structural characterization and membrane interactions of ......ii structural characterization and...
TRANSCRIPT
Structural Characterization and Membrane Interactions of the Amyloid Peptide PrP(106-126)
by
Patrick Walsh
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy
Department of Biochemistry University of Toronto
© Copyright by Patrick Walsh 2013
ii
Structural Characterization and Membrane Interactions of the
Amyloid Peptide PrP(106-126)
Patrick Walsh
Doctor of Philosophy
Department of Biochemistry
University of Toronto
2013
Abstract
The formation of amyloid fibrils is a key characteristic of many neurodegenerative diseases
including Alzheimer’s and Parkinson’s diseases. Similarly prion diseases, those associated with
the prion protein, are neurodegenerative disorders with characteristic protein aggregates
accumulating in the brain of affected individuals. While fibrillar deposits of these disorders have
long been associated with end-stage disease pathology, it is currently hypothesized that protein
oligomers are the cytotoxic structural form of these systems. Residues 106-126 of the human
prion protein have been found to form both amyloid fibrils, as well as toxic amyloid oligomers
and thus provide a suitable model system. This thesis aims to describe the structures of the
amyloid fibrils and oligomers formed by PrP(106-126), how they are interrelated as well as their
interaction with model membranes and cytotoxicity.
Amyloid fibrils of PrP(106-126) contain long, unbranched filaments that contain β-sheet
secondary structure and bind the amyloid-indicating dye, thioflavin-T. These fibrils are
comprised of parallel β-sheets, stacked in an antiparallel fashion.
iii
The non-fibrillar amyloid oligomers are large, spherical structures that contain β-sheets but do
not bind thioflavin-T. It was determined that these oligomers contain parallel β-sheets as well as
the same intersheet packing as fibrils of PrP(106-126).
Finally, the interaction of PrP(106-126) with lipid bilayers and cells was examined. Oligomers of
PrP(106-126) were shown to affect model membranes; with anionic lipids losing integrity and
cholesterol-containing lipid mixtures losing domain structure upon peptide addition.
Additionally, amyloid oligomers of PrP(106-126) cause cell death across a number of cell lines
as well as rat cerebellar slices.
Overall, these results indicate that the conversion of oligomers to fibrils may be facilitated due to
structural similarities between the two. Additionally, the toxicity of PrP(106-126) oligomers may
be attributed to a loss of cholesterol domain structure causing subsequent cell death.
iv
Acknowledgments
I would like to start by thanking my supervisor, Dr. Simon Sharpe, for his constant support and
mentoring over the past 5 years. His patience and constant willingness to invest time in me has
not only made me a better scientist but a better person as a whole. From my starting time in the
lab, Dr. Sharpe has given me the time and skills to succeed, as well as an amazingly interesting
and well thought-out project.
I am very grateful to my graduate committee members, Dr. John Rubinstein and Dr. Avi
Chakrabartty, for their thoughtful input into my project. Also, I wish to thank you both for your
support outside of the committee setting both in matters of science and otherwise.
A great many thanks to the members of the Sharpe lab, past and present. Working with you has
been a great pleasure – I cannot imagine having to do my PhD without you. Your constant
support and friendship has meant a great deal to me and I wish everyone all the best as we go on
with our various careers. I want to especially thank Dave for always being there to lend a helping
hand and a timely joke or two.
A special thanks to my friends, the ones that I knew before I started and the ones that I met along
the way – you have all impacted me for the better and I thank you.
Thank you to my wonderful family – without you, I would be lost. To my mother, thank you for
showing me what it is to work hard by example. Thank you to my sisters for always being just a
phone call away to listen, support and encourage. I am also very grateful to my Grandparents for
their love and support. Thank you to my Aunts and Uncles for always being there to encourage
me. I wish to express special thanks to my Uncle Greg, who introduced to me science from a
young age.
I am most grateful to my wife, Mary Ann, who has stood by me with the utmost love and
dedication. Thank you for your patience and support, especially over the last 5 years.
v
Table of Contents
Acknowledgments .......................................................................................................................... iv
Table of Contents ............................................................................................................................ v
List of Tables ................................................................................................................................. ix
List of Figures ................................................................................................................................. x
List of Equations .......................................................................................................................... xiv
List of Abbreviations .................................................................................................................... xv
1 Introduction ................................................................................................................................ 1
1.1 Protein Misfolding and Amyloids ....................................................................................... 1
1.2 Amyloid Disease ................................................................................................................. 3
1.3 Structural Studies of Amyloid Fibrils ................................................................................. 4
1.4 Prion Protein Structures ...................................................................................................... 8
1.5 Non-fibrillar Oligomers on the Misfolding Pathway of Amyloid Proteins ...................... 10
1.6 Structural Studies of Non-Fibrillar Amyloid Oligomers .................................................. 11
1.7 Non-fibrillar amyloid oligomers as the cytotoxic agents in amyloid disease ................... 16
1.8 PrP(106-126) Peptide as a Model for Amyloid Diseases ................................................. 18
1.9 Biological Applications of Solid State Nuclear Magnetic Resonance .............................. 21
1.9.1 Magic Angle Spinning (MAS) .............................................................................. 21
1.9.2 Chemical Shift ...................................................................................................... 23
1.9.3 Dipolar Coupling .................................................................................................. 24
1.9.4 Solid State NMR in Lipids .................................................................................... 25
2 Core Structure of Amyloid Fibrils Formed by Residues 106-126 of the Human Prion
Protein ...................................................................................................................................... 27
2.1 Abstract ............................................................................................................................. 28
2.2 Introduction ....................................................................................................................... 29
2.3 Materials and Methods ...................................................................................................... 31
vi
2.3.1 PrP(106-126) Fibril Formation ............................................................................. 31
2.3.2 Circular Dichroism Spectroscopy ......................................................................... 32
2.3.3 Thioflavin-T Fluorescence .................................................................................... 32
2.3.4 Transmission Electron Microscopy ...................................................................... 32
2.3.5 Atomic Force Microscopy .................................................................................... 32
2.3.6 Solid State Nuclear Magnetic Resonance ............................................................. 33
2.3.7 NMR Data Analysis .............................................................................................. 34
2.3.8 PrP(106-126) Fibril Modeling .............................................................................. 35
2.4 Results ............................................................................................................................... 37
2.4.1 PrP(106-126) Forms Amyloid Fibrils with Characteristics of a Cross-β
Structure ................................................................................................................ 37
2.4.2 13
C and 15
N Chemical Shifts Reveal an Extended β-Sheet, Spanning Residues
113-123 ................................................................................................................. 39
2.4.3 Amyloid Fibrils of PrP(106-126) are Composed of In-Register Parallel β-
Sheets .................................................................................................................... 44
2.4.4 Quaternary Structure of PrP(106-126) Fibrils from 13
C Spin Diffusion and
Rotational Resonance Experiments ...................................................................... 46
2.4.5 Structural Model of PrP(106-126) Fibrils Based on Solid State NMR
Measurements ....................................................................................................... 49
2.5 Discussion ......................................................................................................................... 51
3 Morphology and Secondary Structure of Stable β-Oligomers Formed by Amyloid Peptide
PrP(106-126) ............................................................................................................................ 54
3.1 Abstract ............................................................................................................................. 55
3.2 Introduction ....................................................................................................................... 56
3.3 Materials and Methods ...................................................................................................... 58
3.3.1 PrP(106-126) Oligomer Formation ....................................................................... 58
3.3.2 Circular Dichroism Spectroscopy ......................................................................... 58
3.3.3 Thioflavin T Fluorescence .................................................................................... 58
3.3.4 Dynamic Light Scattering ..................................................................................... 59
vii
3.3.5 Transmission Electron Microscopy ...................................................................... 59
3.3.6 Atomic Force Microscopy .................................................................................... 59
3.3.7 Liposome Dye-Release Assay .............................................................................. 60
3.3.8 Solid State Nuclear Magnetic Resonance ............................................................. 61
3.4 Results ............................................................................................................................... 62
3.4.1 PrP(106-126) Forms Stable β-sheet Non-fibrillar Oligomers ............................... 62
3.4.2 PrP(106-126) Non-fibrillar Oligomers Form as a Discrete Size .......................... 65
3.4.3 PrP(106-126) Non-fibrillar Oligomers Disrupt Model Membranes ..................... 67
3.5 Discussion ......................................................................................................................... 69
4 Structural Properties and Dynamic Behaviour of Non-Fibrillar Oligomers Formed by
PrP(106-126) ............................................................................................................................ 71
4.1 Abstract ............................................................................................................................. 72
4.2 Introduction ....................................................................................................................... 73
4.3 Materials and Methods ...................................................................................................... 76
4.3.1 Solid State NMR ................................................................................................... 76
4.3.2 Solution NMR ....................................................................................................... 76
4.4 Results ............................................................................................................................... 78
4.4.1 PrP(106-126) Oligomers Contain In-register Parallel β-sheets ............................ 78
4.4.2 PrP(106-126) Oligomers Contain Quaternary Contacts Between β-sheets
Similar to those in PrP(106-126) Amyloid Fibrils ................................................ 82
4.4.3 Identification of Structured Monomeric PrP(106-126) in Fast Exchange with
Non-fibrillar Oligomers from 1H-
1H and
1H-
13C Solution NMR Spectra ............ 84
4.4.4 MAS NMR Paramagnetic Relaxation Enhancement (PRE) of PrP(106-126)
Fibrils and Oligomers ........................................................................................... 90
4.4.5 Proposed Structural Model for Non-fibrillar Oligomers of PrP(106-126) ........... 92
4.5 Discussion ......................................................................................................................... 95
5 Membrane Interactions of PrP(106-126) Oligomers .............................................................. 100
5.1 Abstract ........................................................................................................................... 101
viii
5.2 Introduction ..................................................................................................................... 102
5.3 Materials and Methods .................................................................................................... 104
5.3.1 Preparation of PrP(106-126) Non-fibrillar Oligomers ........................................ 104
5.3.2 Formation of Large Unilamallar liposomes ........................................................ 104
5.3.3 Transmission Electron Microscopy .................................................................... 104
5.3.4 Atomic Force Microscopy .................................................................................. 104
5.3.5 AFM-TIRF .......................................................................................................... 105
5.3.6 Solid State NMR of Liposomes .......................................................................... 106
5.3.7 Brain Slice and Cell Culture ............................................................................... 106
5.4 Results ............................................................................................................................. 108
5.4.1 PrP(106-126) Oligomers Disrupt Anionic Lipid Bilayers .................................. 108
5.4.2 PrP(106-126) Causes Loss of Lipid Domain Order in Cholesterol-Containing
Bilayers ............................................................................................................... 110
5.4.3 PrP(106-126) Oligomers are Cytotoxic to Cultured Cells .................................. 113
5.4.4 Oligomers of PrP(106-126) are Toxic to Rat Cellebellar Brain Slices ............... 116
5.5 Discussion ....................................................................................................................... 117
6 Summary and Future Directions ............................................................................................ 119
6.1 Summary ......................................................................................................................... 119
6.2 Future Directions ............................................................................................................ 121
6.2.1 Continuing Studies on PrP(106-126) Oligomer Interactions with Lipid
Bilayers ............................................................................................................... 121
6.2.2 PrP(106-126) Structures Formed in the Presence of the Bilayer ........................ 123
6.2.3 Additional Toxicity Studies ................................................................................ 124
6.3 Final Conclusions ............................................................................................................ 126
References ................................................................................................................................... 127
Appendices .................................................................................................................................. 147
ix
List of Tables
Table 2-1 - Amino acid sequence and isotope labelling schemes for PrP(106-126) peptides. ..... 31
Table A-1 - 13
C and 15
N chemical shift assignments for PrP(106-126) fibrils. .......................... 147
Table A-2 - Backbone and torsion angles predicted for the sheet-forming region of PrP(106-
126) using TALOS analysis of 13
C and 15
N chemical shifts. ...................................................... 148
Table A-3 – 1H Chemical Shifts for PrP(106-126) Oligomers ................................................... 152
Table A-4 – 13
C chemical shifts for PrP(106-126) Oligomers ................................................... 154
x
List of Figures
Figure 1-1 – Possible Structures Formed Along the Folding/Misfolding Pathway ........................ 2
Figure 1-2 – Possible arrangements of β-strands in an amyloid fibril ............................................ 5
Figure 1-3 – Structure of amyloid β(1-40) fibrils as determined by solid state NMR ................... 7
Figure 1-4 – Solution NMR Structure of the Human Prion Protein (hPrP) .................................... 8
Figure 1-5 – Structures formed by aggregated prion proteins from human and yeast. ................ 10
Figure 1-6 – Toxicity and common structural architecture of amyloid oligomers ....................... 12
Figure 1-7 – Structures of non-fibrillar amyloid oligomers .......................................................... 15
Figure 1-8 – Various Methods of Membrane Disruption by Amyloids ........................................ 17
Figure 1-9 - The relationship between amyloid fibril length and toxicity. ................................... 18
Figure 1-10 – PrP(106-126) fibril model of peptide stacking based on mutagenesis ................... 19
Figure 1-11 – Schematic Representation of Magic Angle Spinning ............................................ 23
Figure 2-1 - Ultrastructural characterization of PrP(106-126) fibrils. .......................................... 38
Figure 2-2 - MAS NMR spectra of PrP(106-126)GAVL
fibrils. ..................................................... 40
Figure 2-3 - Secondary 13
C chemical shifts and line widths measured for PrP(106-126) fibrils. 41
Figure 2-4 - Differences in 13
C NMR line widths between dry and hydrated fibrils of PrP(106-
126). .............................................................................................................................................. 43
Figure 2-5 - PITHIRDS recoupling curves for PrP(106-126) fibrils. ........................................... 45
Figure 2-6 - Long-range 13
C-13
C internuclear contacts observed in 2D 13
C-13
C NMR spectra of
PrP(106-126)GAVL
fibrils. .............................................................................................................. 46
xi
Figure 2-7 - Long-range 13
C-13
C internuclear contacts observed in 2D 13
C-13
C NMR spectra of
PrP(106-126)AVG
fibrils. ............................................................................................................... 47
Figure 2-8 - Experimental 13
C rotational resonance data for PrP(106-126)GAVL
fibrils and
simulated polarization transfer curves. ........................................................................................ 48
Figure 2-9 - Structural models of PrP(106-126) fibrils. ............................................................... 50
Figure 3-1 – Transmission electron microscopy of PrP(106-126) oligomers ............................... 62
Figure 3-2 - Negative stain TEM images of oligomeric PrP(106-126) sample with different
histories do not show evidence of fibril formation or changes in morphology. ........................... 63
Figure 3-3 - AFM of PrP(106-126) oligomers. ............................................................................. 64
Figure 3-4 – Dynamic Light Scattering of PrP(106-126) Non-fibrillar Oligomers ...................... 65
Figure 3-5 – Spectroscopic analysis of PrP(106-126) Oligomers ................................................ 66
Figure 3-6 - 13
C NMR linewidths for PrP(106-126) oligomers. ................................................... 67
Figure 3-7 – Release of the fluorescent dye calcein from 3:1 POPC:POPG liposomes induced by
PrP(106-126) oligomers. ............................................................................................................... 68
Figure 4-1 - Comparison of 13
C secondary chemical shifts and NMR linewidths of hydrated
versus lyophilized PrP(106-126) oligomers. ................................................................................. 79
Figure 4-2 - 13
C-13
C chemical shift correlation spectra of dry versus hydrated PrP(106-126)AVG2
oligomers. ...................................................................................................................................... 80
Figure 4-3 - PITHIRDS recoupling curves for non-fibrillar PrP(106-126) oligomers. ................ 81
Figure 4-4 - Long-range 13
C-13
C internuclear contacts observed in 2D 13
C-13
C NMR spectra of
PrP(106-126)GAVL
oligomers. ....................................................................................................... 83
Figure 4-5 - Long-range 13
C-13
C internuclear contacts are maintained in hydrated PrP(106-
126)GAVL
oligomers. ...................................................................................................................... 84
xii
Figure 4-6 Translational diffusion of PrP(106-126) non-fibrillar oligomers. ............................... 85
Figure 4-7 – 1H-
1H TOCSY and 1H-13C HSQC NMR spectra of PrP(106-126) monomers in
equilibrium with non-fibrillar oligomers ...................................................................................... 86
Figure 4-8 - Sequential and intermolecular NOEs observed in a solution containing non-fibrillar
oligomers of PrP(106-126). .......................................................................................................... 87
Figure 4-9 - φ and ψ backbone torsion angles predicted for PrP(106-126) oligomers and
structured monomers. .................................................................................................................... 89
Figure 4-10 - Mn2+
paramagnetic relaxation enhancement effects in 13
C cross-polarization
spectra of PrP(106-126)AVG2
fibrils and oligomers. ...................................................................... 91
Figure 4-11 - Structural models for non-fibrillar oligomers formed by PrP(106-126). ................ 93
Figure 4-12 - Schematic representation of intramolecular and intermolecular restraints used in
structure calculations and model building. ................................................................................... 94
Figure 4-13 - 13
C spin relaxation times obtained under MAS for non-fibrillar oligomers and
amyloid fibrils formed by PrP(106-126). ...................................................................................... 97
Figure 5-1 – AFM of 3:1 POPC:POPG Supported Bilayers ....................................................... 108
Figure 5-2 – Negative Stain TEM of Large Unilamellar Vesicles ............................................. 109
Figure 5-3 – 31
P Static NMR spectra of anionic large unilamellar vesicles ............................... 110
Figure 5-4 - AFM of 1:1:1 DSPC:DOPC:Cholesterol Supported Bilayers ................................ 111
Figure 5-5 – Static 31
P Spectra of cholesterol-containing LUVs ................................................ 112
Figure 5-6 – Polarized TIRF and AFM Images of 1:1:1 DOPC:DSPC:Cholesterol .................. 113
Figure 5-7 – Toxilight cell-death assay ...................................................................................... 114
Figure 5-8 – MTS Reduction Assay ........................................................................................... 115
Figure 5-9 – Exposure of rat cerebellar slices to PrP(106-126) oligomers ................................. 116
xiii
Figure A-1 - 13
C and 15
N chemical shift correlation spectra for PrP(106-126)AVG
fibrils. ......... 149
Figure A-2 - 13
C and 15
N chemical shift correlation spectra for PrP(106-126)AVG2
fibrils ......... 150
Figure A-3 – Thioflavin-T fluorescence of PrP(106-126) Oligomers Under Various Conditions
..................................................................................................................................................... 151
xiv
List of Equations
Equation 1-1 – Anisotropic Chemical Shift ………….………………………………………….22
Equation 1-2 – Dipolar Coupling Constant……………………………………………………...24
Equation 2-1 – PITHIRDS Weighted Sum of Squared Residuals……………………………….35
xv
List of Abbreviations
Aβ – Alzheimer β peptide
ACN - Acetonitrile
AFM – Atomic Force Microscopy
Ala – Alanine
BSE – Bovine Spongiform Encephalopathy
CD – Circular Dichroism
CJD – Creutzfeldt-Jakob Disease
CP – Cross Polarization
CSA – Chemical Shift Anisotropy
CWD – Chronic Wasting Disease
DNA – Deoxyribonucleic Acid
EDTA – Ethylenediaminetetraacetic Acid
FID – Free Induction Decay
FMOC - 9-Fluorenylmethoxycarbonyl chloride
FTIR – Fourier Transform Infrared Spectroscopy
Gly – Glycine
GSS- Gerstmann-Sträussler-Scheinker
HFIP – 1,1,1,3,3,3-Hexafluoroisopropanol
HPLC – High Performance Liquid Chromatography
xvi
List of Abbreviations Continued…
HSQC – Heteronuclear Single Quantum Correlation
IAPP – Islet Amyloid Polypeptide
LUV – Large Unilamellar Vesicle
Lys – Lysine
MAS – Magic Angle Spinning
MD – Molecular Dynamics
NMR – Nuclear Magnetic Resonance
NOESY – Nuclear Overhauser Effect Spectroscopy
POPC - 1-Palmitoyl-2-Oleoyl-sn-Glycero-3-Phosphocholine
POPG - 1-Palmitoyl-2-Oleoyl-sn-Glycero-3-Phosphoglycerol
PrP – Prion Protein
PrPC – Cellular Prion Protein
PrPSc
– Scrapie Prion Protein
RAD – Radiofrequency Assisted Dipolar Recoupling
RMSD – Root Mean Square Deviation
RNA – Ribonucleic Acid
RR – Rotational Resonance
SDS – Sodium Dodecyl Sulphate
TEM – Transmission Electron Microscopy
xvii
List of Abbreviations Continued…
ThT – Thioflavin-T
TFA – Trifluoroacetic Acid
TMS – Trimethyl Silane
TOCSY – Total Correlation Spectroscopy
TPPM – Two-pulse Phase Modulation
Val - Valine
1
1 Introduction
Selections from this chapter were originally published in the book Advanced Understanding of
Neurodegenerative Disease. Patrick Walsh and Simon Sharpe. Structure-toxicity relationships of
amyloid peptide oligomers. Advanced Understanding of Neurodegenerative Disease. pp. 89-114
copyright Intech 2011.
