-
wide RNA interference screen identifies previously undescribed regulators of
polyglutamine aggregation." Proc Natl Acad Sci U S A. 101: 6403-6408.
O'Brien, T. D. (2002). "Pathogenesis of feline diabetes mellitus." Mol Cell
Endocrinol 197(1-2): 213-9.
O'Brien, T. D., P. C. Butler, et al. (1995). "Human islet amyloid polypeptide
expression in COS-1 cells. A model of intracellular amyloidogenesis." Am J
Pathol 147(3): 609-16.
Ono, K. and M. Yamada (2006). "Antioxidant compounds have potent anti-
fibrillogenic and fibril-destabilizing effects for alpha-synuclein fibrils in vitro." J
Neurochem 97(1): 105-15.
Padrick, S. B., and Miranker, A. D (2002). "Islet amyloid: phase partitioning and
secondary nucleation are central to the mechanism of fibrillogenesis."
Biochemistry 41: 4694-703.
Padrick, S. B. and A. D. Miranker (2001). "Islet amyloid polypeptide: identification
of long-range contacts and local order on the fibrillogenesis pathway." J Mol Biol
308(4): 783-94.
Padrick, S. B. and A. D. Miranker (2002). "Islet amyloid: phase partitioning and
secondary nucleation are central to the mechanism of fibrillogenesis."
Biochemistry 41(14): 4694-703.
Pandey, N., J. Strider, et al. (2008). "Curcumin inhibits aggregation of alpha-
synuclein." Acta Neuropathol 115(4): 479-89.
130
Pannier M, V. S., Godt A, Jeschke G, Spiess HW (2000). "Dead-time free
measurement of dipole-dipole interactions between electron spins." J Magn
Reson 142: 331-340.
Pearlman D, e. a. (1995). "AMBER, a package of computer programs for
applying molecular mechanics, normal mode analysis, molecular dynamics and
free energy calculations to simulate the structural and energetic properties of
molecules." Comp Phys Comm 91: 1-41.
Puchtler, H., Sweat, F. and Levine, M. (1961). "On the binding of Congo Red by
Amyloid." Journal of Histochemistry and Cytochemistry 10: 355-364.
Rescher, U., and Gerke, V. (2004). "Annexins--unique membrane binding
proteins with diverse functions." J Cell Sci 117: 2631-9.
Rescher U, G. V. (2004). "Annexins--unique membrane binding proteins with
diverse functions." J Cell Sci 117(Pt 13): 2631-9.
Reutelingsperger, C. P. (2001). "Annexins: key regulators of haemostasis,
thrombosis, and apoptosis." Thromb Haemost 86(413-9).
Ritzel, R. A., Butler, A. E., Rizza, R. A., Veldhuis, J. D., and Butler, P. C. (2006).
"Relationship between beta-cell mass and fasting blood glucose concentration in
humans." Diabetes Care 29: 717-8.
Ritzel, R. A., J. J. Meier, et al. (2007). "Human islet amyloid polypeptide
oligomers disrupt cell coupling, induce apoptosis, and impair insulin secretion in
isolated human islets." Diabetes 56(1): 65-71.
Rocken, C., R. P. Linke, et al. (1992). "Immunohistology of islet amyloid
polypeptide in diabetes mellitus: semi-quantitative studies in a post-mortem
series." Virchows Arch A Pathol Anat Histopathol 421(4): 339-44.
Romisch, J., Schuler, E., Bastian, B., Burger, T., Dunkel, F. G., Schwinn, A.,
Hartmann, A. A., and Paques, E. P. (1992). "Annexins I to VI: quantitative
determination in different human cell types and in plasma after myocardial
infarction." Blood Coagul Fibrinolysis 3: 11-7.
Ross CA, P. M. (2004). "Protein aggregation and neurodegenerative disease."
Nat Med 10: Supple S10-7.
Saafi, E. L., Konarkowska, B., Zhang, S., Kistler, J., and Cooper, G. J. (2001).
"Ultrastructural evidence that apoptosis is the mechanism by which human
amylin evokes death in RINm5F pancreatic islet beta-cells." Cell Biol Int 25: 339-
50.
131
Saibil, S. C.-K. a. H. R. (2007). "Three-Dimensional Structural Analysis of
Amyloid Fibrils by Electron Microscopy." Protein Reviews 4: 303-313.
Sawaya, M. R., S. Sambashivan, et al. (2007). "Atomic structures of amyloid
cross-beta spines reveal varied steric zippers." Nature 447(7143): 453-7.
Schmidt, V., Giacomelli,C, Brisson, A.R., Borsali, R. (2008). "Towards an easy
access to Annexin-A5 protein binding block copolymer micelles." Materials
science & engineering C. Biomimetic and supramolecular systems 28(4): 479-
488.
Serpell LC, S. J. (2000). "Direct visualisation of the beta-sheet structure of
synthetic Alzheimer's amyloid." J Mol Biol 299(1): 225-31.
Sharma MR, K. L., Ownbey RT, Tuszynski GP, Sharma MC. (2006).
"Angiogenesis-associated protein annexin II in breast cancer: selective