1
1.1 Protein Misfolding and Amyloids
Protein folding is a complex process through which proteins adopt their native, functional
structure. The formation of properly folded proteins is dependent on many factors, not the least
of which is the primary amino acid sequence of a given protein. In normal conditions, protein
folding proceeds with little to no errors; any mistakes are corrected by chaperones or misfolded
proteins are discarded. In many disease states including Alzheimer’s and Parkinson’s diseases
proteins misfold and form aberrant, toxic protein structures. In the cases of amyloid diseases,
there is a characteristic formation of amyloid fibrils – long unbranched filaments deposited in
affected tissues. These fibrillar aggregates bind dyes such as congo red and thioflavin-T giving
rise to birefringence or fluorescence upon interaction with these dyes respectively. The
distinctive feature of amyloid fibrils is a very well ordered arrangement of β-sheets in which the
individual polypeptide chains run perpendicular to the long axis of the fibril – a structure known
as cross-β. Along the misfolding pathway, there are a number of intermediates that can form as
well as a number of ways that a given protein can misfold, as summarized in Figure 1-1. In one
case, unfolded polypeptide chains can be broken down by proteolytic cleavage into smaller
chains or aggregate in a disordered state. In another example, intermediates can assemble into
prefibrillar aggregates. The formation of pre-fibrillar aggregates is associated with the formation
of amyloid fibrils; these prefibrillar structures can include amyloid oligomers, protofibrils,
protofilaments and annular oligomers (Lashuel et al. 2002; Dobson 2003). These intermediates
can go on to form amyloid fibrils and plaques which can deposit inside cells, however, most of
these aggregates will be deposited in the extracellular space as is the case in Parkinson’s and
Alzheimer’s diseases (Dobson 2003).
2
Figure 1-1 – Possible Structures Formed Along the Folding/Misfolding Pathway
A summary of the possible structures formed during protein folding or misfolding. Each structure along the folding
pathway has the ability to form a higher-order structure, including natively folded protein. Unfolded and folding
intermediates can form pre-fibrillar aggregates or amyloid fibrils while native proteins can assemble into oligomers
or fibers comprised of native- protein monomers. Reprinted with permission from Dobson 2003. Copyright Nature
Publishing Group.
3
1.2 Amyloid Disease
The accumulation of misfolded proteins as insoluble, fibrillar aggregates is characteristic of
several degenerative diseases. Examples include the proteins involved in amyloid diseases such
as Alzheimer’s disease (Aβ) (Glenner and Wong 1984), type II diabetes (amylin) (Cooper et al.
1987) and Parkinson’s disease (α-synuclein) (Spillantini et al. 1997), as well as the mammalian
prion diseases (prion protein) which include BSE, CJD and GSS (Prusiner 1982). While
transmissibility and disease onset differ between amyloid and prion diseases, recent evidence
suggests that soluble protein oligomers, rather than fibrils, are the cytotoxic species in each case
(Lambert et al. 1998; Bucciantini et al. 2002; Kayed et al. 2003; Walsh and Selkoe 2004; Silveira
et al. 2005; Baglioni et al. 2006; Simoneau et al. 2007). It has been suggested that these non-
fibrillar assemblies may be a common element of all amyloid diseases, and non-fibrillar
oligomers formed by several amyloid proteins have been identified in vivo (Walsh et al. 2000;
Walsh et al. 2002) or produced in vitro (Uversky et al. 2001; Kayed et al. 2003). Regardless of
protein sequence, these oligomers share several key features, including reactivity to
conformation-specific antibodies, the ability to permeabilize model membranes, and cytotoxicity
to cultured neurons (Kayed et al. 2003; Kayed et al. 2004). However, despite their potential
importance in the pathogenesis of amyloid diseases, the details of the molecular structure of
these non-fibrillar oligomers are only now beginning to emerge, as is their relationship to mature
fibrils, and to the onset of disease.
The mechanism or mechanisms through which these oligomeric species induce cell death and
contribute to the pathology of amyloid diseases remains a matter of some debate. Current
hypotheses include a physical disruption of cellular membranes (Demuro 2005), formation of
amyloid pores or channels (Kostka et al. 2008), induction of oxidative stress (Ebenezer et al.), or
interactions with receptor proteins on the cell surface leading to either altered protein function, or
the initiation of a signaling event (Chong et al. 2006; Um et al. 2012). Defining the link between
the structure of misfolded protein aggregates and the concurrent gain of a toxic functionality is
inhibited by the inherent difficulties of studying aggregative proteins, and is further complicated
by the ability of amyloid proteins and peptides to form several distinct types of oligomers and
fibrils, which often exist as heterogeneous mixtures. Each species of aggregate may exhibit
varied biological activity, different local structure or gross morphology and typically contains
different numbers of monomers per assembly. Despite these challenges, there has been
4
significant recent progress in obtaining high-resolution structural details of amyloid fibrils and
non-fibrillar oligomers, and in defining their biological mode of action. Given the relationship
between amyloid fibrils and oligomers along the misfolding pathway, it is important to gain as
much structural information as possible.
1.3 Structural Studies of Amyloid Fibrils
As the final stage in the assembly pathway for misfolded amyloid proteins, accumulation of
fibrils has long been seen as the hallmark of amyloid diseases. Since they were the only readily
detectable amyloid assembly present in disease tissue, early work suggested that fibrils were
likely to be the mediators of cell death and disease progression (Shirahama and Cohen 1967). In
addition, preparation of stable mature amyloid fibrils has generally been more accessible than the
potentially transient non-fibrillar oligomers, facilitating biophysical and structural analysis. With
recent advances in methodology and instrumentation, high-resolution structural details have been
reported for amyloid fibrils formed by several proteins and peptides, based on data from x-ray
crystallography, electron cryomicroscopy and solid state NMR studies (Petkova et al. 2002;
Jaroniec et al. 2004; Sawaya et al. 2007; Lee et al. 2008; Sachse et al. 2008; Mizuno et al. 2012).
While the details of each structure differ, based on sequence and solution conditions used for
assembly, these studies have confirmed the presence of a cross-β architecture within the core of
all amyloid fibrils studied to date. This structural motif is characterized by having protein or
peptide strands form extended β-sheets running perpendicular to the long axis of the filament,
and was initially identified from x-ray fiber diffraction studies of amyloid fibrils (Eanes and
Glenner 1968; Geddes et al. 1968; Jahn et al. 2009). The cross-β diffraction pattern contains
intense reflections at 4.7-4.8 Å (meridional) and 10 Å (equatorial) due to the characteristic
spacing between β-strands along the long axis and between the perpendicularly stacked β-sheets,
respectively.
In general, the core of most amyloid fibrils is considered to contain a dehydrated interface
between adjacent β-sheets. This result from packing of hydrophobic residues in a water-
excluded core, giving rise to one of 8 possible steric zipper arrangements, as first proposed by
Sawaya et al. (Sawaya et al. 2007) (Figure 1-2). These permutations arise from the fact that
there are 2 possible types of β-sheet (parallel or antiparallel), 2 stacking possibilities (parallel or
anti-parallel) and 2 surfaces for inter-sheet packing (face-to-face or face-to-back). The presence
5
of steric zipper motifs was initially observed in X-ray structures of fibril-like crystals formed by
short amyloidogenic peptides (Sawaya et al. 2007), and a subset of these classes of intersheet
packing have been observed in solid-state NMR structures of amyloid fibrils (Nielsen et al.
2009). It is important to note, however, that recent NMR studies have revealed some possible
structural differences between the crystalline and fibrillar forms of the GNNQQNY peptide
derived from the yeast prion Sup35 (van der Wel et al. 2007), such that more structures of
amyloid fibrils are required to confirm the crystallographic data.
Figure 1-2 – Possible arrangements of β-strands in an amyloid fibril
Eight permutations exist, four containing parallel β-sheets and four containing anti-parallel β-sheets, each with the
possibility of parallel or antiparallel stacking of the two sheets, which may align in a face-to-face or face-to-back
manner. In each case, the interface between the sheets forms a so-called steric zipper, with opposing side chains
interdigitating to exclude water. Reprinted with permission from Nielsen et al., 2009. Copyright 2009 Angewandte
Chemie.
Additional complexity in fibril structure comes from quaternary interactions in which
protofilaments containing a basic building block (for example a filament formed by extended
arrangement of a pair of stacked β-sheets) are bundled or twisted together to form the mature
amyloid fibril. It is clear from electron microscopy studies of fibrils formed by numerous
6
amyloid peptides that significant heterogeneity can exist even between fibrils formed by the
same peptide (Fandrich et al. 2009). This can be rationalized by variations in the interchain,
intersheet, and inter-protofilament packing, as well as conformational heterogeneity between
peptide chains. The heterogeneous nature of many fibril preparations has been supported by
solid state NMR for Aβ(1-40) (Petkova et al. 2005), α-synuclein (Heise et al. 2005), GNQQNY
fibrils (van der Wel et al.), and amylin (Madine et al. 2008).
Probably the best characterized fibril structures are those formed by fragments of the
Alzheimer’s Aβ protein. In particular, several structures for fibrils formed by Aβ(1-40) have
been reported, based primarily on solid state NMR or electron microscopy (Petkova et al. 2002;
Sachse et al. 2008; Chan 2011; Tycko 2011). The fibril morphology and subunit peptide
structure in each case is dependent on the incubation conditions during in vitro fibrillization, and
can exhibit significant heterogeneity in both TEM and NMR experiments. An example structure
for Aβ(1-40) fibrils is shown in Figure 1-3. Each peptide adopts a β-turn-β conformation,
forming parallel in-register β-sheets with neighboring peptides down the long axis of the fibril.
The two sheets pack into an internal class 1 steric zipper motif within the protofilament. In this
structural model, quaternary interactions between two protofilaments were determined using
intermolecular dipolar couplings from solid state NMR, giving rise to the depicted structure for
the mature fibril. These quaternary interactions vary between fibrils with different morphology,
such as the three-fold symmetric fibrils reported by Paravastu et al., (Paravastu et al. 2006) or
those studied by cryoelectron microscopy (Sachse et al. 2008; Schmidt et al. 2009).
7
Figure 1-3 – Structure of amyloid β(1-40) fibrils as determined by solid state NMR
This structure contains a class 1 steric zipper with parallel β-sheets stacked in a face-to-face antiparallel
arrangement. The upper image shows the backbone of several monomers, arranged with the fibril axis extending
into the page, while the lower image focuses on a representative pair of peptides, showing the interdigitation of
sidechains within the hydrophobic core, as well as depicting quaternary contacts between adjacent protofilaments.
Reprinted with permission from Petkova et al., 2002. Copyright 2002 Proceedings of the National Academy of
Science of the United States.
By contrast, only a single well-defined structure has been reported so far for protofilaments
formed by the far more neurotoxic and more aggregative Aβ(1-42) peptide, which is a less
abundant form of Aβ, but which correlated more closely with pathogenesis (Burdick et al. 1992;
Jarrett et al. 1993; Luhrs et al. 2005; Kumar-Singh et al. 2006). This structure is similar to that
of Aβ(1-40), but rather than intramolecular contacts forming the steric zipper, the top strand from
one monomer makes side chain contacts with the bottom strand from an adjacent monomer.
Modeling of the mature fibril based on cryoelectron microscopy and hydrogen/deuterium
exchange measurements has suggested a distinctly different quaternary assembly for Aβ(1-42)
fibrils, but the potential relationship between these structures and the varied biological activity of
the two Aβ peptides remains (Olofsson et al. 2007; Zhang, R. et al. 2009; Miller et al. 2010).
Numerous solid state NMR structures of small amyloid-forming peptides have now been
reported, including short fragments of Aβ (Balbach et al. 2000; Tycko and Ishii 2003), amylin
8
(Luca et al. 2007; Madine et al. 2008), transthyretin (Jaroniec et al. 2002; Jaroniec et al. 2004),
calcitonin (Naito et al. 2004) and neurotoxic fragments of PrP (Cheng et al. 2006; Lee et al.
2008). Some short peptides display alternate packing arrangements in the fibrils, such as the
antiparallel β-sheets formed by Aβ(16-22) (Balbach et al. 2000) or the antiparallel heterozipper
arrangement of amylin(20-29) fibrils (Nielsen et al. 2009). Longer amyloid proteins have
remained more challenging, although progress is still being made. For example, initial studies of
full-length α-synuclein by hydrogen—deuterium exchange and solid state NMR have allowed
identification of secondary structure elements and delineation of the fibril core, but a high-
resolution fibril structure is lacking (Heise et al. 2005; Vilar et al. 2008).
1.4 Prion Protein Structures
Prion diseases are a class of neurodegenerative diseases associated with the accumulation and
associated cell death related to proteinaceous and infectious proteins. These diseases are
observed across various mammals with certain species such as equine and porcine maintaining or
acquiring resistance to prion disease. It is currently accepted that prion disease is caused by
aberrant proteins alone, without any source of DNA or RNA for transmission of the disease.
Figure 1-4 – Solution NMR Structure of the Human Prion Protein (hPrP)
Structure of the non-disease form of the human prion protein (residues 121-231) as determined by solution NMR.
The non-disease form of the prion protein is a 208 residue α-helix rich protein of currently unknown function (Zahn
et al. 2000).
9
In humans, prion diseases can manifest due to genetic factors, spontaneously or as a result of
ingestion of infected protein. While most cases of the disease are genetic they are ultimately,
including spontaneous cases, due to a misfolding of the prion protein. Mutations in the primary
sequence of the protein can cause destabilization of the native state causing the formation of self-
propagating aggregates. For example, D178N, P102L, E200K and V210I are all mutations that
result in disease phenotypes (Beck et al. 2010). In yeast, prions formed by proteins such as
Ure2p (Wickner 1994) and Sup35 (Wickner et al. 1995) infer specific phenotypes on dividing
cells (Colby and Prusiner 2011). The same is true for the fungal prion of HET-s, where infection
does not equate to a degenerative state but rather to signal incompatibility of heterkaryons
(Coustou et al. 1997). These protein systems in fungi and yeast are considered prions because of
their ability to form ordered rod or fibrillar aggregates as well as for their infectivity. Although
these yeast and fungal systems differ from mammalian systems in that they do not cause cell
death, they offer structural biologists a functional prion to work with for structural studies.
The prion protein (PrP) is the major causative agent of neurodegenerative prion diseases, such as
scrapie in sheep, BSE in cattle, and CJD in humans. The protein converts from a monomeric,
primarily helical cellular form (PrPC) shown in Figure 1-4, to an infectious, oligomeric, scrapie
form (PrPSc
), with increased -structure. In addition, there are several known fungal prion
proteins, unrelated to PrP in amino acid sequence, but sharing the ability to adopt a fibrillar,
infectious, β-sheet rich structure. While sharing some common structural elements with fibrils
formed by amyloid proteins, some striking differences have been observed. For example, the
fungal Het-S prion protein structure solved by solid state NMR contains a β-solenoid structure
with two protein molecules per “rung” of the solenoid ladder, rather than the cross-β packing
typical of amyloid (Figure 1-5C) (Wasmer et al. 2008). By contrast, amyloid fibrils formed by
PrP in vitro were shown by electron paramagnetic resonance (EPR) to contain amyloid-like in-
register parallel β-sheet structure (Cobb et al. 2007) (Figure 1-5A), similar to the yeast prion
proteins Ure2 (Baxa et al. 2007) and Sup35 (Shewmaker et al. 2006). Interestingly, it has been
shown through electron crystallography, X-ray fiber diffraction, and molecular dynamics
simulations that the infectious PrPSc
form of PrP from infected brains likely differs from in vitro
fibrils and may contain a β-helix or β-solenoid structure (Figure 1-5B) (Govaerts et al. 2004),
similar to Het-S.
10
Figure 1-5 – Structures formed by aggregated prion proteins from human and yeast.
(A) Structure of amyloid fibrils formed by PrP, showing parallel in-register β-sheets. The structure is also stabilized
by a disulphide bond. Reprinted with permission from Cobb et al., 2007. Copyright2007 Proceedings of the National
Academy of Science of the United States. (B) β-helical structure formed by human PrP taken from infectious
material. Reprinted with permission from Govaerts et al., 2004. Copyright 2004 Proceedings of the National
Academy of Science of the United States. (C) The Het-S prion structure from solid state NMR showing residues218-
289 in a β-solenoid. Reprinted with permission from Wasmer et al., 2008. Copyright 2008 Science
1.5 Non-fibrillar Oligomers on the Misfolding Pathway of Amyloid Proteins
The relationship between the formation of non-fibrillar oligomers and the misfolding pathway
leading to amyloid fibril formation has not been definitively determined. While there have been
conflicting reports (Necula et al. 2007), most evidence points to the spherical, cytotoxic
oligomers existing as on-pathway intermediates. In particular, various prefibrillar oligomers of
Aβ have been shown to be transient, disappearing as they reorganize into mature fibrils (Chimon
and Ishii 2005). Similarly, pore forming oligomers of α-synuclein are considered to be on-
pathway for fibrillization (Kim et al. 2009). From a mechanistic standpoint, the structural data on
prefibrillar oligomers suggests early adoption of an extended β-structure, followed by formation
of tertiary and quaternary contacts as the oligomers increase in size. The precise steps involved
in the transition from discoidal or spherical oligomers to an extended amyloid fibril have not
been determined, but likely involve an increase in the tightness of lateral associations between
strands, with optimized hydrophobic packing and hydrogen bond formation driving the final
steps of assembly. Taken together, the transient nature and fibril-like structure show that these
entities exist on the aggregation pathway toward fibrils. In contrast, annular oligomers do not
appear to exist as productive intermediates, but may instead represent off-pathway assembly. In
11
the case of Aβ, it has been shown that in the presence of lipid membranes, prefibrillar oligomers
are capable of rearranging to form annular oligomers, suggesting that in this case they may
represent an alternate end-stage of the misfolding pathway (Kayed et al. 2009). This may also
present a possible mechanism for formation of membrane-disrupting entities from the on-
pathway non fibrillar oligomers.
1.6 Structural Studies of Non-Fibrillar Amyloid Oligomers
While a wealth of structural information is becoming available for the fibrillar forms of many
model and disease related amyloid proteins and peptides, relatively little is known about the
molecular structure of non-fibrillar oligomers formed by the same polypeptides. Structural
characterization has been made particularly challenging by the transient nature of many of these
assemblies, which are widely considered to form as intermediates along the amyloid misfolding
pathway. Thus, the difficulty of obtaining highly pure samples of non-fibrillar oligomers which
are sufficiently long-lived for biophysical studies has significantly slowed progress in this field.
A number of studies have used small molecules, including detergents or lipids, to trap or
stabilize oligomeric states of amyloid proteins (Laurents et al. 2005; Yu et al. 2009), but this
approach risks formation of off-pathway or non-biological assemblies, rather than the on-
pathway intermediates likely to play a role in amyloid disease (Kayed et al. 2003).
Despite these challenges, however, a number of low-resolution studies have been reported, using
TEM, atomic force microscopy (AFM), hydrogen/deuterium exchange, and fluorescence
spectroscopy-based approaches (Huang, T. H. et al. 2000; Williams et al. 2005; Losic et al. 2006;
Ono et al. 2009). Microscopy and size exclusion chromatography have shown that, similar to
amyloid fibrils, there are a wide range of non-fibrillar oligomers that can be categorized based on
their size (ranging from dimers of Aβ(1-40) to large spherical assemblies containing hundreds of
peptide monomers) or morphology (Haass and Selkoe 2007; Walsh and Selkoe 2007). In terms
of the latter, most oligomers reported have either exhibited a roughly globular appearance by
AFM and TEM, or have been annular in nature – exhibiting a pore or ring shaped structure
(Janson et al. 1999; Conway et al. 2000; Lashuel et al. 2002). These two morphologies appear to
exhibit different degrees of biological activity, with spherical oligomers, but not annular
oligomers, increasing membrane conductance and inducing apoptosis in cell culture (Kayed et al.
2009). The large (3-10 nm diameter), spherical oligomers formed by several amyloid proteins
12
have been shown to bind to a single conformational antibody, suggesting that a common
structural motif exists in these assemblies, despite having no sequence similarity. Antibody
binding was also shown to inhibit the inherent cytotoxicity of these large amyloid oligomers
(Figure 1-6).
Figure 1-6 – Toxicity and common structural architecture of amyloid oligomers
A graph showing the viability of neuroblastoma SH-SY5Y cells, monitored by the 3-(4,5-dimethyl-2-thiazoyl)-2,5-
diphenyltetrasodium bromide (MTT) reduction assay, as a function of treatment with preparations of several
amyloid proteins. The toxicity of non-fibrillar oligomers formed by each peptide is shown to significantly decrease
cell survival (black bars) relative to the control, soluble (presumed monomeric) peptide and mature amyloid fibrils.