expression in invasive breast cancer and contribution to tumor invasion and
progression." Exp Mol Pathol 81(2): 146-56.
Shirahama T, C. A. (1967). "High-resolution electron microscopic analysis of the
amyloid fibril." The Journal of Cell Biology 33: 679-708.
Soto, C. (2003). "Unfolding the role of protein misfolding in neurodegenerative
diseases." Nature Reviews Neuroscience 4: 49-60.
Sumner Makin, O. and L. C. Serpell (2004). "Structural characterisation of islet
amyloid polypeptide fibrils." J Mol Biol 335(5): 1279-88.
Sunde, M., Blake C.C.F. (1998). "From the globular to the fibrous state: Protein
structure and structural conversion in teh amyloid formation." Q. Rev. Biophys.
31: 1-39.
Swairjo, M. A., Concha, N. O., Kaetzel, M. A., Dedman, J. R., and Seaton, B. A
(1995). "Ca(2+)-bridging mechanism and phospholipid head group recognition in
the membrane-binding protein annexin V." Nat Struc Biol 2(968-74).
Tomiyama, T., H. Kaneko, et al. (1997). "Rifampicin inhibits the toxicity of pre-
aggregated amyloid peptides by binding to peptide fibrils and preventing amyloid-
cell interaction." Biochem J 322 ( Pt 3): 859-65.
Torok, M., Milton, S., Kayed, R., Wu, P., McIntire, T., Glabe, C. G., and Langen,
R. (2002). "Structural and dynamic features of Alzheimer's Abeta peptide in
amyloid fibrils studied by site-directed spin labeling." J Biol Chem 277: 40810-5.
Török M, M. S., Kayed R, Wu P, McIntire T, Glabe CG, Langen R. (2002).
"Structural and dynamic features of Alzheimer's Abeta peptide in amyloid fibrils
studied by site-directed spin labeling." J Biol Chem 277(43): 40810-5.
132
Tycko, R. (2006). "Molecular Structure of amyloid fibrils: Inishgts from solid state
NMR,." Q.Rev. Biophys. 39: 1-55.
van Ham TJ, T. K., Breitling R, Hofstra RM, Plasterk RH, Nollen EA. (2008). "C.
elegans model identifies genetic modifiers of alpha-synuclein inclusion formation
during aging." PLoS Genet. 4(3).
Verchere, C. B., D. A. D'Alessio, et al. (1996). "Islet amyloid formation associated
with hyperglycemia in transgenic mice with pancreatic beta cell expression of
human islet amyloid polypeptide." Proc Natl Acad Sci U S A 93(8): 3492-6.
Westermark, P., C. Wernstedt, et al. (1987). "Amyloid fibrils in human insulinoma
and islets of Langerhans of the diabetic cat are derived from a neuropeptide-like
protein also present in normal islet cells." Proc Natl Acad Sci U S A 84(11): 3881-
5.
Westermark, P., E. Wilander, et al. (1987). "Islet amyloid polypeptide-like
immunoreactivity in the islet B cells of type 2 (non-insulin-dependent) diabetic
and non-diabetic individuals." Diabetologia 30(11): 887-92.
Wilhelmus MM, O.-H. I., Wesseling P, de Waal RM, Boelens WC, Verbeek MM
(2006). "Specific association of small heat shock proteins with the pathological
hallmarks of Alzheimer's disease brains." Neuropathol Appl Neurobiol 32(2): 119-
30.
Wiltzius, J. J., S. A. Sievers, et al. (2008). "Atomic structure of the cross-{beta}
spine of Islet Amyloid Polypeptide (Amylin)." Protein Sci.
Wimley, W. C., Selsted, M. E., and White, S. H. (1994). "Interactions between
human defensins and lipid bilayers: evidence for formation of multimeric pores."
Protein Sci 3: 1362-73.
Wisniewski T, S. M. (2008). "Preventing beta-amyloid fibrillization and deposition:
beta-sheet breakers and pathological chaperone inhibitors." BMC Neurosci
9(Suppl 2:S5).
Wu Y, L. Y. (2005). "Transgenic C. elegans as a model in Alzheimer's research."
Curr Alzheimer Res 2(1): 37-45.
Yang, F., G. P. Lim, et al. (2005). "Curcumin inhibits formation of amyloid beta
oligomers and fibrils, binds plaques, and reduces amyloid in vivo." J Biol Chem
280(7): 5892-901.
Yong W, L. A., Kirkitadze MD, Teplow DB, Chen SH, Benedek GB. (2002).
"Structure determination of micelle-like intermediates in amyloid beta -protein
133
fibril assembly by using small angle neutron scattering." Proc Natl Acad Sci U S
A 99(1): 15-4.
Young, A. A., Gedulin, B.B., Vine, W., Percy A., Rink, T.J. (1995). "Gastric
emptying is accelerated in diabetic BB rats and is slowed by subcutaneous
injections of amylin." Diabetologia 38(6): 642-8.
Zhang, S., J. Liu, et al. (1999). "Induction of apoptosis by human amylin in
RINm5F islet beta-cells is associated with enhanced expression of p53 and
p21WAF1/CIP1." FEBS Lett 455(3): 315-20.