In each case, the effects of the oligomers on cell survival are attenuated by the addition of an amyloid oligomer
specific antibody (A11, white bars). Non-specific IgG is shown in hatched bars, and exerts no effect on the system.
Reprinted with permission from Kayed et al., 2003. Copyright 2003 Science
Likewise, annular oligomers formed by Aβ(1-42), amylin and α-synuclein are all recognized by
an antibody that does not bind to monomeric or fibrillar material, and that shows only weak
binding to spherical oligomers, indicating that these contain distinct structural elements from the
other assemblies (Kayed et al. 2009).
13
More recently, solid state NMR has been successfully used to obtain high-resolution structural
details of non-fibrillar oligomers formed by Aβ (Chimon and Ishii 2005; Chimon et al. 2007) and
α-synuclein (Kim, H. Y. et al. 2009), and solution NMR has been used to investigate the
structure of small detergent stabilized oligomers of Aβ(1-42) (Yu et al. 2009). Advances in
computational infrastructure and methodologies have also led to an increased use of molecular
dynamics simulations to investigate the structure and assembly of non-fibrillar amyloid
oligomers.
The non-fibrillar oligomers formed by Aβ(1-40) and Aβ(1-42) have been implicated as the main
toxic species associated with Alzheimer’s disease, and as such have been the focus of the
majority of studies on amyloid oligomers reported to date. Structural characterization has been
impeded by the wide spectrum of oligomeric states that can be adopted by these peptides along
their aggregation pathways. As indicated above, species ranging in size from dimers to
oligomers containing hundreds of peptides have been reported, both in vitro, and in material
isolated from the brains of Alzheimer’s patients (Haass and Selkoe 2007; Walsh and Selkoe
2007).
The larger oligomers can also be subdivided into spherical, so-called pre-fibrillar oligomers and
ring-shaped annular oligomers, each with different antibody reactivity. From a high-resolution
standpoint, most experimental progress has been made in defining the molecular structures of
small and large pre-fibrillar oligomers formed by Aβ, although numerous molecular dynamics
simulations have been carried out on membrane-bound amyloid channels or pores that closely
resemble the overall morphology of annular protofibrils as seen in TEM and AFM images (Jang
et al. 2007; Zheng et al. 2008). These annular oligomers are 8-20 nm in diameter by TEM and
AFM, and like the spherical oligomers, circular dichroism (CD) spectroscopy shows that they
contain high levels of β-sheet (Kayed et al. 2009). The anti-annular oligomer antibodies also
bind to the β-barrel pores formed by the bacterial toxin α-hemolysin, such that they may share
the same general architecture (Kayed et al. 2009). Interestingly, preformed annular oligomers did
not permeabilize membranes, instead converting to prefibrillar oligomers upon interaction with
membranes. This may suggest that any pore like structure formed by Aβ would need to
assemble within the membrane, rather than acting through insertion of a preformed assembly.
14
In the pre-fibrillar oligomers, the structural data that has emerged from recent studies suggests
that even at the earliest stages of aggregation they share common features with the fibrillar forms
of Aβ. For example, Yu et al., used 0.05% SDS to stabilize very small pre-globulomers and
globulomers of Aβ(1-42), with molecular weights of 16 and 64 kDa respectively (Yu et al.
2009). These were assumed to represent very early points in the amyloid aggregation pathway,
and structural studies were conducted using solution NMR. The intra-chain and inter-chain
contacts in these oligomers share similarities with the Aβ (1-40) and Aβ(1-42) fibril structures
reported to date. Both contain similar secondary structure elements with the fibrillar form, and
contain intermolecular contacts reminiscent of the fibrils, although in the small oligomers, the N-
terminal strand folds back on itself, rather than participating in intermolecular β-sheet formation
(Figure 1-7A). In a similar vein, DSS was used to stabilize very large (764kDa) Aβ(1-40)
oligomers, and subsequent structural analysis suggested the presence of micelle-like assemblies
containing a radial arrangement of Aβ monomers in an extended β-sheet conformation (Figure 1-
7C) (Laurents et al. 2005). The nature of intermolecular or intramolecular β-sheets was not
determined in this study, so it is difficult to relate the resulting models to the fibrillar form of the
protein.
For both of the aforementioned studies, it is important to note that the effect of detergents and
other small molecules on the structure and assembly of amyloid peptides remains unclear.
Addition of cofactors may lead to formation or stabilization of otherwise unpopulated structures.
Recent studies on on-pathway prefibrillar oligomers of Aβ(1-40) and Aβ(1-42) have
circumvented this requirement by using either gel filtration and lyophilization (Chimon and Ishii
2005; Chimon et al. 2007) or careful modulation of solution salt and pH conditions to trap non-
fibrillar oligomers for structural studies (Ahmed et al. 2010).
Solid state NMR of large (15-35 nm) spherical oligomers of Aβ(1-40) prepared by freeze-
trapping revealed fibril-like secondary and quaternary structures, leading to a model in which the
location and intermolecular assembly of β-sheets is shared between the two forms (Chimon et al.
2007). A schematic for the proposed architecture of these oligomers is shown in Figure 1-7B,
along with a model of the Aβ(1-40) protofilament structure determined by Petkova et al.
(Petkova et al. 2002). This micelle-like arrangement is reminiscent of that proposed for DSS-
stabilized oligomers (Figure 1-6C), potentially validating the use of small molecules to trap
15
transient amyloid oligomers. These large oligomers were shown to exhibit neurotoxicity, and
based on their transient nature can be assumed to lie on the fibril assembly pathway.
Figure 1-7 – Structures of non-fibrillar amyloid oligomers
(A) Pre-globulomer (top) and globulomer (bottom) structures formed by Aβ(1-42) are shown. Both structures show
similarities to the basic Aβ(1-40) fibril subunit shown in Figure 1-2B. Reprinted with permission from Yu et al.,
2009. Copyright 2009 The American Chemical Society. (B) Structural model of large spherical Aβ(1-40) oligomers
obtained using solid state NMR. Reprinted with permission from Chimon et al., 2008. Copyright 2008 Nature
Publishing Group. (C) A structural model of large, DSS stabilized Aβ(1-40) oligomers shown as extended micelle-
like structures, approximately 35 nm in diameter. Significant structural similarity with the solid state NMR derived
model shown in (B) is evident. Reprinted with permission from Laurents et al., 2005. Copyright 2005 Journal of
Biological Chemistry.
Ahmed et al. have used altered solution conditions to trap discoidal pentamers and decamers of
Aβ(1-42) with potent neurotoxicity (Ahmed et al. 2010). When incubated at 37°C for several
hours, these oligomers convert to amyloid fibrils, suggesting that they are productive
intermediates on the assembly pathway. In contrast to the large Aβ(1-40) oligomers studies by
Chimon et al. Fourier transform infrared spectroscopy (FTIR) and solid-state NMR studies of
these small oligomers indicated the presence of significantly increased disorder and solvent
accessibility relative to fibrils of Aβ(1-42), and showed that the oligomers lack the in-register
parallel β-sheet architecture of the fibrillar form (Chimon et al. 2007). The oligomeric peptides
do, however contain the same β-loop-β secondary and tertiary fold observed in Aβ(1-42) and
Aβ(1-40) fibrils. This is supported by molecular dynamics and hydrogen-deuterium exchange
studies from several other groups, and leads to an overall picture in which Aβ peptides adopt a β-
loop-β structure as a common element of all oligomeric states, with intermolecular contacts and
solvent accessibility varying between different types of oligomers. These results also lead to the
16
general concept that early intermediates formed during Aβ assembly may be more solvent
accessible and potentially more labile, and that conformational flexibility is likely to play an
important role in their biological activity (Cheon et al. 2007; Zhang, A. et al. 2009; Yu et al.
2010; Pan et al. 2011; Yu and Zheng 2011).
1.7 Non-fibrillar amyloid oligomers as the cytotoxic agents in amyloid disease
While early studies focused on the amyloid fibrils or plaques as the causative agents of
neurotoxicity in Alzheimer’s disease, more recently it has become evident that small non-fibrillar
oligomers correlate much more closely with loss of neuronal function and neurodegenerative
disease progression (Kayed et al. 2003; Haass and Selkoe 2007; Walsh and Selkoe 2007) This
finding has been echoed for non-fibrillar oligomers formed by a broad array of disease related
and non-disease related amyloid proteins (Baglioni et al. 2006). Given the potential for some
amyloid oligomers to have similar structural properties, regardless of amino acid sequence, it is
possible that many of these may act via a similar toxic mechanism. The conformations accessible
to aggregative proteins may create interactions with components of the cellular ion transport
system or may allow them to form channels or pores in cell membranes (Lin, H. et al. 2001;
Kayed et al. 2004; Demuro et al. 2005). This may represent a general mechanism through which
cytotoxic effects are exerted during the early stages of protein aggregation. Supporting this
hypothesis, soluble amyloid oligomers with spherical morphology, induce vesicle leakage, and
are toxic to cultured cells, possibly through disruption of calcium homeostasis (Thellung et al.
2000; Demuro et al. 2005; Ferreiro et al. 2008).
Alternatively, membranes may enhance amyloid aggregation and membrane binding of many
amyloid peptides has been described extensively (McLaurin and Chakrabartty 1996; Yip et al.
2002; Kayed et al. 2004). Once bound to the membrane surface, or inserted into the bilayer,
non-fibrillar oligomers would have the potential to rearrange into channels, pores, or non-
specific aggregates at the membrane surface. Any of these mechanisms are likely to cause
membrane destabilization and cell death, and it has recently been demonstrated for Aβ oligomers
that increased membrane conductance can occur in the absence of channel formation (Sokolov et
al. 2006). Physical disruption such as the introduction of membrane defects, possibly through
insertion of oligomers, or through membrane-catalyzed fibril formation, would also be sufficient
17
to induce leakage of cell contents and ultimately lead to cell death (McLaurin and Chakrabartty
1997; Yip et al. 2002).These interactions with membranes are summarized in Figure 1-8
including the formation of a pore, as well as fibrillization into the membrane in a raft-like
manner (Berthelot et al. 2013).
Figure 1-8 – Various Methods of Membrane Disruption by Amyloids
A schematic showing various ways that amyloid proteins or peptides can interact with lipid membranes. They
include simple binding causing disruption; carpeting – whereby aggregation at the surface of them membrane causes
disruption; pore formation; loss of lipids due to proteins or peptides acting as detergent; and a raft-like disruption
where fibrillization into the membrane causes defects in the phospholipid membrane. Reprinted with permission
from Berthelot et al 2013. Copyright 2013 Biochimie.
While the oligomer fold is distinct from that of fibrils, as determined by differential antibody
reactivity, a common theme emerging from structural studies of non-fibrillar amyloid oligomers
is the presence of local fibril-like structure. While it does not speak to the actual mechanism
through which toxicity is exerted, this observation may suggest that small fibril-like assemblies
are the key element required for cytotoxicity. A similar phenomenon has been reported by Xue
et al. (Xue et al. 2009), who demonstrated that fragmentation of mature amyloid fibrils formed
by α-synuclein, β2-microglobulin and lysozyme leads to an increase in membrane disruption and
cytotoxicity (Figure 1-9). Likewise amyloid fibrils formed by hexapeptides gained cytotoxicity
towards primary neuronal cell culture after physical disruption (Pastor et al. 2008). In both
18
studies, it is likely that the increase in active ends allows improved interactions with cellular
targets – membranes or other cell surface molecules, where they are able to rearrange to form
active, toxic entities. Oligomeric species, which are known to be more conformationally flexible
and less stable than their fibrillar counterparts, may act through a similar mechanism, carrying
active fibril-like segments to the site of toxic activity.
Figure 1-9 - The relationship between amyloid fibril length and toxicity.
As the concentration of small fragments increases, increased membrane disruption and cellular toxicity are
observed. As the fragments become smaller, it is proposed that they will be increasingly toxic to the cell. Reprinted
with permission from Xue et al., 2009. Copyright 2009 Journal of Biological Chemistry.
1.8 PrP(106-126) Peptide as a Model for Amyloid Diseases
Regions of the mammalian prion protein thought to be important in aggregation and conversion
were identified by Tagliavini et al. using secondary structure propensity and hydropathy
(Tagliavini et al. 1993). Two sections of the prion protein, both contained in the sequence found
19
to cause GSS were identified. Peptides comprised of residues 106-126 and 127-147 were found
to form amyloid fibrils based on x-ray fiber diffraction. They also both caused green
birefringence when exposed to the dye congo red. While both of these peptides are able to form
amyloid fibrils, PrP(106-126) was ultimately determined to be more amyloidogenic and, as such,
was focused on more intensely for structural, mutagenesis, membrane interactions and toxicity.
Early studies of the peptide found that PrP(106-126) formed β-sheet containing structures across
a variety of pH’s and conditions and was neurotoxic. Salmona et al. were able to describe an
initial model for how peptide monomers were arranged in the fibrils formed by PrP(106-126)
(Figure 1-10) (Salmona et al. 1999). They found that the amino acid change A117V decreased
fibril formation, suggesting that this residue must be in the inward facing core of the stacked β-
sheets with the reasoning that a bulkier side-chain would prevent the core from forming.
Additionally, they found that amidation of the C-terminus prevented fibril formation and from
this, that the C-terminus must form a salt-bridge with Lys110. Hydrogen-deuterium exchange
and molecular dynamics simulations performed on mouse PrP(106-126) showed the presence of
parallel β-sheets stacked antiparallel (Kuwata et al. 2003).It should be mentioned that human and
mouse sequences differ at position 112, with mouse containing a valine in place of methionine.
Figure 1-10 – PrP(106-126) fibril model of peptide stacking based on mutagenesis
A model of how peptide monomers are stacked by Salmona et al. Formation of the salt bridge was postulated based
on the observation that amidation of the C-terminus abolished fibril formation (Salmona et al, 1999). Reprinted with
permission from Salmona et al. Copyright Biochemical Journal. 1999.
The PrP derived peptide PrP(106-126) poses an interesting structure-toxicity relationship given
the ability of the peptide to form both amyloid fibrils and cytotoxic oligomers, making it a useful
model for studying the structural and mechanistic details of non-fibrillar amyloid oligomers
20
(Forloni et al. 1993; Selvaggini et al. 1993; Jobling et al. 1999; Salmona et al. 1999). For
example, in studies by Kayed et al., non-fibrillar oligomers of PrP(106-126) were shown to form
large (10-20 nm diameter) spherical oligomers with similar morphology to Aβ, amylin, and
several other amyloid proteins (Kayed et al. 2003). These oligomers cause increased membrane
conductance and were cytotoxic to neuronal cell cultures, and have also been shown to disrupt
model-membranes (Kayed et al. 2004).
There have been conflicting reports on the toxicity of PrP(106-126) largely due to confounding
effects of its ability to form amyloid oligomers as well as potentially playing a role in conversion
of full-length PrP to the infectious PrPSc
form (Gu et al. 2002). PrP(106-126) has been shown to
be toxic in a number of different ways. Reports initially characterized PrP(106-126) as requiring
full-length PrP for toxicity in cerebral endothelial cells (Deli et al. 2000). There is also
significant evidence for PrP-independent cytotoxicity, but it is important to note that in most
studies of PrP(106-126), the aggregation state of the peptide was not clearly defined, so the
activity of prefibrillar oligomers is implicit rather than explicit in the results. PrP(106-126) has
been shown to interact with L-type voltage sensitive calcium channels, causing apoptosis (Florio
et al. 1998; Silei et al. 1999; Thellung et al. 2000). It has also been demonstrated that this
peptide causes the activation of the JNK-c-Jun pathway, rapidly leading to apoptosis (Carimalo
et al. 2005).
There have been several reports of direct membrane destabilization by PrP(106-126), including
the formation of ion channels (Lin, M. C. et al. 1997), or alterations in membrane viscosity
across a number of different cell types including human granulocytes and rat glial cells in vitro
(Salmona et al. 1997). More recent work using prefibrillar oligomers of PrP(106-126) have
shown that it permeabilizes membranes (Kayed et al. 2004) and induces cytotoxicity in
neuroblastoma cell cultures (Kayed et al. 2003). It is well known that PrP(106-126) interacts
with phospholipid membranes even as a monomeric peptide, with lipid composition playing a
role both in interaction and post-binding events. For example, PrP(106-126) has been shown to
cause the aggregation of liposomes containing the ganglioside GM1 (Kurganov et al. 2004).
While there is no direct link to the disruption of calcium channels or activation of the JNK-c-jun
pathway, current evidence supports direct membrane interaction and disruption as a mechanism
for PrP(106-126) cytotoxicity – at least in the absence of cell surface PrP. While recent results
for Aβ show that alterations of the membrane are sufficient to increase conductance without
21
requiring channels, rearrangement of the protein in the membrane to form discrete pores or
channels cannot be ruled out (Eliezer 2006; Sokolov et al. 2006).
1.9 Biological Applications of Solid State Nuclear Magnetic Resonance
Solid state NMR has seen a relatively large increase in application to structural biology in the
last 10 years. This growth in biological solid state NMR can be attributed to a surge in available
hardware and methodologies as well as a need for a technique to study systems which have more
recently come to the forefront of research, such as amyloids and membrane active proteins. For
these systems solid state NMR has a significant advantage; one can study proteins and peptides
in their “native” environment, be it in or on a lipid bilayer or in an aggregated and or crystalline
state. The only real limitation to the size of the particle being studied is the ability to resolve each
individual resonance; the adaptation of sample preparation, selective labeling schemes, and
introduction of new 3-dimensional experiments has allowed for larger and more complex
systems to be studied. While crystallography can require the use of non-native conditions such
as detergents or salts, the technique has proved very useful in the determination of structures of
small amyloid peptides at very high resolution (Sawaya et al. 2007). The same can also be said
about the use of solution NMR for the study of aggregated proteins through the use of hydrogen-
deuterium exchange (Vilar et al. 2012) or saturated transfer difference (Huang, H. et al. 2008).
1.9.1 Magic Angle Spinning (MAS)
The use of solid state NMR offers an interesting approach in that samples in the solid state are
inherently anisotropic, meaning these samples contain distance information that is not time-
averaged by fast molecular tumbling. These same orientation-dependent interactions, namely
chemical shift anisotropy and dipolar coupling can lead to undesirable line broadening, making
the assignment of individual resonances difficult. If we consider a small protein to be studied by
solution NMR, the fast molecular tumbling of that protein averages these interactions giving rise
to resolvable lines. In contrast, a solid state NMR sample containing fibrils, for example, would
tumble very slowly, even if suspended in solution. This gives rise to lines associated with every
orientation of each spin. As an example, if we take the equation for the anisotropic portion of the
chemical shift frequency (Equation 1-1), we see that the orientation of the (axially aligned)
chemical shielding tensor relative to the external magnetic field is dependent on the angle θ.
22
(Equation 1-1) (Duer 2007)
Equation 1-1 shows the anisotropic chemical shift frequency (ωcs) for an axially symmetric
chemical shielding tensor. ω0 refers to the frequency of the rotating frame and σiso is the isotropic
component of the chemical shift. In this case, the chemical shielding tensor is aligned with the z-
axis of the principle axis frame (PAF). The angle θ is the angle between the external magnetic
field and the z-axis of the PAF.
Take for example, a sample of freeze-dried glycine; acquiring a static spectrum of this sample
would reveal a “powder pattern” where each broad component is a sum of the chemical shift for
all orientations of that nucleus in the sample, and contains additional broadening from
homonuclear and heteronuclear dipole couplings. In order to achieve individual, resolved lines
needed for most applications, the orientation dependence of the terms such as CSA and dipolar
couplings must be removed; a feat which is achieved naturally in solution NMR by molecular
tumbling. We therefore employ mechanical rotation of the solid in order to mimic the tumbling
achieved in the solution state. The solution where this 3cos2 θ-1=0 (and thus, for example the
chemical shift anisotropy interaction averages to zero) comes at an angle of 54.74° with respect
to the external magnetic field and is represented in Figure 1-11. Thus, 54.74° is termed the
magic angle and, at increasing spinning speeds, leads to the averaging of orientation dependent
terms such as CSA and removing most dipolar coupling terms resulting in the observation of
isotropic chemical shifts.
23
Figure 1-11 – Schematic Representation of Magic Angle Spinning
In this schematic representation, the NMR sample is spinning at an angle of 54.74° relative to the external magnetic
field (B0). The angle 54.74° allows for the averaging of the orientation-dependent terms and interactions such as
CSA and dipolar couplings. Reprinted with permission from Alia et al. Copyright Photosynthesis Research 2009.
1.9.2 Chemical Shift
Chemical shift in NMR results from the shielding or deshielding of the nucleus from the external
magnetic field created by local electrons. By comparing the NMR frequency of a 13
C carbon in a
given protein or peptide with a reference compound, an absolute scale is created. Chemical shift
itself is comprised of 2 components: isotropic chemical shift – a scalar with no directional
component and anisotropic chemical shift and a second order tensor with directionality (Duer
2007). Isotropic chemical shift is the more commonly thought of entity as observed in solution
NMR. Chemical shift anisotropy (CSA), represented in equation 1-1, can be thought of as an
ellipsoid whose longest axis is along Z that imparts a direction to the chemical shielding
associated with electrons surrounding the nucleus. In solution NMR, this CSA is completely
averaged away by molecular tumbling; however, in solid state NMR, the lack of molecular
tumbling means that this interaction is ever-present. One advantage that solid state NMR has is
its ability to average this CSA away through the use of MAS or retain it during static
experiments to exploit the directionality of the interaction for orientation information. In the end,
chemical shift provides a powerful tool available in solid state NMR through the use of chemical
24
shift analysis. Since the chemical environment of the backbone nuclei of proteins (such as HN,
HA, CA, CB, CO) is dependent on the torsion angles, measurement of chemical shifts can be
made to determine the secondary structure of a protein. Most commonly in solid state NMR, the
measured isotropic chemical shifts for α, β and carbonyl carbons are compared to standard values
for amino acids in a random coil confirmation (Wishart and Sykes 1994). These chemical shifts
can also be used to predict backbone angles using computational methods (Shen et al. 2009).
More recently, chemical shifts have been used as a stand-alone method for protein structure
determination (Robustelli et al. 2008).
1.9.3 Dipolar Coupling
The dipolar interaction between two spins can be described by the dipolar coupling constant
(Equation 1-2) (Duer 2007)
where γI and γS are the gyromagnetic ratios of two spins I and S. The strength of the interaction
is inversely proportional to r3 (r being the distance between the two spins I and S) meaning that
qualitative and quantitative information is accessible between spins through the dipolar
interaction. Dipolar coupling between two spins is also orientation dependent, carrying the same
3cos2 θ-1 term as CSA. As described above, orientation dependent terms such as dipolar
coupling and chemical shift anisotropy are averaged away by MAS. Therefore, in order to utilize
dipolar couplings in rotating solids, we must reintroduce the dipolar interaction through the use
of specific pulse sequences. It is possible to utilize dipolar couplings in both a qualitative or
quantitative fashion for the acquisition of multiple dipolar coupled spins and their relative
distances (Takegoshi et al. 2001; Morcombe et al. 2004). Since there is a strong use for 13
C
nuclei in solid state NMR, we will focus on experiments that utilize 13
C. To do this, protons are
utilized for the transfer of 13
C magnetization; at slow spinning speeds, this can be done in the
absence of 1H RF application during mixing and is called Proton Driven Spin Diffusion (PDSD).
At higher spinning speeds, spin exchange is reduced enough that this is rendered less effective
and RF must therefore be applied during the mixing period to achieve the desired magnetization
transfer and is called RF-aided Diffusion (RAD) (Morcombe et al. 2004). To achieve
magnetization transfer through dipolar coupled protons in RAD (coupled to 13
C as well as other
1H), an RF field is applied to
1H at a frequency of ωr or 2ωr. This interfering field recouples
13C-
25
1H and
1H-
1H dipolar couplings and causes the diffusion of
13C magnetization (Morcombe et al.
2004).
In contrast to 3rd
spin mediated recoupling such as RAD, 13
C or 15
N homonuclear dipolar
interactions can be directly recoupling using rotor synchronized 180° pulses. For example, the
PITHIRDS-constant time experiment utilizes 180° pulses equal to 1/3 of the rotor period. In the
experiment, a set of one-dimensional spectra are acquired for each recoupling time. The pulse
sequence itself is comprised of two blocks – a block that does not allow for dipolar recoupling
and a block which recouples 13
C nuclei close in space. As the recoupling block is increased, the
non-recoupling block is decreased by the same amount of time giving rise to a constant time
experiment and thus alleviating the need for a reference spectrum to be acquired for each
recoupling time. As 13
C dipoles are recoupled, the signal associated with these nuclei will begin
to dephase as a function of the recoupling time and distance apart. The normalized signal
intensity can then be plotted and compared with computer simulated curves to give accurate
distance restraints between two given nuclei.
The use of dipolar recoupling in solids therefore presents a complimentary system for the
acquisition of distance information in biological solids. In RAD, one utilizes uniform or sparse
13C/
15N isotopic labeling to gather qualitative distance information; while PITHIRDS allows for
quantitative distance restraints to be acquired in very sparse or singly labeled proteins.
1.9.4 Solid State NMR in Lipids
The overall anisotropic environment of the lipid bilayer makes the study of lipids by solid state
NMR both feasible and advantageous. 31
P is an NMR active nucleus which is 100% naturally
abundant allowing phosphorus NMR to be performed without the addition or incorporation of
any exogenous nuclei. This allows for the acquisition of both static and MAS spectra of lipid
bilayers. For the purposes of biological relevance, lipid bilayers are usually studied in the fluid
phase where there is a large amount of lipid rotation (Auger 2010). The powder patterns of lipid
bilayers are defined by the phase behaviour and associated dynamic motions of the lipid head
groups. Since the CSA is the dominant interaction in static 31
P spectra, the molecular motions
exhibited by membranes allow bilayers’ phase behaviour to be determined, as well as changes in
size and order of the liposomes. To do this, one examines the change in the powder pattern in
response to external parameters. Comparison of spectra over various conditions can give
26
information on the relative size of the lipid assemblies, the order of the headgroups and the phase
of the lipids. Since different phases of lipids inherently have different molecular motions and
orientations, the observed powder pattern for different phases of lipids will have distinctive
characteristics. For instance, small, fast-tumbling micelles display narrow, isotropic 31
P static
spectra, while lamellar phase shows an asymmetric powder pattern with the major component
upfield from the isotropic value (Seelig 1978; Auger 2010). The effect of the bee venom melittin
on phospholipid membranes was one of the first to be studied by solid state NMR. In this case,
the reduction in the breadth of the 31
P powder pattern in response to the addition of melittin
demonstrated that this venom causes lysis of phospholipid membranes as well as the formation
of melittin-lipid vesicles (Dufourc et al. 1986).
27
2 Core Structure of Amyloid Fibrils Formed by Residues 106-126 of the Human Prion Protein
Sections of this chapter were originally published in the journal Structure. Walsh, P., Simonetti,
K., and Sharpe, S. Core Structure of Amyloid Fibrils Formed by Residues 106-126 of the Human
Prion Protein. Structure. 2009. 17(3). pp 417-426. Reprinted with permission. Copyright AAAS
2009.
All experiments were carried out by P. Walsh with the exception of rotational resonance (RR)
experiments which were carried out by S. Sharpe.
28
2.1 Abstract
Peptides comprising residues 106-126 of the human prion protein (PrP) exhibit many features of
the full-length protein. PrP(106-126) induces apoptosis in neurons, forms fibrillar aggregates,
and is able to mediate the conversion of native cellular PrP (PrPC) to the scrapie form (PrP
Sc).
Despite a wide range of biochemical and biophysical studies on this peptide, including
investigation of its propensity for aggregation, interactions with cell membranes, and PrP-like
toxicity, the structure of amyloid fibrils formed by PrP(106-126) remains poorly defined. In this
study we use solid state NMR to define the secondary and quaternary structure of PrP(106-126)
fibrils. Our results reveal that PrP(106-126) forms in-register parallel -sheets, stacked in an
antiparallel fashion within the mature fibril. The close intermolecular contacts observed in the
fibril core provide a rational for the sequence dependent behaviour of PrP(106-126), and provide
a basis for further investigation of its biological properties.
29
2.2 Introduction
Prion diseases are fatal neurodegenerative disorders characterized by the accumulation of a
misfolded and infectious form of the prion protein (PrP) (Prusiner 1982; Prusiner et al. 1998).
These diseases include bovine spongiform encephalopathy (BSE) in cattle, and Creutzfeldt-
Jakob disease (CJD) in humans. The pathogenic form of PrP (PrPSc
) is generally considered to be
the causative agent for prion disease, due to the ability of PrPSc
to propagate the misfolding of
PrPC in exposed cells, leading to conversion to the disease phenotype in the host organism
(Prusiner 1996; Reisner 2003; Weissmann 2004). The mechanism through which PrPSc
catalyzes
PrPC conversion from a soluble, helical protein into -sheet containing aggregates has not yet
been elucidated, nor has the structure of PrPSc
. Therefore, structural studies of PrP and related
peptides in non-native conformations represent an important step towards understanding the
molecular details of prion disease.
PrP(106-126) is a 21-amino acid peptide derived from the unstructured N-terminus of PrP, and
exhibits many characteristics of the full-length protein (Singh et al. 2002). For instance, PrP(106-
126) has been shown to form amyloid fibrils and to cause apoptosis of cultured neurons through
mechanisms that appear similar to those induced by exposure to PrPSc
(Forloni et al. 1993;
Ettaiche et al. 2000; Forloni et al. 2000; Thellung et al. 2000). Of particular interest is the
observation that toxicity of PrP(106-126) requires the presence of full-length PrPC in the target
cells, consistent with its reported ability to mediate the conversion of PrPC to PrP
Sc, in vitro and
in vivo, and implying a direct interaction between the peptide and PrPC (Forloni et al. 1993;
Selvaggini et al. 1993; Jobling et al. 1999; Salmona et al. 1999). While the molecular structure of
residues 106-126 in PrPSc
remains unknown, it has been shown that monoclonal antibodies raised
against aggregated forms of PrP(106-126) also recognize PrPSc
in human brain tissue from CJD
patients, strongly suggesting a similarity between the structures adopted by aggregates of the
PrP(106-126) peptide and residues 106-126 in full-length PrPSc
(Jones et al. 2008).
PrP(106-126) also shares common features of the fibril forming proteins observed in amyloid
diseases. In particular, there is evidence for direct formation of cation channels as well as non-
specific disruption of model membranes by oligomeric PrP(106-126) (Lin, M. C. et al. 1997;
Kourie, J.I. and Culverson 2000; Dupiereux et al. 2005), similar to the effects of other amyloid
proteins (Kayed et al. 2003; Demuro et al. 2005; Quist et al. 2005). Thus, studies of PrP(106-
30
126) structure are likely to shed light on common elements of amyloid disease. Models for
amyloid fibrils of this peptide have been proposed based on site-directed mutagenesis studies
(Salmona et al. 1999), and on amide hydrogen exchange data (Kuwata et al. 2003). In addition,
recent FTIR and solid state nuclear magnetic resonance (NMR) investigations have revealed an
antiparallel -sheet arrangement in fibrils of an amidated form of the related peptide PrP(109-
122) (Silva et al. 2003; Lee et al. 2008). However, despite extensive biochemical
characterization, no high resolution structural studies of aggregated states of PrP(106-126) have
been reported.
Here we present a detailed structural model for amyloid fibrils formed by PrP(106-126) based on
solid state NMR, transmission electron microscopy (TEM), and atomic force microscopy (AFM).
In particular, we have identified intermolecular contacts between peptides which are
characteristic of parallel -sheets, as well as contacts arising from an antiparallel packing of
sheets within the fibril. These measurements allow unambiguous definition of the key
interactions which define the structure of the fibril core. In the resulting structural model, the
hydrophobic residues pack in a manner similar to the recently described class 1 steric zipper
motif (Sawaya et al. 2007). This core structure is further stabilized by intersheet salt bridges
between the C-terminus and the side chain of Lys110. This model provides a basis for
understanding the effects of sequence alterations and modifications on the biological and
biophysical behaviour of this peptide.
31
2.3 Materials and Methods
2.3.1 PrP(106-126) Fibril Formation
PrP(106-126) peptides were prepared by solid phase peptide synthesis, using standard FMOC
chemistry (APTC, Hospital for Sick Children). For NMR experiments, six selective labelling
schemes were devised, with incorporation of 13
C and 15
N amino acids (Spectra Stable Isotopes
and Cambridge Isotope Laboratories) at the sites indicated in Table 2-1. The final product was
purified by reverse phase HPLC, using an 11 x 300 mm C8 peptide column (Vydac) and a
gradient of 0 to 54% acetonitrile (ACN) with 0.1% TFA. PrP(106-126) eluted at 32% ACN
(confirmed by mass spectrometry) and was freeze-dried.
Peptide Amino acid sequence
PrP(106-126)AVG
KTNMKHMA113
GAAAAGAV121
VG123
GLG
PrP(106-126)GAVL
KTNMKHMAG114
AA116
AAGAVV122
GGL125
G
PrP(106-126)AAGG
KTNMKHMAG(2-13
C)A115
A(1-13
C)A117
A(1-13
C)G119
AVVG(2-13
C)G124
LG
PrP(106-126)AVG2
KTNMKHMAGAAAA118
GAV121
VGGLG126
PrP(106-126)ACO
KTNMKHMAGAAAAG(1-13
C)A120
VVGGLG
PrP(106-126)GCO
KTNMKHMA(1-13
C)G114
AAAAGAVVGGLG
Table 2-1 - Amino acid sequence and isotope labelling schemes for PrP(106-126) peptides.
13C,
15N labelled amino acids were incorporated at the sites indicated in bold, with uniform isotope labelling, except
for sites with selective incorporation of 13
C as indicated in parentheses.
To form fibrils, 25 mg of peptide was dissolved in 1ml of 1,1,1,3,3,3-hexafluoroisopropanol
(HFIP, Fluka), vortexed and sonicated for 5min, then allowed to stand at room temperature for
10 min. HFIP was evaporated under a stream of N2(g), and the resulting peptide film was
resuspended in 5 ml of 20 mM Tris buffer, pH 8.0, vortexed to mix and briefly sonicated. In
order to promote complete and rapid fibril formation, the reaction was self-seeded. 100 μL of the
initial fibril solution was removed, briefly probe-sonicated and returned. Fibrillization reactions
were allowed to stand at room temperature until a noticeable change in viscosity was achieved,
32
as indicated by the formation of a gel-like suspension, usually occurring after 5-7 days. Samples
to be analyzed by solid state NMR were dialyzed against 4L of Millipore water overnight with a
3500 Da cutoff membrane (Spectrapore), to remove residual peptide monomers, and
subsequently lyophilized.
2.3.2 Circular Dichroism Spectroscopy
The secondary structure of PrP(106-126) was analyzed by circular dichroism (CD) using a Jasco
J-810 spectropolarimeter and a quartz cuvette with a 1.0 mm path length. The CD spectrum of
monomeric peptide was obtained using 0.5 mg/ml peptide in HFIP. Fibrillar samples were
briefly probe sonicated prior to measurement, in order to reduce light scattering. Reported
spectra are the sum of 3 wavelength scans from 190-250 nm, recorded at 100 nm/min.
2.3.3 Thioflavin-T Fluorescence
Binding of ThT to PrP(106-126) fibrils was assayed by adding freshly prepared 20 μM ThT
(Sigma) to a solution of 3, 30 or 60 μM fibrils. Fluorescence was monitored using a Photon
Technology International (PTI) C60 spectrofluorimeter with excitation and emission slit widths
set to 2 and 5 nm respectively. Spectra were obtained by scanning the fluorescence emission
from 430 to 600 nm, with excitation at 442 nm. Experimental spectra were compared to a control
spectrum obtained using a solution of 20 μM ThT in fibrillization buffer (20 mM Tris, pH 8.0).
2.3.4 Transmission Electron Microscopy
Samples for electron microscopy were deposited on fresh continuous carbon films prepared from
copper rhodium grids (Electron Microscopy Sciences). Prior to adding samples, the grids were
charged using a glow discharger for 15 s at 30 mA negative discharge. Fibril solutions of 0.5
mg/ml were adsorbed to grids for 2 minutes prior to rinsing with 10 μL water for 10s. Samples
were blotted using No. 2 Whatman filter paper and stained with freshly filtered 2% uranyl
acetate for 15 s. TEM images were obtained using a Jeol 1011 microscope operating at 80 kV.
2.3.5 Atomic Force Microscopy
A 0.5mg/ml solution of fibrils was adsorbed to a freshly cleaved mica surface on a solid support,
blotted to remove excess material, and air dried. Fibrils were analyzed using a Nanoscope IIIa
Multimode scanning probe microscope (Digital Instruments/Veeco) operating in tapping mode to
33
acquire images with an area of 1, 2 and 5 m2. Height profiles along the length of individual
fibrils were obtained, as well as a series of profiles taken perpendicular to the fibril long axis. An
average height for PrP(106-126) fibrils was calculated from this data, with error reported as
standard deviation from the mean.
2.3.6 Solid State Nuclear Magnetic Resonance
Lyophilized fibrils were packed into standard 22 l 3.2 mm MAS rotors. For experiments on
hydrated fibrils, an equal mass of water was added to the dry sample in the rotor, followed by
centrifugation and incubation for at least 1 hour to ensure uniform incorporation of water. Solid
state NMR measurements were carried out on a narrow bore Varian VNMRS spectrometer,
operating at a 1H frequency of 499.82 MHz. All experiments were carried out using Varian
triple-resonance 3.2 mm T3 MAS and BioMAS probes. Sample heating in standard T3 probes
was alleviated by delivering high-flow rates of ambient temperature dry air to the sample. All
spectra were externally referenced to the downfield 13
C resonance of adamantane at 38.56 ppm
relative to TMS (Morcombe and Zilm 2003).
13C and
15N cross polarization was implemented using a linear ramped radio frequency (rf) field
centered around 40-60 kHz on the low channel, with a 50-80 kHz field on the 1H channel and
contact times of 1-1.5 ms. /2 pulse widths were typically 2.2-3 s for all channels on the T3
MAS probe, or 2.5 s (1H) and 5.5 s (
13C,
15N) on the BioMAS probe.
1H decoupling fields of
110 kHz were applied during all t1 and t2 periods, using the TPPM decoupling scheme (Bennett
et al. 1995). In all cases, a 2 s delay was used between scans.
Two-dimensional (2D) 13
C-13
C NMR spectra were obtained using a radio frequency assisted
diffusion (RAD) recoupling sequence (Takegoshi et al. 2001; Morcombe et al. 2004). RAD
mixing times of 10, 250 and 500 ms were used, at an MAS frequency of 10 kHz. 2D 15
N-13
C
correlation spectra were obtained using a double cross polarization pulse sequence in which 15
N
to 13
C cross polarization after the t1 period was achieved with rf fields of 40 to 60 kHz on each
channel and a linear ramp on the 15
N channel with a contact time of 1.5 ms. 200 points were
taken in t1 with a dwell of 25 μs with a total of 512 scans per FID.
Constant time 13
C recoupling experiments were performed using the PITHIRDS pulse sequence
as described by Tycko (Tycko 2007), using with k1 equal to 4, and k2 and k3 decremented and
34
incremented from 23 to 0 and 0 to 23 respectively, giving 50.4ms of total dipolar recoupling.
PITHIRDS spectra were obtained at an MAS rate of 20 kHz, such that the 16.67 s pulses
were used during the recoupling period. Each spectrum was taken at a sweep width of 20161 Hz
with 1600 scans per FID. The PITHIRDS pulse sequence was tested using carboxylate labelled
L-alanine, giving identical dephasing curves at various recoupling times - similar to the curves
reported by Tycko (Tycko 2007).
Rotational resonance (RR) experiments (Raleigh et al. 1988) were performed at the n = 1 RR
condition, using a 1-4 ms Gaussian pulse for selective inversion of the C1 peak. For each RR
time point (RR) in the experimental S1 curves, a reference signal (S0) was also recorded, without
inversion of the C1 signal. After obtaining a difference spectrum (S0 – S1) for each time point,
polarization transfer was calculated as the difference in peak areas (C1 - C2), normalized to the
C1 peak area at RR = 0. For accurate fits of RR polarization transfer curves, 13
C zero-quantum
coherence relaxation time (T2ZQ
) values for each spin pair C1 and C2 were estimated from
single-quantum T2 values using the equation (T2ZQ
)-1
= (T2C1
)-1
+ (T2C2
)-1
. Spin echo spectra for
T2 measurements were recorded under similar conditions of 1H decoupling and MAS rate to the
RR experiments.
2.3.7 NMR Data Analysis
All 2D spectra were processed in NMRPipe and visualized using NMRDraw (Delaglio et al.
1995), while 1D spectra were processed using VNMRJ (Varian Inc.). The number of points used
for Fourier transformation in each dimension was doubled by zero filling, and an exponential
line broadening function of 50-150 Hz was typically applied to each FID. TALOS (Cornilescu et
al. 1999) was used to obtain predicted and backbone torsion angles for PrP(106-126) in
amyloid fibrils, based on 13
C and 15
N chemical shift data.