Asset Metadata

Creator Bedrood, Sahar (author)

Core Title Structural features and modifiers of islet amyloid polypeptide: implications for type II diabetes mellitus

Contributor Electronically uploaded by the author (provenance)

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Biochemistry and Molecular Biology

Degree Conferral Date 2009-08

Publication Date 06/23/2009

Defense Date 05/07/2009

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

Tag amyloid,amyloid structure,annexin V,curcumin,IAPP,islet amyloid polypeptide,OAI-PMH Harvest,protein misfolding,type II diabetes mellitus

Language English

Advisor Chen, Jeannie (committee member), Chow, Robert HP. (committee member), Haworth, Ian S. (committee member)

Creator Email bedrood@usc.edu,carpemane@graffiti.net

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m2315

Unique identifier UC1492780

Identifier etd-Bedrood-2945 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-243190 (legacy record id),usctheses-m2315 (legacy record id)

Legacy Identifier etd-Bedrood-2945.pdf

Dmrecord 243190

Document Type Thesis

Rights Bedrood, Sahar

Type texts

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

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email uscdl@usc.edu

Abstract (if available)

Abstract Protein misfolding is a common motif in a number of human diseases, including Alzheimer disease, Parkinson disease and type II diabetes mellitus (TTDM). In TTDM, over 90% of patients are found to have pancreatic amyloid deposits upon autopsy. These deposits are primarily composed of a 37-residue human islet amyloid polypeptide (hIAPP). Evidence suggests an association between these amyloid plaques and pancreatic beta-cell dysfunction. Elucidating the structure of these deposits and the effects modifiers have on the misfolding pathway can help further the understanding of its toxicity and could be of use in the design of drug inhibitors for amyloid diseases. I used site-directed spin-labeling and electron paramagnetic resonance (EPR) spectroscopy to analyze spin-labeled derivatives of hIAPP to determine structural features of the peptide in its fibrillar form. Using continuous-wave EPR and four-pulse DEER coupled with computational modeling, I determined a detailed structural model of hIAPP fibrils. The N-terminal and C-terminal regions are less ordered and more mobile and there is a turn region between residues 19-30 located between two β-strands (13-18 and 31-36). My findings also show that the β1 and β2 strands from one molecule of hIAPP are staggered in relation to one another. The staggered peptide stacks directly on another staggered peptide, forming a protofilament with a twist that I aptly call a β-spiral motif. In addition to these structural findings, I have studied the effects of modifiers, such as curcumin and annexin proteins, to the misfolding of amyloid fibrils. Using EPR, ThT and EM, I found that these molecules or proteins seem to alter hIAPP, Aβ and α-synuclein fibril formation and fibril morphology. Additionally, studies of these modifiers in tissue culture and animal models showed reduced toxicity and protein aggregation.