Simulations of NMR data were carried out using Spinevolution (Veshtort and Griffin 2006).
PITHIRDS dephasing curves were calculated using a linear arrangement of five equidistant 13
C
atoms and an explicit treatment of the pulse sequence, observing only the central spin. Prior to
fitting, the natural abundance contribution at = 0 ms was subtracted from all experimental data.
This was calculated as 16.5% of the initial peak area for A120 (due to overlapping contributions
of 15 natural abundance carbonyl shifts), 4.4% of the initial peak area for G114 and G124 (4
35
additional glycine residues), and 11% for A115 (overlap from alanine and valine methyl
carbons). RMS noise was calculated from baseline regions of each spectrum, using integral
widths identical to those for the peak of interest, and was normalized against the area of the first
peak. Experimental PITHIRDS data were fit to simulations using the weighted sum of squared
residuals (equation 2-1)
wyyS 22 ˆ Equation 2-1
where y is the value of the experimental data point (normalized intensity), ŷ is the corresponding
simulated data point (normalized intensity) and w is a weighting factor; in this case w
corresponds to the span of the RMS noise.
RR polarization transfer was simulated using 4 spins chosen to represent the C1 and C2 nuclei
and one directly bonded 13
C spin for each nucleus. There is a negligible dependence of the RR
transfer on CSA values and relative tensor orientations under our sample conditions (finite T2ZQ
values and internuclear distances greater than 2.5 Å), such that only the C1-C2 distance (r12) and
T2ZQ
values are required for fitting. Both variables were explicitly included in the Spinevolution
calculations.
2.3.8 PrP(106-126) Fibril Modeling
An initial model of PrP(106-126) fibrils was built in MOLMOL (Koradi et al. 1996) as an ideal
parallel in-register -sheet composed of 10 strands. Two sheets were stacked in an antiparallel
arrangement using close intersheet contacts consistent with 13
C RAD and RR data. Multiple
rounds of restrained energy minimization and molecular dynamics (MD) were carried out using
Tinker/Forcefield explorer (available at http://dasher.wustl.edu/tinker/), resulting in an ensemble
of 10 structures. The CHARMM27 forcefield was used for all calculations. Structural constraints
on internuclear distances and backbone torsion angles made use of harmonic potential energy
functions to restrain models as described below. An initial round of energy minimization was
performed with torsion angles from residues 115-123 of = -120° and = 113°, representing an
ideal parallel β-sheet conformation. Subsequently, torsion angles were restrained to the values
obtained from TALOS for a second round of energy minimization and for restrained molecular
dynamics. Hydrogen bonding between strands was enforced as 2.15 Å distances between
backbone carbonyl oxygens and amide hydrogens from residues 115-123. Intersheet contacts
36
obtained from 13
C spin diffusion experiments were defined as < 8 Å distances between pairs of
atoms in G114/V122, A116/V122 and A113/G123. The stability of the final model obtained was
confirmed using unrestrained molecular dynamics simulations. Key interstrand contacts were
preserved for longer than 100 ps at temperatures of up to 400 K, and the fibril structure remained
intact. Some global distortions were observed at higher temperatures, likely due to the artificially
short fibril being simulated.
37
2.4 Results
2.4.1 PrP(106-126) Forms Amyloid Fibrils with Characteristics of a Cross-β Structure
Fibrillar aggregates formed by PrP(106-126) were visualized by TEM and AFM, with
representative images shown in Figure 2-1. Negative stain TEM images (Figure 2-1A) revealed
straight, untwisted, unbranched fibrils 0.5–1 m in length and 5-7 nm in width. The morphology
of PrP(106-126) fibrils was uniform throughout all samples imaged, and is consistent with
previous reports of amyloid fibrils formed by this peptide (Forloni et al. 1993; Selvaggini et al.
1993; Salmona et al. 1999). AFM measurements (Figure 2-1B) revealed a height of 2.4 ±0.4 nm
for all fibrils measured. In comparison with the 4-5 nm heights measured for fibrils formed by
amylin and -amyloid (Petkova et al. 2002; Luca et al. 2007), each of which contains a core of 4
stacked -sheets, we expect PrP(106-126) fibrils to be composed of 2 sheets.
CD spectra obtained for suspensions of PrP(106-126) fibrils are shown in Figure 2-1C, and
exhibit a minimum at 216 nm, characteristic of -sheet secondary structure. This is in contrast to
spectra of PrP(106-126) in HFIP, which indicate that the peptide is unstructured prior to fibril
formation. Thioflavin-T (ThT) binding was used to confirm the amyloid nature of fibrils formed
by our PrP(106-126) peptides. Binding of this dye to polypeptide chains is specific for the cross-
structure of amyloid fibrils. As a function of the amount of fibrils added to the solution, an
increase in ThT fluorescence emission at 480 - 490 nm is observed, as shown in Figure 2-1D,
supporting a cross- structure for PrP(106-126) fibrils.
38
Figure 2-1 - Ultrastructural characterization of PrP(106-126) fibrils.
(A) TEM images of PrP(106-126) fibrils stained with uranyl acetate, obtained at 60,000x magnification. The inset
shows the details of an individual fibril, viewed at 250,000x magnification. (B) An AFM image of PrP(106-126)
fibrils, obtained using tapping mode in air. (C) Circular dichroism spectra of PrP(106-126) dissolved in HFIP
(dashed line), and of PrP(106-126) fibrils suspended in pH 8.0 Tris buffer (solid line). (D) Fluorescence emission
spectra obtained for 20 μM ThT in the presence or absence of PrP(106-126) fibrils at the concentrations indicated.
39
2.4.2 13C and 15N Chemical Shifts Reveal an Extended β-Sheet, Spanning Residues 113-123
Lyophilized PrP(106-126) fibrils, labelled with 13
C and 15
N as indicated in Table 2-1, were
studied by MAS NMR. 1D 13
C and 15
N spectra of PrP(106-126)GAVL
are shown in Figures 2-2A
and 2-2B, respectively. 13
C chemical shift assignments were made directly from 1D spectra in
cases where spectral overlap was minimal or non-existent (PrP(106-126)AAGG
, PrP(106-126)ACO
and PrP(106-126)GCO
fibrils). Otherwise, assignments were made by identifying the amino-acid
specific spin systems in 2D 13
C-13
C chemical shift correlation spectra. An example of 13
C
chemical shift assignment is shown in Figures 2-2C and 2-2D, with the readily identified spin
systems labelled in each case. 15
N chemical shifts were assigned using the N-C cross peaks in
15N-
13C heteronuclear correlation spectra. The same approach was used to assign all
13C and
15N
resonances in fibrils of PrP(106-126)AVG
and PrP(106-126)AVG2
(see appendix , Figures A1 and
A2). Chemical shifts for all labelled sites are provided in appendix Table A-1.
13C secondary chemical shifts for all carbonyl, and carbons were calculated as the difference
between the observed chemical shifts and those reported for unstructured peptides in solution
(Wishart and Sykes 1994). Due to the dependence of backbone chemical shift values on local
structure, these values can be used to define secondary structure elements in proteins (Saito
1986; Wishart and Sykes 1994). Results of this analysis are shown in Figure 2-3A, with residues
113-125 exhibiting upfield shifts of carbonyl and carbons, and downfield shifts of carbons,
characteristic of an extended -structure. Gly126 exhibits random coil chemical shift values,
with broad CO and C resonances indicative of significant structural heterogeneity. Backbone
chemical shift data were used to predict backbone and torsion angles for residues 113-125,
using TALOS (Cornilescu et al. 1999). These are listed in appendix Table A-2, and support the
presence of an extended -strand in the hydrophobic segment of PrP(106-126).
40
Figure 2-2 - MAS NMR spectra of PrP(106-126)GAVL
fibrils.
(A) 13
C cross polarization spectrum of lyophilized fibrils, obtained at an MAS spinning frequency of 10 kHz.
Spinning sidebands are indicated by asterisks. (B) 15
N cross polarization spectrum of the same sample, at an MAS
frequency of 10 kHz. (C) 2D 13
C-13
C correlation spectrum obtained at 10 kHz MAS, using a 10 ms RAD mixing
period. In the direct dimension, 1024 complex points were acquired in the direct dimension, with 200 increments at
a dwell time of 25 μs acquired in t1. A 10 kHz 1H field was applied during the RAD period and 64 scans were taken
per FID. (D) Expanded view of the 2D spectrum in (C), showing the details of the aliphatic region. In (C) and (D),
the resonance assignments for each amino acid are shown.
41
Figure 2-3 - Secondary 13
C chemical shifts and line widths measured for PrP(106-126) fibrils.
(A) The differences in 13
C chemical shifts from the random coil values reported by Wishart et al. are shown for each
CO, C, and C resonance assigned in this study, measured using 13
C-13
C 2D spectra obtained with a RAD mixing
period of 10ms, or using 1D 13
C spectra, where appropriate. (B) 13
C line widths for each assigned CO, C, and C
resonance.
42
13C NMR line widths were measured for PrP(106-126) fibrils, and are reported in Figure 2-3B.
The line widths observed for the majority of sites in the palindromic AGAAAAGA region of
PrP(106-126) fall into the range of 1.5-2.5 ppm, suggesting a well-ordered protein structure
(Petkova et al. 2002; Sharpe et al. 2004), with the exception of the A113 side chain. Line widths
for V122 are also ~2-2.5 ppm, while significantly broader 13
C lines were observed for residues
V121, G123, L125 and G126 (3.5-4.5 ppm for some sites), suggesting reduced structural order or
multiple conformations. Side chain atoms exhibiting significant line broadening relative to the
backbone are likely exposed on the surface of the fibril, as opposed to forming well-defined
interactions in the core.
Upon hydration of PrP(106-126) fibrils, the 13
C line widths for A113 and G123 narrow by 0.5-1
ppm, while L125 and G126 lines are up to 2 ppm narrower Figure 2-4). This suggests an
increase in local order upon hydration, such that all residues exhibit line widths from 1.5-2.25
ppm, and is consistent with an extended -structure encompassing these residues. Overall, our
chemical shift and line width data are consistent with H/D exchange rates reported for similar
fibrils (Kuwata et al. 2003), and support a model in which the hydrophobic residues 114-124
form the core of PrP(106-126) fibrils. One 13
C labelled site showed evidence of structural
polymorphism, giving rise to two distinct resonances. In 10 ms RAD spectra, two resonances are
observed for the methyl carbon of A118, at a 3:1 ratio. Lee et al. have reported a 1:1 ratio of two
methyl resonances for all alanine residues in PrP(109-122) fibrils, likely due to the front to back
packing of sheets indicated by their data (Lee et al. 2008). In PrP(106-126) fibrils, only A118
shows a doubling of the methyl peak, and the 3:1 ratio does not support two equally populated
sites. At 250 and 500 ms RAD mixing, exchange is observed between the two A118 methyl
resonances, suggesting close proximity of methyl groups with different packing interactions in
the same fibril. Since only this site gives rise to two distinct peaks, it is unlikely that a second
fibril morphology exists in our PrP(106-126) fibrils.
43
Figure 2-4 - Differences in 13
C NMR line widths between dry and hydrated fibrils of PrP(106-126).
The difference in 13
C line width observed for lyophilized and hydrated fibrils is plotted for the carbonyl, C and C
atoms of each labelled amino acid. Positive values indicate an increase in measured line width for dry fibrils
relative to hydrated fibrils. These results suggest an increase in structural order (resulting in a decrease in line
width) for sites at the ends of the hydrophobic core when hydrated. In dry fibrils, these sites are relatively
disordered, giving rise to significantly broadened lines. Residues 114-122 within the fibril core remain relatively
unaffected by hydration state, as expected based on their location within the dehydrated core of the amyloid fibril.
44
2.4.3 Amyloid Fibrils of PrP(106-126) are Composed of In-Register Parallel β-Sheets
The constant-time PITHIRDS experiment (Tycko 2007) was used to measure the average
distance between a single 13
C atom on one PrP(106-126) peptide and the same site on adjacent
peptides within the fibril. As a function of the recoupling period, typically 0 to 60 ms, the
magnetization will dephase at a rate proportional to the 13
C-13
C dipolar coupling constant. The
dipolar coupling between two nuclei is in turn proportional to 1/r3, where r is the internuclear
distance. This has been shown to be highly sensitive to interstrand packing with 4.8 Å distances
typically observed between backbone atoms in adjacent strands of an in-register parallel -sheet.
The integrated 13
C NMR peak intensities obtained as a function of PITHIRDS recoupling time,
for PrP(106-126)GCO
, PrP(106-126)ACO
, the C of G124, and the C of A115 are shown in
Figure 2-5. These are plotted along with dephasing curves calculated using Spinevolution, and
are best fit to 6.6, 6.2, 5.6 and 5.6 Å (+/- 0.25 Å) internuclear distances for samples labelled at
G114, A120, G124 and A115 respectively. For two of the three sites, the observed distances are
within the 5.5-6.0 Å range observed for parallel -sheets in fibrils formed by -amyloid (Petkova
et al. 2005; Chimon et al. 2007) and amylin (Luca et al. 2007). The slower decay of 13
C
recoupling data relative to that expected for an ideal -sheet has previously been attributed to the
effects of transverse relaxation (Balbach et al. 2002), and might also be symptomatic of minor
structural heterogeneities within PrP(106-126) fibrils. These data provide strong evidence in
favour of an in-register parallel -sheet arrangement in PrP(106-126) fibrils. The longer (6.6 Å)
interstrand distance observed at G114 may indicate increased disorder of the sheet at the N-
terminus, in keeping with decreased H/D exchange protection factors at this site. This may be a
result of repulsion between the positive charges of K106 and K110. Despite the slightly longer
distance between G114 residues, the A120, G124 and A115 data rule out an antiparallel
arrangement of strands within the PrP(106-126) fibrils. In addition, the short distance measured
between the A115 methyl carbons strongly suggests the presence of an in-register -sheet, since
a shift in registry would significantly increase the separation of A115 side chains in adjacent
strands.
45
Figure 2-5 - PITHIRDS recoupling curves for PrP(106-126) fibrils.
Dipolar dephasing curves obtained using the PITHIRDS homonuclear recoupling scheme are shown for the carbonyl
13C resonance in fibrils formed by PrP(106-126)
GCO (blue squares), PrP(106-126)
ACO (green), the A115 C
resonance of PrP(106-126)AAGG
fibrils (magenta), and the G124 C resonance of PrP(106-126)AAGG
fibrils (red).
Simulated curves corresponding to interatomic distances from 5.4 to 7.0Å shown in 0.2 Å increments with the most
dephasing (lowest curve) being 5.4 Å and the least dephasing (highest curve) at 7.0Å. Error bars for the PrP(106-
126)GCO
and A115 C data are smaller than the points in the graph.
46
2.4.4 Quaternary Structure of PrP(106-126) Fibrils from 13C Spin Diffusion and Rotational Resonance Experiments
13C-
13C correlation spectra recorded using long (250 - 500 ms) RAD spin diffusion periods allow
cross peaks to be observed between nuclei up to 6 Å apart (Petkova et al. 2006). Note that
transfer between nuclei with longer internuclear distances may be observed due to sequential
transfer during the spin diffusion period. In Figure 2-6, a series of 13
C spin diffusion spectra
obtained with 10, 250 and 500 ms RAD mixing are shown for PrP(106-126)GAVL
fibrils. As the
mixing time is increased, cross peaks resulting from long-range interresidue contacts between
V122 and G114, as well as V122 and A116 (Figure 2-6D, E, F) are observed.
Figure 2-6 - Long-range 13
C-13
C internuclear contacts observed in 2D 13
C-13
C NMR spectra of PrP(106-
126)GAVL
fibrils.
(A) 13
C-13
C correlation spectra obtained with a RAD mixing time of 10 ms. Similar spectra obtained with mixing
times of 250 and 500 ms are shown in (B) and (C), respectively. Horizontal slices at the G114C , A114C, and
V122C frequencies are shown for 10 ms (D), 250 ms (E) and 500 ms (F) mixing times. Interresidue cross peaks are
indicated on the slices, and discussed in the text.
Due to the distance between these residues in an extended -strand, these peaks can only arise
from proximity of these residues in two closely packed sheets. At 250 ms all G114 or A116 to
V122 contacts are present with the exception of A116Cα to V122Cα, which is only present at
500 ms mixing. These data strongly suggest an arrangement of sheets within the fibrils such that
V122 is packed between G114 and A116. This is supported by similar experiments carried out
47
on PrP(106-126)AVG
fibrils (Figure 2-7), in which strong cross peaks between the side chain and
backbone 13
C atoms of A113 and G123 are observed at both 250 and 500 ms RAD mixing. No
long-range contacts were observed at up to 500 ms RAD mixing for either the AVG2 or AAGG
samples.
Figure 2-7 - Long-range 13
C-13
C internuclear contacts observed in 2D 13
C-13
C NMR spectra of PrP(106-
126)AVG
fibrils.
(A) 13
C-13
C correlation spectra obtained with a RAD mixing time of 10 ms. Similar spectra obtained with mixing
times of 250 and 500 ms are shown in (B) and (C), respectively. Horizontal slices at the A113Cβ , A113C, and
G123CO frequencies are shown for 10 ms (D), 250 ms (E) and 500 ms (F) mixing times. Interresidue cross peaks
are indicated on the slices.
To further characterize these intersheet packing interactions, 13
C rotational resonance (RR)
experiments were carried out on the PrP(106-126)GAVL
fibrils. The results of RR polarization
transfer between the carbonyl carbon of G114 and the C, C and Cγ atoms of V122 are shown
in Figure 2-8. When fit to simulated curves, the distance between the G114CO and the C or C
of V122 were found to be approximately 4.5 and 5.0 Å, respectively. While the two methyl
resonances of V122 cannot be independently resolved, the fit of the dephasing curve to 4.3-4.5 Å
indicates close proximity of the G114CO to at least one of the V122 methyl groups. Shorter
distances observed between the G114CO and V122C, relative to V122C, support an
arrangement in which the valine side chain is packed against the backbone of A115, between the
G114 and A116 side chains.
48
Figure 2-8 - Experimental 13
C rotational resonance data for PrP(106-126)GAVL
fibrils and simulated
polarization transfer curves.
The differences in NMR peak areas (C1 – C2) are plotted for each polarization transfer experiment as a function of
RR, along with curves simulated using r12 distances of 4, 4.5, 5 and 5.5 Ǻ and T2ZQ
values of 1.2 ms (solid lines) or
1.0 and 1.4 ms (upper and lower dashed lines, respectively). The T2ZQ
values obtained from single quantum T2
relaxation times for experimental C1, C2 spin pairs are 1.28 0.17 ms (G114CO, V122C), 1.05 0.17 ms
(G114CO, V122C), and 1.07 0.15 ms (G114CO, V122C).
49
2.4.5 Structural Model of PrP(106-126) Fibrils Based on Solid State NMR Measurements
An initial model for PrP(106-126) fibrils based on the intermolecular contacts identified by solid
state NMR is depicted in Figures 2-9A and 2-9B. Figure 2-9A shows the parallel arrangement of
strands dictated by the PITHIRDS experiments, while Figure 2-9B shows the close intersheet
packing consistent with 13
C spin diffusion and rotational resonance data. These measurements,
and the absence of other close contacts in our fibril samples, allow unambiguous alignment of
PrP(106-126) in parallel -sheets, which are stacked in an antiparallel manner to form the fibril.
and backbone torsion angles obtained from TALOS were also included in the restraints used
during energy minimization and MD simulations.
A representative minimized structure for a PrP(106-126) fibril composed of 10 peptide chains in
each of two stacked sheets is shown in Figure 2-9C. The N-terminal residues from 106-112
show significant disorder, although the salt bridge formed between K110 and the C-terminal
carboxyl group of the opposing strand restricts the movement of this segment. Residues 114-123
form the tightly packed core of the fibril, and remain in an in-register parallel -sheet after
energy minimization. A slight twist is evident in the minimized fibril structure, as expected for
an extended -structure. In Figure 2-9D, the ensemble of structures obtained for 2 strands in the
middle of the fibril model are shown, along with an average structure. Hydrophobic residues
forming the core of the fibril are tightly interdigitated with those from the opposing sheet. At
each end the M112 side chain is accommodated by a slight deformation/bulge at G124/L125,
with L125 on the exterior of the fibril. Surface exposure of L125 is supported by the decrease in
13C line widths at this site upon hydration of dry fibrils. The register of opposing sheets in this
model places the C-terminus in a good position to form a salt bridge with K110, as seen in the
minimized structures.
50
Figure 2-9 - Structural models of PrP(106-126) fibrils.
(A) Schematic of a single parallel sheet formed by PrP(106-126), with residues G114, A120, and G124 shown in
blue, green, and red respectively. (B) Ball and stick model showing the key intersheet packing interactions
consistent with 13
C spin diffusion and RR NMR experiments. (C) A typical structural model obtained after energy
minimization and restrained molecular dynamics of PrP(106-126) fibrils. The two parallel -sheets are shown in
orange and blue, with the disordered N- and C-termini shown in grey. (D) Cross section of the fibril shown in (C),
highlighting the packing interactions between two peptides in opposing -sheets. The upper panel shows 12 overlaid
structures for each strand, as obtained from multiple rounds of MD and minimization. The lower panel shows a
single average structure (calculated by MOLMOL), with selected residues from the lower strand labelled as a guide.
Amino acids are color coded by type, with green, blue, red and magenta indicating hydrophobic, positively charged,
negatively charged (here the C-terminal glycine) and polar side chains, respectively.
51
2.5 Discussion
The peptide comprising residues 106-126 of mammalian PrP has been extensively studied as
both a neurotoxic amyloid peptide, and as a potential mediator of PrP conversion to the scrapie
form. In particular, the highly conserved palindromic sequence AGAAAAGA has been
implicated in the assembly of amyloid fibrils and neurotoxicity of PrP(106-126) (Jobling et al.
1999; Lee et al. 2008). In the present work we have used solid state NMR to examine the
molecular structure of amyloid fibrils formed by PrP(106-126). Combined with TEM and AFM
measurements, we have produced an experimentally constrained structural model for the tightly
packed core of these fibrils. The overall morphology of our fibrils is consistent with the
observations of other groups (Forloni et al. 1993; Selvaggini et al. 1993; Salmona et al. 1999).
After restrained energy minimization and molecular dynamics, the structure obtained for the
fibril core is consistent with the formation of a class 1 steric zipper motif identified by Sawaya et
al., and closely resembles the quaternary packing observed for fibrils of the -amyloid protein
(Petkova et al. 2006). The extended hydrophobic segment spanning residues G114 to G123 is
involved in an extensive and tightly interdigitated interface between the two parallel -sheets,
including the V122 to G114/A116 contacts observed in 13
C spin diffusion experiments. Two
other factors which may stabilize the fibril structure can be identified in the structural model.
One is the presence of a salt bridge between the C-terminus and the K110 side chain which may
explain the effects of C-terminal amidation on PrP(106-126) fibril formation, as discussed below.
The second is the potential for the formation of an extended chain of - interactions between
the side chains of H111, which are closely aligned along the fibril long axis.
The structural model presented here provides a rational basis for understanding the effects of
sequence alterations on PrP(106-126) assembly. For instance, altering the polarity of the
palindromic sequence by replacing the alanine residues with serine dramatically reduces
fibrillization, likely due to reduced stability of the hydrophobic core (Jobling et al. 1999).
Likewise G114A / G119A double mutants form fibrils much less readily (Florio et al. 2003). In
our model, a G114A mutation would add a methyl group to the pocket already occupied by V122
from the opposing sheet, potentially inhibiting packing. A related effect may account for
observations that methionine oxidation also reduces aggregation and alters the morphology of the
52
resulting fibrils (Bergstrom et al. 2007). Certainly increasing the polarity of the tightly packed
M112 side chain would destabilize the hydrophobic interface in our fibrils. The A117V mutation
observed in some GSS patients actually enhances fibril formation, which is consistent with its
location on the exterior of our fibrils, rather than in the core. Similarly, it has been observed that
PrP(106-126) forms fibrils more readily at pH 7 than at pH 5, an effect most readily explained by
a change in the protonation state of H111 (Salmona et al. 1999). Due to the close proximity of
H111 side chains in our fibrils, a positive charge in this position would likely have a strong
destabilizing effect.
It is important to note that in several of the cases mentioned above, there is not only a reduced
fibril formation but an apparent change in fibril morphology. It is evident from examining the
sequence of PrP(106-126) that the extended hydrophobic interface can accommodate more than
one favourable mode of packing. Thus, subtle mutations may promote the formation of non-wild
type interactions. A similar phenomenon may account for the different structural models derived
for fibrils formed by three closely related PrP peptide sequences.
In particular, the molecular dynamics simulations performed by Kuwata et al. (Kuwata et al.
2003) were performed using an amidated form of the mouse PrP(106-126), corresponding to
human PrP residues 107-127, and containing valine residues in place of M109 and M112. While
our structure is consistent with their reported H/D exchange data, their simulations suggested a
parallel stacking of parallel -sheets to be the energetically preferred structure. It is likely that
amidation of the C-terminus, which removes the charge pairing with K110 in our model, would
significantly alter peptide packing. By contrast, our model matches a previous proposal in which
the presence of a salt bridge at this position was suggested (Salmona et al. 1999). In fact, it has
been noted by several groups that amidation of PrP(106-126) inhibits fibril formation, and
significantly alters the mechanism of peptide toxicity in vivo (Salmona et al. 1999; Bergstrom et
al. 2007).
Recent solid state NMR and FT-IR studies of fibrils formed by an amidated and acetylated
PrP(109-122) peptide have presented strong evidence for a class 6, 7 or 8 steric zipper, in which
antiparallel -sheets stack in a face to back manner to form the fibril (Silva et al. 2003; Lee et al.
2008). This peptide is significantly different from the uncapped 106-126 sequence in our work,
and is likely to exhibit different packing arrangement in amyloid fibrils. As noted above,
53
amidation and other modifications appear to dramatically alter the physical and biological
properties of this peptide, such that understanding the factors driving the formation of these
different structures will be of significant value. Similar switches between parallel and antiparallel
-sheets have been observed in peptides based on the Alzheimer’s -amyloid protein, and appear
to stem from changes in alignment of polar residues and different hydrophobic packing in
peptides of slightly differing length and/or sequence (Tycko et al. 2006).
It has been suggested that PrP(106-126) fibrils may have direct relevance to prion biology. In
most cases, toxicity of PrP(106-126) requires expression of PrP in the target cells (Thellung et
al., 2000; Pietri 2006; Forloni et al., 1993; Gong et al., 2007; Ettaiche et al 2000). Brown has
suggested a direct interaction with PrPC
(Brown 2000), while Gu et al. demonstrated that
PrP(106-126) can catalyze formation of protease resistant PrP in neuroblastoma cells (Gu et al.
2002). Along similar lines, deletion of G114-A120 is protective against PrPSc
(Holscher et al.
1998). It is tempting, therefore, to suggest that the structures of aggregated forms of this peptide,
such as the fibril model presented here, may share common elements with PrPSc
, or may reveal
insight into the mechanism of PrP conversion. This concept finds strong support in a recent study
in which antibodies against aggregated, non-amidated PrP(106-126) are able to selectively bind
to PrPSc
versus PrPC
(Jones et al. 2008).
In conclusion, we have produced the first experimentally constrained structural model for
amyloid fibrils formed by PrP(106-126). Our data support the presence of a class 1 steric zipper
in the hydrophobic core of these fibrils, giving rise to favourable van der Waals interactions
within the sheet-sheet interface which are supplemented by the presence of a salt bridge between
the C-terminus and the side chain of Lys110. Our model is consistent with the effect of amino
acid substitutions within this sequence on peptide aggregation, and accounts for the reported
inability of PrP(106-126) peptides with amidated C-termini to form fibrils. While the putative
roles of PrP(106-126) in toxicity and in PrP conversion remain undefined, increased knowledge
regarding the aggregated states of this peptide represents an important step towards
understanding these processes.
54
3 Morphology and Secondary Structure of Stable β-Oligomers Formed by Amyloid Peptide PrP(106-126)
Selections from this chapter were originally published in the journal Biochemistry. Walsh, P.,
Yau, J., and Sharpe, S. Morphology and Secondary Structure of Stable β-oligomers Formed by
PrP(106-126). Biochemistry. 2009. 48(25). pp5779-5781. Reprinted with perimission. Copyright
American Chemical Society 2009.
All experiments were carried out by P.Walsh.
55
3.1 Abstract
The formation of non-fibrillar oligomers has been proposed to be a common element of the
aggregation pathway of amyloid peptides. Here we describe the first detailed investigation of the
morphology and secondary structure of stable oligomers formed by a peptide comprising
residues 106-126 of the human prion protein (PrP). These oligomers have an apparent
hydrodynamic radius of approximately 30 nm, and are more membrane-active than monomeric
or fibrillar PrP(106-126). Circular dichroism and solid state NMR data support formation of an
extended β-strand by the hydrophobic core of PrP(106-126), while negative thioflavin-T binding
implies an absence of cross-β structure in non-fibrillar oligomers.
56
3.2 Introduction
The formation of fibrillar protein aggregates characterizes many neurodegenerative diseases,
including Alzheimer’s, Parkinson’s, and mammalian prion diseases. While the accumulation of
amyloid fibrils, and their deposition in plaques, has long been associated with cell death and
disease progression, recent evidence suggests that it is more likely that non-fibrillar protein
oligomers are the cytotoxic species (Bucciantini et al. 2002; Kayed et al. 2003). Soluble
oligomers of the Alzheimer’s β-amyloid protein (Aβ) have been observed in vitro and in vivo,
and induce neuronal cell death more strongly than fibrillar or monomeric protein (Walsh et al.
2002; Haas and Selkow 2007). Similar oligomers formed by several amyloid proteins have been
produced in vitro and share several key features, including a common morphology, the ability to
permeabilize model membranes, and cytotoxicity to cultured neurons (Kayed et al. 2004;
Sokolov et al. 2006; Glabe, Charles G. 2008).
Based on these observations, a general mechanism for amyloid toxicity has been proposed in
which cell death results from the accumulation of non-fibrillar aggregates formed during the
early stages of protein misfolding prior to the appearance of amyloid fibrils. However, despite
their potential importance in the pathogenesis of amyloid diseases, relatively little is known
about the molecular structure of non-fibrillar oligomers. In particular, no structural data have
been reported for those formed by peptides other than Aβ.
The structure of amyloid fibrils formed by several proteins has been described in detail, based
primarily on high-resolution solid state nuclear magnetic resonance (NMR) studies (Jaroniec et
al. 2004; Petkova et al. 2006; Luca et al. 2007). X-ray crystallography has also revealed the
details of fibril structure for a number of short amyloidogenic peptides (Sawaya et al. 2007). In
all cases, these structures share a common cross-β architecture, in which β-strands run
perpendicular to the long axis of the fibril.
The best characterized non-fibrillar oligomers are those formed by Aβ, for which spherical
aggregates ranging from 5-35 nm in diameter have been reported, as well as 20-200 μm diameter
“β-amy balls” (Walsh and Selkoe 2007; Glabe, C. 2008). While some of these oligomers have
been shown to contain β-sheet structure, their relationship to amyloid fibrils has not clearly been
determined. In particular, it remains unclear if the spherical oligomers represent intermediate
57
stages on the pathway to fibril formation, or if they are the products of alternate misfolding
pathways. Additionally, amyloid oligomers from an 11-residue segment of αB crystallin have
recently been characterized. These structures were found to form a cylindrical structure of 6
antiparallel β-strands termed a cylindrin (Laganowsky et al. 2012). The discovery of this new
amyloid structure in this model system provided the first example of a cytotoxic, barrel-like
structure that is also able to convert to amyloid fibrils.
Direct evidence for fibril-like local structure in late stage intermediates of Aβ has been obtained
from solid state NMR. Large (>650 kDa), transient oligomers were shown to share a quaternary
structure with amyloid fibrils of the same peptide (Chimon et al. 2007). More recently, solution
NMR spectroscopy has been used to define the local structure of Aβ (1-42) in the context of
small (16-64 kDa) SDS-stabilized oligomers, revealing extended β-strands forming a mixture of
fibril-like and non-fibril-like intersheet contacts (Yu et al. 2009). Furthermore, the structural
characterization of a pentamer of Aβ1-42 has been described as having a β-turn-β motif similar to
that of the mature fibril but with fewer intramolecular contacts (Ahmed et al. 2010). This model
gives a more detailed understanding of how these disc-like pentamers could undergo a transition
to mature Aβ1-42 fibrils (Ahmed et al. 2010). The ability of Aβ to form oligomers with very
different structural properties suggests that diverse non-fibrillar assemblies are accessible to
other amyloidogenic peptides.
Here we describe the morphology and secondary structure of non-fibrillar oligomers formed by
PrP(106-126), an amyloidogenic fragment of the mammalian prion protein (PrP). This peptide
forms amyloid fibrils and soluble oligomers, induces apoptosis in cultured neurons, and may
play a role in catalyzing the conversion of cellular prion (PrPC) to the scrapie form (PrP
Sc)
(Forloni et al. 1993; Kaneko et al. 1995; Gu et al. 2002). Chapter 2 describes the structure of
amyloid fibrils formed by PrP(106-126). These are composed of parallel β-sheets stacked in an
antiparallel class 1 steric zipper motif, resulting in 5-7 nm wide untwisted fibrils with the cross-β
architecture typical of amyloid. The cytotoxicity of non-fibrillar oligomers of PrP(106-126) has
been reported (Kayed et al. 2003; Kayed et al. 2004), but the details of their structure or
morphology have not.
58
3.3 Materials and Methods
3.3.1 PrP(106-126) Oligomer Formation
PrP(106-126) was obtained from the Advanced Protein Technology Centre at the Hospital for
Sick children with 13
C and 15
N amino acids obtained from Spectra Stable Isotopes and
Cambridge Isotope Laboratories. PrP(106-126) peptides were prepared by solid phase peptide
synthesis, using standard FMOC chemistry. For the purpose of NMR experiments, five different
labelling schemes were devised, with incorporation of 13
C and 15
N-labeled amino acids at
selected sites as indicated in chapter 1 (Table 2-1). The final product was purified in each case by
reverse phase HPLC, using an 11 x 300 mm C8 peptide column (Vydac). A gradient of 0 to 54%
acetonitrile (ACN) with 0.1% TFA was used, and the desired product eluted at 32% ACN and
was freeze-dried. The purity of PrP (106-126) was confirmed by MALDI-TOF mass
spectrometry.
For oligomer formation, 10mg of peptide was dissolved in 2.5 ml of 1,1,1,3,3,3-
hexafluoroisopropanol (HFIP, Fluka), vortexed and sonicated for 5 min, then allowed to stand at
room temperature for 10 min. 10 ml of 10 mM sodium acetate was then added to the peptide in
HRIP. HFIP was then removed by evaporation under a stream of N2 (g). Samples to be analyzed
by solid state NMR were dialyzed against 4L of Millipore water overnight with a 3500 Da cutoff
membrane (Spectrapore) to remove residual monomers and subsequently lyophilized.
3.3.2 Circular Dichroism Spectroscopy
The secondary structure of PrP(106-126) peptides was analyzed by circular dichroism (CD)
using a Jasco J-810 spectropolarimeter. The CD spectrum of monomeric peptide was obtained
using 0.5 mg/ml peptide in HFIP. A 0.5 mg/ml sample of oligomer in 10 mM sodium acetate
buffer was obtained. In each case, the spectrum was obtained as a scan of wavelengths from
190-250 nm, using a 1.0 mm path length cuvette.
3.3.3 Thioflavin T Fluorescence
Binding of ThT (Sigma) to PrP(106-126) fibrils and oligomers was assayed by adding 20 μM
freshly prepared ThT to 3, 30 or 60 μM peptide in 10mM acetate buffer pH 4.6 for oligomers and
20 mM Tris buffer pH 8.0 for fibrils (as described in chapter 2). Additionally, ThT to oligomers
59
was assayed in 20 mM Tris pH 8.0 with no change in binding from pH 4.6. Fluorescence
emission spectra were measured using a Photon Technology International (PTI) C-60
spectrofluorimeter with 2 nm excitation and 5 nm emission slit widths. The excitation
wavelength was 442 nm, and emission was measured from 450 to 600 nm. Background ThT
fluorescence was measured using a control sample containing only 20 μM ThT in 10mM acetate
buffer. Upon binding of ThT to proteins or peptides containing cross- structures, there is an
increase in fluorescence emission at 482 nm (Naiki et al. 1989). Unbound dye has excitation and
emission maxima at 350 nm and 450 nm, respectively.
3.3.4 Dynamic Light Scattering
Dynamic light scattering measurements were performed using a Dynapro instrument (Protein
Solutions) at 20 °C. Samples were prepared by reconstituting 3 mg of lyophilized oligomers in
10 mM acetate buffer and centrifuging at 14,000 g for 10 min, in order to remove any small
particulate. Data were collected and processed using Dynamics 5.26.38 software.
3.3.5 Transmission Electron Microscopy
Samples for single or double carbon negative stain transmission electron microscopy (TEM)
were deposited on fresh continuous carbon films prepared from copper rhodium grids (Electron
Microscopy Sciences, EMS Hatfield, PA.). Prior to adding samples, the grids were charged
using a glow discharger (EMS) for 15 s at 30 mA negative discharge. Oligomer solutions
between 0.1 and 0.004 mg/ml were allowed to adsorb to the grids for 2 minutes prior to rinsing
in a 10 μL drop of water for 10 s. Samples were blotted using No. 2 Whatman filter paper.
Single carbon TEM samples were stained with a 10 μL drop of freshly filtered 2% uranyl acetate
(EMS) for 15s before blotting excess stain. Double carbon TEM samples were prepared by
depositing a layer of carbon floating on 2% uranyl acetate on top of freshly washed samples,
such that the specimen is embedded in a layer of stain sandwiched between two carbon films.
Negative stain TEM images were obtained using a Jeol JEM-1011 microscope with an
acceleration voltage of 80 kV.
3.3.6 Atomic Force Microscopy
A 0.5 mg/ml solution of PrP(106-126) oligomers was adsorbed to a freshly cleaved mica surface
on a solid support, blotted to remove excess material, and air dried. AFM images were then
60
obtained using a Nanoscope IIIa Multimode scanning probe microscope (Digital
Instruments/Veeco) operating in tapping mode to acquire images with areas of 1, 2 and 5 μm2.
3.3.7 Liposome Dye-Release Assay
The liposome dye leakage assay was modified from (Kayalar and Duzgunes 1986), and exploits
the concentration-dependent self-quenching of the fluorescent dye calcein. When trapped in
liposomes above the threshold concentration (35 mM), calcein has a weak fluorescence emission
at 516 nm (excitation at 490 nm). Upon disruption of the liposomal membrane, calcein is
released and diluted into the bulk solution, and emission increases. Large unilamellar vesicles
(LUVs) were prepared by codissolving 50 mg of 3:1 1-palmitoyl-2-oleoyl-sn-glycero-3-
phosphocholine (POPC): 1-palmitoyl-2-pleoyl-sn-glycero-3-phosphoglycerol (POPG) in 1:1
methanol:chloroform. The lipids were dried to a thin film under N2 in a glass tube, suspended in
water, and lyophilized to remove residual solvent. Lipids were then resuspended in assay buffer
(10 mM Tris HCl, 150 mM NaCl, 0.1 g/L EDTA, 1 mM NaN3, pH 8.0), containing 150mM
calcein. The resulting multilamellar liposomes were then freeze-thawed 5 times with liquid
nitrogen to ensure complete mixing, and then extruded through a 0.4 µm membrane. LUVs were
separated from free calcein on a Sephadex G25 size exclusion column equilibrated with assay
buffer. Fractions giving the highest increase in fluorescence intensity over background, upon
addition of 0.05% Triton X-100, were used in all assays.
Soluble oligomers or amyloid fibrils formed by PrP(106-126) were added to a 1 cm path length
fluorescence cuvette containing calcein-loaded LUVs. The fluorescence emission intensity at
516 nm was monitored over 1min, with constant stirring. Monomeric PrP(106-126) dissolved in
HFIP (unstructured peptide by CD) or TFE (α-helical peptide by CD) was also tested for
membrane-disrupting activity. All assays were completed in triplicate, and were initially
processed with baseline subtraction and then normalized relative to 100% disruption, measured
as the fluorescence intensity after addition of Triton X-100. In order to normalize the data for
peptide:lipid ratio in the final samples, phospholipid concentrations in the LUV suspensions
were quantified using the colorimetric assay described by Stewart et al. (Stewart 1980). Briefly,
unused liposome suspensions from the dye release assay were subjected to chloroform separation
and dried. Afterwards, they were resuspended in chloroform followed by the addition of an
61
aqueous solution ammonium ferrothiocyanate which forms a complex in the chloroform layer
which was then subjected to spectrophotometric analysis and compared to a standard curve.
3.3.8 Solid State Nuclear Magnetic Resonance
Lyophilized oligomers were packed into standard 22 l 3.2 mm MAS rotors. Solid state NMR
measurements were carried out on a narrow bore Varian VNMRS spectrometer, operating at a 1H
frequency of 499.82 MHz. All experiments were carried out using a Varian triple-resonance 3.2
mm T3 MAS probe or a triple-resonance 3.2mm BioMAS probe. Sample heating in standard T3
probes was alleviated by delivering high-flow rates of ambient temperature dry air to the sample.
All spectra were externally referenced to the downfield 13
C resonance of adamantane at 38.56
ppm relative to TMS (Morcombe and Zilm 2003).
1H -
13C cross polarization was implemented using a ramped radio frequency (rf) field centered
around 40-60 kHz on the low channel, with a 50-80 kHz field on the 1H channel. A linear ramp
on the 13
C channel was used, and contact times were typically 1-1.5 ms in length. /2 pulse
widths were typically 2.5-3 s for all channels on the T3 MAS probe, or 2.5 s (1H) and 5.5 s
(13
C, 15
N) on the BioMAS probe. 1H decoupling fields of 110 kHz were applied during all t1 and
t2 periods, using the TPPM decoupling scheme (Bennett et al. 1995). In all cases, a 2 s delay was
used between scans. Two-dimensional (2D) 13
C-13
C NMR spectra were obtained using a radio
frequency assisted diffusion (RAD) pulse sequence for homonuclear recoupling (Takegoshi et al.
2001; Morcombe et al. 2004), with a mixing time of 10 ms and an MAS frequency of 10 kHz.
200 points were taken in t1 at a sweep width of 25 μs, and a total of 48 scans per FID.
All spectra were processed using NMRPipe (Delaglio et al. 1995) and visualized with nmrDraw.
13C assignments were made directly from one-dimensional spectra where spectral overlap was
not present; otherwise assignments were made from two-dimensional 13
C-13
C correlations.
Chemical shifts were compared to the previously reported values for random coil (Wishart and
Sykes 1994). The comparison of experimental shifts with known values yields an accurate
prediction of secondary structure elements within proteins and peptides (Wishart and Sykes
1994). 13
C chemical shifts and linewidths obtained for PrP(106-126) oligomers were compared
with chemical shifts reported for amyloid fibrils discussed in chapter 2.
62
3.4 Results
3.4.1 PrP(106-126) Forms Stable β-sheet Non-fibrillar Oligomers
PrP(106-126) oligomers appear as 5-30 nm spheres by negative stain transmission electron
microscopy (TEM) (Figure 3-1A)., TEM images of samples embedded in a layer of stain
sandwiched between two thin carbon films reveal an asymmetric oligomer morphology, with
particles approximately 12-20 nm x 30 nm (Figure 3-1B-D).
Figure 3-1 – Transmission electron microscopy of PrP(106-126) oligomers
(A) Peptide oligomers stained with uranyl acetate and imaged at 100,000 x magnification. (B) and (C) Lower
magnification images of two samples prepared using a double-carbon technique, with the uranyl acetate stain
sandwiched between two carbon films. Due to technical limitations of the specimen preparation for these images,
use of more dilute samples was necessary. Examples of peptide oligomers are indicated by the arrows. Apparently
elongated objects in panels (B) and (C) arise from clustering of two or more small oligomers. (D) Representative
TEM images of individual PrP(106-126) oligomers from samples shown in (B) and (C).
63
This method may reduce some artifacts associated with staining, and may therefore provide a
more accurate representation of PrP(106-126) oligomers. Even after several weeks of incubation,
no fibrils or morphological changes are observed in TEM images (Figure 3-2B). Likewise,
samples lyophilized for NMR exhibit similar TEM morphology after reconstitution (Figure 3-
2C).
Figure 3-2 - Negative stain TEM images of oligomeric PrP(106-126) sample with different histories do not
show evidence of fibril formation or changes in morphology.
(A) TEM image of freshly prepared oligomeric PrP(106-126). (B) PrP(106-126) oligomers aged in solution at
ambient temperatures for more than 60 days prior to preparing the TEM sample. (C) Oligomeric PrP(106-126)
reconstituted from a lyophilized NMR sample. (D) For comparison, a negative stain TEM image of amyloid fibrils
formed by PrP(106-126) is shown. All images were prepared using standard single-carbon methods and stained with
uranyl acetate. The scale bar represents 100 nm in all panels.
Additionally, approximately spherical PrP(106-126) oligomers were also observed in unstained
samples using atomic force microscopy (AFM) (Figure 3-3) which agrees well with both single
and double carbon negative stain TEM.
64
Figure 3-3 - AFM of PrP(106-126) oligomers.
An image obtained using tapping mode in air. The general appearance of small spherical structures is in agreement
with our TEM data, and is distinct from the appearance of amyloid fibrils formed by this peptide, although
quantitative analysis of the AFM images was not performed due to potential artifacts arising from drying, surface
interactions, or sample deformation by the AFM probe tip.
65
We have also defined the secondary structure of PrP(106-126) oligomers using circular
dichroism (CD) and solid state NMR. The CD spectra shown in Figure 3-5A are characteristic of
a β-sheet secondary structure, in contrast to the unstructured monomeric peptide.
3.4.2 PrP(106-126) Non-fibrillar Oligomers Form as a Discrete Size
Figure 3-4 – Dynamic Light Scattering of PrP(106-126) Non-fibrillar Oligomers
The figure shows the hydrodynamic radius, R(nm) versus percent mass for a solution of non-fibrillar oligomers of
PrP(106-126)
Analysis of the non-fibrillar oligomers of PrP(106-126) by dynamic light scattering shows that
they form with a relatively small size distribution. Figure 3-4 shows the hydronamic radii present
in a solution of PrP(106-126) oligomers with 81.4 % of the particles analyzed having a
hydrodynamic radius of 30 nm.
66
Figure 3-5 – Spectroscopic analysis of PrP(106-126) Oligomers
(A) Circular dichroism spectra of non-fibrillar oligomers compared with monomeric peptide in HFIP (monomer
spectrum as reported in chapter 2). (B) 13
C secondary NMR chemical shifts for PrP(106-126) oligomers. (C) ThT
fluorescence emission spectra of PrP(106-126) fibrils and soluble oligomers as a function of peptide concentration.
Likewise, the 13
C NMR chemical shifts obtained for residues 113-126 of PrP(106-126) are
consistent with the presence of an extended β-strand spanning this region of the peptide (Figure
3-5B). As shown in Figure 3-6, the 13
C linewidths observed for residues 113-124 are less than 2
ppm, suggestive of a well ordered system. A slight increase in NMR linewidth is observed for
13C resonances from L125 and G126, at the C-terminus, likely indicating increased disorder at
the end of the β-strand. Overall, these results suggest a similar secondary structure to PrP(106-
126) fibrils described in chapter 2.
67
Figure 3-6 - 13
C NMR linewidths for PrP(106-126) oligomers.
The half-height linewidths are shown for each assigned 13
C resonance in soluble oligomers formed by PrP(106-126).
Values less than 2.0 ppm are indicative of a well ordered structure. Slightly broader lines are observed for L125 and
G126, possibly suggesting increased disorder towards the C-terminus.
3.4.3 PrP(106-126) Non-fibrillar Oligomers Disrupt Model Membranes
To assess the reported cytotoxicity of PrP(106-126) oligomers, we utilized a dye –release assay
to determine whether these oligomers affect the integrity of model membranes. As shown in
Figure 3-7, addition of oligomeric PrP(106-126) causes leakage of the dye, calcein, from pre-
loaded vesicles indicating a loss in structural integrity. Additionally, the addition of fibrillar,
monomeric (unstructured) or monomeric (α-helical) peptides has no effect on the leakage of dye.
68
Figure 3-7 – Release of the fluorescent dye calcein from 3:1 POPC:POPG liposomes induced by PrP(106-126)
oligomers.
Data points represent the average of 3 independent measurements of fluorescence emission at 516nm, and are
normalized against background fluorescence. Error bars indicate the standard deviation between the triplicate runs.
Monomeric peptide and amyloid fibrils do not significantly increase membrane permeability to calcein, while
soluble oligomers cause significant dye release, in a concentration-dependent manner.
69
3.5 Discussion
PrP(106-126) has been studied as both a model amyloid and prion peptide. In this chapter, we
describe large, β-sheet containing amyloid oligomers that do not bind thioflavin-T. These
spherical structures appear, by TEM, to have dimensions of 12-20 nm by 30 nm. The narrow size
distribution of PrP(106-126) oligomers is supported by dynamic light scattering (DLS)
measurements in which >97% of the sample mass resides in particles with a hydrodynamic
radius of 30-40 nm. This suggests that the oligomers have a mass over 1 MDa, corresponding to
more than 500 peptide monomers per oligomer which is close to the size of large Aβ oligomers
recently studied by solid state NMR (Chimon et al. 2007), as opposed to the 3-5 nm diameter
previously reported for oligomers of PrP(106-126) (Kayed et al. 2004). Our results are also
consistent with sedimentation velocity experiments performed on the amyloid peptide amylin
(Vaiana et al. 2008), in which no small oligomers were detected, leading the authors to estimate
that amylin predominantly forms aggregates of at least 390 kDa.
The cytotoxicity of PrP(106-126) has been debated in the literature – the conflicting reports
likely stem from the different aggregated states accessible to this peptide. The oligomers studied
here exhibit a potent ability to disrupt model membranes. Soluble PrP(106-126) oligomers
increase membrane permeability to the self-quenching fluorescent dye calcein (Figure 3-3).
Amyloid fibrils, unstructured monomers (in HFIP), or helical monomers (in TFE) exhibit no
activity in this assay, supporting the formation of large soluble oligomers as a potentially
important step in the cytotoxicity of PrP(106-126).
In contrast to the amyloid fibrils formed by PrP(106-126), addition of non-fibrillar oligomers to
thioflavin-T solutions does not result in increased fluorescence emission at 482nm. (Figure 3-
5C). This strongly suggests that the PrP(106-126) oligomers described here lack the
characteristic cross-β structure of the fibrillar form. Alternatively, it is possible that the ThT
binding sites may be occluded in the non-fibrillar oligomers. In either scenario, the lack of dye
binding indicates significant differences in peptide packing relative to amyloid fibrils. This result
contrasts with the fibril-like nature of large Aβ oligomers, as monitored by NMR and ThT
binding (Chimon et al. 2007).
While the relationship between the soluble oligomers of PrP(106-126) and amyloid fibrils is
unclear, it is remarkable that no conversion to larger aggregates is observed. The soluble
oligomers form rapidly in solution, and remain unchanged on a timescale of at least weeks under
70
the solution conditions reported here. No additional loss of peptide mass is observed with
extensive dialysis, suggesting that reversion to monomeric peptide does not occur. The absence
of ThT binding at up to 520 μM oligomer indicates a lack of fibril formation (Appendix figure
A-3), and is supported by TEM analysis of aged samples (Figure 3-2B). Removal of residual
monomeric peptide by dialysis results in samples containing only the oligomeric species.
Previous structural studies of amyloid oligomers have relied on trapping a transient state or
stabilizing oligomers with detergents (Chimon et al. 2007; Yu et al. 2009). Here we demonstrate
the formation of a stable, membrane-disrupting oligomeric species by a model amyloid peptide.
This system will facilitate investigations of the structure and mechanism of action of non-fibrillar
oligomers of PrP(106-126), and may provide insight into the assembly of other amyloids.
Based on the importance of non-fibrillar oligomers in the pathogenesis of amyloid diseases, it is
essential to develop a detailed understanding of the factors governing formation, molecular
architecture and activity of non-fibrillar assemblies. In addition to its utility as a model amyloid
peptide, PrP(106-126) has been extensively investigated for its potential role in mediating
conversion of PrPC to PrP
Sc in mammalian prion disease. Therefore, a detailed understanding of
the aggregated states accessible to this peptide may improve our understanding of the events
underlying prion conversion.
71
4 Structural Properties and Dynamic Behaviour of Non-Fibrillar Oligomers Formed by PrP(106-126)
Selections from this chapter were previously published in the Journal of the American Chemical
Society. Walsh, P., Neudecker, P., and Sharpe, S. Structural Characterization and Dynamic
Behaviour of non-fibrillar oligomers formed by PrP(106-126). Journal of the American Chemical
Socity. 2010. 132(22). pp7684-7695. Reprinted with permission. Copyright American Chemical
Society 2010.
All experiments were carried out by P. Walsh with the exception of solution NMR studies
(including pulse-field gradient diffusion experiments and hydrodynamic radius calculations)
which were done in collaboration with Dr. P. Neudecker.
72
4.1 Abstract
The formation of non-fibrillar oligomers has been proposed as a common element of the
aggregation pathway of proteins and peptides associated with neurodegenerative diseases such as
Alzheimer’s and Creutzfeldt - Jakob disease. While fibrillar structures have long been considered
indicators of diseases linked with the accumulation of amyloid plaques, it has more recently been
proposed that amyloid oligomers are in fact the cytotoxic form. Here we describe the local
structure and dynamics of stable oligomers formed by a peptide comprising residues 106-126 of
the human prion protein (PrP). Structural constraints from solid state NMR reveal quaternary
packing interactions within the hydrophobic core, similar to those previously reported for
amyloid fibrils formed by this peptide, and consistent with structural studies of oligomers formed
by the Alzheimer’s β-amyloid peptide. However, a hydration-dependent increase in disorder is
observed for non-fibrillar oligomers of PrP(106-126). In solution NMR spectra we observe
narrow 1H and
13C resonances corresponding to a monomer in exchange with the ~30 nm
diameter non-fibrillar oligomers, giving additional information on the molecular structure of
these species. Taken together, our data support a model in which the local structure of the
oligomers contains the basic elements of amyloid fibrils, but with long-range disorder and local
mobility that distinguishes these assemblies from the fibrillar form of PrP(106-126). These
characteristics may provide a basis for the differing biological activities of amyloid fibrils and
oligomers.
73
4.2 Introduction
Neurodegeneration associated with protein misfolding is characteristic of several human diseases
including Alzheimer’s, Parkinson’s and prion diseases. In each case, one of the primary
pathological markers is the presence of proteinaceous plaques containing amyloid fibrils formed
by the misfolded protein. Despite initial suggestions that formation of fibrils directly results in
cytotoxicity, it is currently hypothesized that non-fibrillar oligomers are responsible for neuronal
cell death and disease progression (Bucciantini et al. 2002; Caughey and Lansbury 2003; Kayed
et al. 2003). Non-fibrillar oligomers have been observed for a number of amyloid proteins and
peptides in vivo and in vitro, including the Alzheimer’s β-amyloid protein (Aβ), α-synuclein and
IAPP (Kayed et al. 1999; Conway et al. 2000; Walsh et al. 2002; Haass and Selkoe 2007). The
discovery of antibodies that recognize oligomers formed by several amyloid peptides suggests
that these assemblies share common structural elements (Kayed et al. 2003). In addition, non-
fibrillar oligomers exhibit a marked increase in toxicity relative to their fibrillar counterparts,
which is largely attributed to their membrane-disrupting ability, and which has been proposed to
represent a common mechanism for the degenerative nature of amyloid diseases (Kayed et al.
2004; Sokolov et al. 2006; Glabe, C. G. 2008).
Recent studies on amyloid forming peptides and proteins have shed light on the structural
properties of amyloid fibrils (Jaroniec et al. 2004; Luca et al. 2007; Sawaya et al. 2007). The
cross-β motif, in which the protein forms β-strands perpendicular to the long axis of the fibril,
has been observed in several fibril structures, including Aβ1-40 (Petkova et al. 2002), amylin
(Luca et al. 2007; Wiltzius et al. 2008), and crystals of several short amyloid peptides (Sawaya et
al. 2007). In each case, the core of the protein contains a dehydrated interface between stacked β-
sheets, creating a ‘steric-zipper’, as described by Sawaya et al. (Sawaya et al. 2007). Illustrating
the potential for variations on this theme, a recently reported structure for the Het-S yeast prion
revealed β-solenoid or β-helical packing rather than a steric zipper in the core of the protein
(Wasmer et al. 2008). A similar arrangement has been proposed for the amyloid-like filaments
formed by the E. coli curli protein (Shewmaker et al. 2009).
In contrast to amyloid fibrils, relatively little is known regarding the molecular structure of non-
fibrillar amyloid oligomers. A detailed characterization is of considerable interest since
oligomers formed by several amyloid proteins have been shown to cause cell death in cultured
74
neurons (Lambert et al. 1998) as well as endothelial cells (Bhatia et al. 2000; Zhu et al. 2000).
The observation that these species cause disruption of calcium regulation, in combination with
several experiments suggesting that amyloid peptides can form conductive channels in planar
bilayers (Lin, M. C. et al. 1997; Kourie, J.I. and Culverson 2000; Kourie, J. I. et al. 2001), has
led to the concept that channel formation is an important aspect of oligomer cytotoxicity (Pollard
et al. 1995; Lin, M. C. et al. 1997; Kayed et al. 2004). Other possible mechanisms which have
been proposed include a non-specific disruption of membranes, possibly through insertion of
hydrophobic segments into the plasma membrane, or through surface-mediated nucleation of
fibril formation (McLaurin and Chakrabartty 1996; McLaurin and Chakrabartty 1997; Yip et al.
2002). The induction of apoptosis in cultured neurons by amyloid oligomers has been reported,
although a mechanism for this has not been revealed (O'Donovan et al. 2001; Carimalo et al.
2005). High-resolution structural studies of non-fibrillar amyloid oligomers have been limited to
Aβ1-40 and Aβ1-42, and have provided strong evidence that these species are comprised of -
sheets with significant intermolecular strand formation (Chimon et al. 2007; Yu et al. 2009).
Solid state NMR studies reported by Chimon et al. revealed fibril-like packing within the -sheet
containing core of large Aβ1-40 oligomers (Chimon et al. 2007). A significantly different
organization of inter-and intramolecular -sheets was observed for small 16-64 kDa detergent
stabilized globulomers of Aβ1-42 (Yu et al. 2009). In contrast, NMR data reported for pore-
forming oligomers of α-synuclein suggest a poorly ordered assembly with secondary structure
distinct from α-synuclein fibrils (Kim, H. Y. et al. 2009). Thus questions remain regarding the
potential for common elements that result in the cytotoxicity of non-fibrillar amyloid oligomers.
As a model for investigating the structures accessible to amyloid peptides we have focused on a
21-residue peptide derived from the mammalian prion protein (PrP). This peptide, PrP(106-126),
forms amyloid fibrils (Forloni et al. 1993; Selvaggini et al. 1993; Salmona et al. 1999) as well as
cytotoxic non-fibrillar oligomers (Kayed et al. 2003; Kayed et al. 2004), the latter of which have
been proposed to either form ion channels (Lin, M. C. et al. 1997; Florio et al. 1998; Kourie, J.I.
and Culverson 2000) or to induce a disruption of cellular membranes (Salmona et al. 1997). In
chapter 2, solid-state NMR was used to determine the structure of amyloid fibrils formed by
PrP(106-126), revealing a class 1 steric zipper motif. Chapter 3 is an initial biophysical
characterization of membrane-disrupting, non-fibrillar oligomers of PrP(106-126), in which
dynamic light scattering (DLS), transmission electron microscopy (TEM), and solid state NMR
75
were used to define the morphology and secondary structure of these species. They were
observed to be roughly spherical structures with largely β-sheet secondary structure and a
hydrodynamic radius of approximately 30 nm.
Here, we present a comprehensive NMR investigation of the local structure and dynamics of
PrP(106-126) in non-fibrillar oligomers. We find that these assemblies contain subunits with
secondary and quaternary structure with strong similarity to the fibrillar form of this peptide. In
particular, dipolar recoupling experiments indicate the presence of a parallel, in-register -sheet
structure, with sheets stacked in an antiparallel fashion to form fibril-like subunits. Solution
NMR reveals the presence of a small population of structured monomers in rapid equilibrium
with the large oligomers. Additional data suggest increased local motions and potential long-
range disorder in the non-fibrillar oligomers relative to amyloid fibrils. Based on these data, a
putative model for the oligomeric assembly of PrP(106-126) is presented.
76
4.3 Materials and Methods
4.3.1 Solid State NMR
13C-
13C RAD and PITHIRDS experiments were carried out as described in Chapter 2.
13C T1
relaxation times were obtained from cross polarization spectra recorded using the spin-
temperature inversion method described by Torchia (Torchia 1978), with the time between the
two π/2 pulses arrayed from 0 to 7.5 s. T2 relaxation time measurements were performed using a
CPMG spin-echo experiment (Carr and Purcell 1954; Meiboom and Gill 1958), with an echo
period (including π-pulse) arrayed from 50 to 5000 μs. In both T1 and T2 experiments, 1H-
13C
cross-polarization was achieved as described above with an MAS frequency of 10 kHz. Peak
intensity was plotted as a function of the recovery time and in each case was fitted to a single
exponential decay using Origin software.
To directly observe solvent accessible sites Mn2+
, a paramagnetic shift reagent, was added to
both fibril and oligomer samples. Briefly, lyophilized fibrils and oligomers were hydrated with
excess water in a 36 μl MAS rotor, with the water being added after weighing the dry sample.
1D 13
C cross polarization spectra of the hydrated samples were recorded in the presence and
absence of MnEDTA (0.2 mole per mole peptide). A 4 s recycle delay was used in these
experiments, and was sufficient to prevent sample heating.
4.3.2 Solution NMR
Solution NMR spectra were recorded on samples containing 2.0 mM not isotope-enriched
PrP(106-126) oligomers. 2D [1H,
1H]-TOCSY with a 10 kHz DIPSI-2 mixing scheme (Rucker
and Shaka 1989) (45 ms mixing time), 2D [1H,
1H]-NOESY (Jeener et al. 1979) (600 ms mixing
time), and natural abundance [1H,
13C]-HSQC (Bodenhausen and Ruben 1980) spectra were
recorded at 25 °C on a Varian Unity INOVA 500 MHz NMR spectrometer equipped with a
room-temperature probe with z-axis pulsed field gradient capabilities. The H2O resonance was
suppressed by WATERGATE (Piotto et al. 1992) with quadrature detection in the indirect 1H
dimension achieved by States-TPPI (Marion et al. 1989) in the homonuclear experiments; in the
[1H,
13C]-HSQC the H2O resonance was suppressed by gradient coherence selection with
quadrature detection in the indirect 13
C dimension achieved by the echo-antiecho method (Kay et
al. 1992; Schleucher et al. 1993).
77
All solution NMR spectra were processed with NMRPipe (Delaglio et al. 1995) software and
analyzed with NMRViewJ 8.0.b16 (Johnson and Blevins 1994). 1H chemical shifts were
referenced with respect to external DSS in D2O and 13
C chemical shifts were referenced
indirectly (Markley et al. 1998). Spin systems were initially identified using the 1H-
1H TOCSY
in conjunction with the 1H-
13C HSQC. Intra-residue and sequential inter-residue connectivities
were then assigned using the 1H-
1H NOESY.
The diffusion coefficient of structured PrP(106-126) monomers was measured using 1D 1H pulse
gradient stimulated echo longitudinal encode-decode (PG-SLED) translational diffusion
experiments (Altieri et al. 1995; Choy et al. 2002) with suppression of the H2O resonance by
WATERGATE (Piotto et al. 1992) at 500 MHz, 25 °C. Three non-overlapping methyl group
regions in the 1D 1H spectra were integrated independently and the resulting intensities as a
function of gradient strength fit independently by three Gaussian decays (Stejskal and Tanner
1965). The decay constants from these fits were converted into diffusion coefficients based
(Stejskal and Tanner 1965) on the absolute strength of the pulse field gradients, which had been
calibrated carefully using two independent methods (Altieri et al. 1995) in agreement with each
other to better than 0.5%. The resulting diffusion coefficients were in turn converted into
hydrodynamic radii based on the Stokes-Einstein equation assuming a viscosity of 0.900×10-
3 Pa s interpolated for 5% D2O at 25 °C (Cho et al. 1999). The diffusion coefficient and
hydrodynamic radii are reported as mean ± standard error over the three non-overlapping methyl
group regions.
78
4.4 Results
4.4.1 PrP(106-126) Oligomers Contain In-register Parallel β-sheets
As reported in chapter 3, analysis of the solid state 13
C and 15
N chemical shifts for residues
within the hydrophobic core sequence of non-fibrillar PrP(106-126) oligomers shows they are
very similar to those observed for fibrils formed by this peptide as shown in chapter 2, and
strongly suggest the presence of an extended β-strand from residues 113-125. In order to confirm
that these results were not due to freeze-drying the oligomers, chemical shifts were assigned for
rehydrated PrP(106-126) oligomers using 1D and 2D 13
C MAS NMR spectra. For most sites, the
13C linewidths in the dry samples are 2ppm, suggesting a well ordered and homogeneous
structure. For comparison, the secondary chemical shifts for both lyophilized and rehydrated
oligomers are shown in Figure 4-1A, and confirm that the same secondary structure is present in
each case. Likewise, the 13
C linewidths (Figure 4-1B) for most sites remain unchanged upon
addition of excess water to the PrP(106-126) oligomers, although some broadening is observed
for G119-V122 upon hydration, suggesting an increase in disorder or mobility.
79
Figure 4-1 - Comparison of 13
C secondary chemical shifts and NMR linewidths of hydrated versus lyophilized
PrP(106-126) oligomers.
(A) Deviations of CO, Cα and Cβ chemical shifts from random coil values (Wishart and Sykes 1994) are shown for
hydrated and lyophilized non-fibrillar oligomers of PrP(106-126). (B) 13
C NMR linewidths for the same resonances.
Where possible, chemical shifts and linewidths were obtained from 1D 13
C spectra. All others were obtained from
2D 13
C-13
C correlation spectra recorded using a RAD mixing period of 10 ms.
80
Residues 123-126 exhibit slightly broadened lines in both lyophilized and rehydrated oligomers,
again suggestive of some conformational disorder of the peptide C-terminus in both assemblies.
This is supported by a loss of signal intensity for the G126 CO-Cα crosspeak in radio frequency
assisted diffusion (RAD) spectra of hydrated PrP(106-126)AVG2
oligomers (Figure 4-2) which is
likely a result of reduced dipolar couplings due to motional averaging at this site.
Figure 4-2 - 13
C-13
C chemical shift correlation spectra of dry versus hydrated PrP(106-126)AVG2
oligomers.
13C-
13C correlation spectra obtained using a 10 ms RAD mixing time are shown for dry (A) and hydrated (B)
oligomers of the PrP(106-126)AVG2
peptide. Horizontal slices from the G126 C frequency are shown below each
spectrum, normalized to the intensity of the diagonal peak for this resonance. The position of the intraresidue G126
CO - C crosspeak is indicated on each slice. In the hydrated sample there is a significant loss of crosspeak
intensity relative to the diagonal Cα peak, indicating a decrease in the effective dipolar coupling between these
nuclei. This likely results from increased motional averaging in the presence of bulk water.
Based on the presence of parallel, in-register β-sheets in amyloid fibrils formed by PrP(106-126)
discussed in chapter 2, we used the PITHIRDS homonuclear recoupling scheme (Tycko 2007)
to test for a similar arrangement in non-fibrillar oligomers. This experiment reports on the
average internuclear distance between different copies of a given atom within a homooligomeric
β-sheet. Shorter distances between labeled sites will result in an increased rate of signal
dephasing due to homonuclear dipolar couplings. PITHIRDS data obtained for oligomers
81
formed by PrP(106-126)GCO
, PrP(106-126)ACO
and PrP(106-126)AAGG
, are shown in Figure 4-3A.
Fitting to the simulated dephasing curves gives distances of 5.6, 6.2 and 5.6 Å for the A115 Cβ,
A120 carbonyl and G124 Cα atoms respectively. These distances are consistent with our
previous measurements on PrP(106-126) fibrils, and indicate that the β-strands in non-fibrillar
oligomers of this peptide adopt a similar in-register, parallel β-sheet structure. In particular, only
this arrangement can account for the 5.6 Å distance between A115 β-carbons on adjacent strands.
The best fit for the experimental G114 carbonyl data gives a distance of 7.2 Å between adjacent
G114 residues, indicating an increased average interstrand distance at the N-terminus of the β-
sheet, potentially a result of reduced structural order at this site.
Figure 4-3 - PITHIRDS recoupling curves for non-fibrillar PrP(106-126) oligomers.
(A) 13
C dipolar dephasing curves obtained using PITHIRDS-CT (Tycko 2007) are shown for oligomers formed by
PrP(106-126)GCO
(blue squares), PrP(106-126)ACO
(red), the A115 C resonance of PrP(106-126)AAGG
(magenta),
and the G124 C resonance of PrP(106-126)AAGG
(green). The best fits to internuclear distances simulated using
Spinevolution are 7.2, 6.2, 5.6 and 5.6Å, respectively. (B) A comparison of PITHIRDS recoupling curves for dry
(blue) and hydrated (orange) PrP(106-126) GCO
oligomers or hydrated amyloid fibrils (black). The corresponding
internuclear distances are 7.2Å (dry oligomers), 7.6Å (hydrated oligomers) and 7.0Å (hydrated fibrils). Simulated
data are shown in both panels as solid lines from 5.4Å (lowest curve) to 7.8 Å (highest curve) in 0.2Å increments.
Error bars for experimental data were calculated from the RMS noise in the PITHRIDS recoupling spectra. Note that
in some cases, the error is smaller than the size of the symbol.
While the data shown in Figure 4-3 are for lyophilized oligomers, we have also recorded
PITHIRDS curves for hydrated oligomers. The slight increase in the measured G114 carbonyl
distance, from 7.2 Å in the dry oligomers to 7.6 Å in hydrated samples, is indicative of either an
increased average internuclear distance upon hydration or an increase in motional averaging of
82
the dipolar couplings in the presence of excess water (Figure 4-3B). In both cases, this is slightly
longer than the 7.0 Å distance observed in hydrated PrP(106-126) fibrils, and is suggestive of
increased disorder at the N-terminus of oligomers relative to fibrillar assemblies.
4.4.2 PrP(106-126) Oligomers Contain Quaternary Contacts Between β-sheets Similar to those in PrP(106-126) Amyloid Fibrils
Radio frequency assisted diffusion (RAD) spectra with various 13
C-13
C spin diffusion mixing
times were recorded for dry and hydrated PrP(106-126) oligomers. At shorter mixing times,
crosspeaks are observed between directly bonded 13
C atoms, permitting identification of amino
acid spin systems for resonance assignments. At longer mixing times of 250 – 500 ms,
crosspeaks are seen between all 13
C nuclei with internuclear distances of less than 6-7 Å. Thus
these data provide valuable information regarding tertiary and quaternary structure in proteins.
Figure 4-4 shows RAD spectra of PrP(106-126)GAVL
oligomers, obtained with mixing times of
10, 250 and 500 ms. At the longer mixing times, crosspeaks corresponding to long-range
contacts between atoms in G114/A116 and sites within V122 are observed. Since these residues
are at opposite ends of an extended parallel β-sheet, the observed connectivies must arise from
intersheet contacts within the oligomer, as we have previously observed for amyloid fibrils
formed by this peptide.
83
Figure 4-4 - Long-range 13
C-13
C internuclear contacts observed in 2D 13
C-13
C NMR spectra of PrP(106-
126)GAVL
oligomers.
(A) 13
C-13
C correlation spectra obtained with a RAD mixing time of 10 ms. Similar spectra obtained with mixing
times of 250 and 500 ms are shown in (B) and (C), respectively. Horizontal slices at the G114C , A116C, and
V122C frequencies are shown for 10 ms (D), 250 ms (E) and 500 ms (F) mixing times. Interresidue cross peaks are
indicated on the slices.
These contacts are maintained in the presence of excess water, as shown in the 500 ms RAD
spectrum of hydrated PrP(106-126)GAVL
oligomers (Figure 4-5), and provide strong evidence for
an antiparallel arrangement of opposing -sheets within the oligomeric assembly.
84
Figure 4-5 - Long-range 13
C-13
C internuclear contacts are maintained in hydrated PrP(106-126)GAVL
oligomers.
A 13
C-13
C correlation spectrum of hydrated PrP(106-126)GAVL oligomers, obtained using a 500 ms RAD mixing
time is shown in the upper panel. Horizontal slices from the G114C and V122Cγ frequencies are shown below.
Interresidue crosspeaks are identified on the slices as in Figure 4-4.
4.4.3 Identification of Structured Monomeric PrP(106-126) in Fast Exchange with Non-fibrillar Oligomers from 1H-1H and 1H-13C Solution NMR Spectra
In order to further probe the structure and dynamics of non-fibrillar PrP(106-126), 1H-
1H
TOCSY, NOESY and 1H-
13C HSQC spectra were recorded for a solution of oligomers formed
85
by unlabeled PrP(106-126). The resulting solution spectra gave surprisingly sharp, well resolved
resonances, which were unlikely to arise from the amyloid oligomers, which have an estimated
molecular weight of approximately 1 MDa. To address this issue, pulse-field gradient NMR
experiments were carried out, and the correlation time of the molecular species giving rise to the
sharp resonances was determined. The PrP(106-126) species observed by solution NMR exhibits
a translational diffusion coefficient of D = 2.15×10-6
cm2/s ± 0.11×10
-6 cm
2/s at 25 °C (Fig 4-6).
Figure 4-6 Translational diffusion of PrP(106-126) non-fibrillar oligomers.
Integrated intensity I(Grel
) of the spectral region comprising alanine methyl groups (1.31..1.43 ppm) in a 1D 1
H PG-
SLED translational diffusion experiment with a diffusion delay of T = 100 ms between two rectangular pulse field
gradients of duration δ = 2.0 ms each at 500 MHz, 25°C, as a function of relative gradient strength Grel
. Grel
= 1
corresponds to an absolute gradient strength of 66.4 G/cm. The solid line indicates the best-fit Gaussian decay
(diffusion coefficient D = 1.96×10-6
cm2
/s) to the experimental data (filled circles).
86
This corresponds to a hydrodynamic radius of Rh = 11.3 Å ± 0.6 Å, which is identical within
error to the hydrodynamic radius predicted (Wilkins et al. 1999) for a partially folded (11.5 Å ±
1.3 Å) or denatured (12.5 Å ± 1.1 Å) 21-residue monomer.
The aliphatic regions of the TOCSY and HSQC spectra are presented in Figure 4-7, and contain
a number of well-resolved crosspeaks.
Figure 4-7 – 1H-
1H TOCSY and 1H-13C HSQC NMR spectra of PrP(106-126) monomers in equilibrium with
non-fibrillar oligomers
1H and
13C assignments for several sites are shown in 2D correlation spectra of PrP(106-126) oligomers in deuterated
10 mM acetate buffer (pH 4.6), recorded under solution NMR conditions. A portion of the aliphatic region of a 1H-
1H TOCSY is shown in (A), while the corresponding region of a
1H-
13C HSQC spectrum is shown in (B). Identified
1H-
13C correlations are labeled in (B), with ambiguous assignments, for which only the amino acid type is known,
indicated by asterisks. Connections to the corresponding spin systems in the 1H-
1H TOCSY spectrum are indicated
by dashed lines. Specific assignments were made as described in the text.
87
A portion of the amide region of the HN-Hα region of the TOCSY spectrum is shown in Figure
4-8A. Spin systems corresponding to T107, N108, H111 and L125 were unambiguously
identified using only their characteristic connectivity patterns in the TOCSY spectrum, allowing
assignment of most 1H and
13C resonances from these residues. In addition, two inequivalent
Lys and Met spin systems were identified, along with two Val, three Gly, and two Ala residues.
A number of broad and poorly resolved resonances were also observed in the TOCSY spectrum,
which were tentatively assigned to alanines, based on their chemical shift and spin systems.
Figure 4-8 - Sequential and intermolecular NOEs observed in a solution containing non-fibrillar oligomers of
PrP(106-126).
The HN-Hα region of a 1H-
1H TOCSY spectrum of PrP(106-126) oligomers is shown in (A), with backbone
assignments indicated. As in Figure 4-7, ambiguous peak assignments are indicated by asterisks, and are labeled by
the amino acid spin system identified. The same expansion of a 1H-
1H NOESY spectrum is shown in (B), with intra
and interresidue peak assignments as indicated. A sample interresidue NOE cross peak is indicated by the dashed
lines connecting the Hα resonance of G119 with the amide of A120. Negative cross peaks are shown in red in the
NOESY spectrum.
Relatively few interresidue crosspeaks were observed in the NOESY spectrum (HN-Hα region
shown in Figure 4-8B, suggesting a relative lack of well-ordered secondary structure in the
regions of PrP(106-126) exhibiting sharp 1
H resonances. A small number of Hαi – HNi+1 NOEs
were identified (M112 – A113, A113 – G114, G119 – A120, A120 – V121, V122 – G123),
allowing unambiguous assignments for K106 – G114 and G119 – G123, with the exception of
K106 and K110, which cannot be identified with certainty. Two remaining glycines were
88
ambiguously assigned to G120 and G126. Additional interresidue NOEs were observed between
two unassigned alanine residues with very broad (TOCSY) and weak (NOESY) NMR signals.
These likely represent resonances within the central A115-A118 sequence. Based on the 1H
assignments, 13
C resonances in the HSQC spectrum were assigned where possible.
Several weak crosspeaks were also observed in the NOESY spectrum between the H111 and
M109 sidechains with L125 Hα/HN resonances. These connectivities are very low intensity, and
may result from transient intermolecular contacts between these residues during the assembly or
exchange with the oligomer. It is also important to note that the HA, CA and HN resonances of
G114, 119 and 123 are significantly broadened in the spectra of the monomeric peptide, possibly
suggesting some minor structural heterogeneity towards the core β-sheet region. This may also
result from decreased mobility due to transient interactions with the larger oligomeric assembly.
Some additional evidence for structural heterogeneity is seen in the presence of three distinct sets
of resonances for N108, and broadening for resonances in one of the Lys spin systems. Overall,
the results support the presence of a structured but somewhat disordered monomer in equilibrium
with the oligomeric PrP(106-126).
All unambiguous solution and solid state NMR 1H and
13C shift assignments for a solution
containing PrP(106-126) oligomers are shown in Appendix Table A-3 and A-4. While significant
deviations are observed between the 13
C shifts measured for several sites in the monomer relative
to those obtained under MAS conditions for the oligomer, in both species the chemical shifts are
largely consistent with an extended -sheet containing structure. Deviations in shift could stem
from differences in local structure and or from different degrees of exposure to the bulk solvent.
Despite both the relatively large changes in chemical shift at specific sites, as well as having
monomers in exchange with non-fibrillar oligomers, the secondary structure appears to be
relatively unaffected, based on secondary chemical shift analysis and the extended structure
indicated by the presence of Hαi - HNi+1 NOEs. TALOS (Cornilescu et al. 1999) prediction of
the backbone ψ and torsion angles for residues 107-125 was performed using solid state 13
C
chemical shifts for the oligomers and solution 1H and
13C chemical shifts for the structured
monomer. In the case of ambiguous shifts (G124/126 and K106/110, and A115-118 in the
monomer), calculations were performed using all permutations of these data, resulting in
negligible changes in the predicted torsion angles.
89
Figure 4-9 - φ and ψ backbone torsion angles predicted for PrP(106-126) oligomers and structured
monomers.
Angles are shown for TALOS calculations performed using only 13
C and 15
N chemical shifts obtained from MAS
NMR (G114-L125 only) and for calculations performed using only the 1H and
13C shifts from solution NMR. For
monomer sites with only ambiguous solution assignments the calculation was repeated with all possible
combinations of assigned shifts, with no significant change in the resulting torsion angles. Therefore a representative
set of angles is presented here. For comparison, the torsion angles previously reported for the fibrillar form of this
peptide are also given for residues 114-125.
The predicted backbone torsion angles for the non-fibrillar oligomers are shown in Figure 4-9,
along with the results for the monomeric PrP(106-126) and and ψ values for amyloid fibrils of
PrP(106-126), reported in chapter 2. In each case, the results are consistent with a primarily β-
strand secondary structure for all residues, with the possible exception of a turn at H111/M112 in
the monomeric peptide. TALOS predicts two distinct possibilities at these sites – resulting in
either an extended or bent structure. The absence of supporting NOEs, as well as the likely
mobility at the N-terminus of PrP(106-126) suggest that the TALOS prediction for these residues
may not be entirely accurate. However, the diffusion measurements described above are more
consistent with a somewhat compact form of the peptide, so the corresponding TALOS results
90
are reported here. The chemical shift datasets for the fibrillar and oligomeric forms of this
peptide do not include K106-M112, preventing a direct comparison of the N-terminal structure in
these morphologies.
4.4.4 MAS NMR Paramagnetic Relaxation Enhancement (PRE) of PrP(106-126) Fibrils and Oligomers
PRE by Mn2+
was used to probe solvent exposure in MAS NMR spectra of selectively labeled
PrP(106-126) fibrils and non-fibrillar oligomers. This technique has been used as a means to
probe intermolecular distances in solution (Iwahara and Clore 2006), and in solid state NMR as a
method to probe the depth of protein insertion into model bilayer membranes(Su et al. 2008).
Since various sites within PrP have been proposed to chelate divalent metal ions, we used
MnEDTA to reduce the likelihood of direct protein-metal interactions. One-dimensional 13
C
spectra are shown in Figure 4-10 for PrP(106-126)AVG2
fibrils and oligomers, in the presence and
absence of MnEDTA (for clarity only the aliphatic region is shown). Even at relatively low
concentrations of MnEDTA (1:5 relative to protein concentration), the fibrils show ~ 90% loss of
signal intensity for the C-terminal G126 CO and Cα resonances, and a 20-25% reduction in
signal from the V121 methyls. The latter residue is located on the outer surface of PrP(106-126)
fibrils, although we cannot exclude the possibility that a significant portion of the fibril surface
is occluded due to lateral association of fibrils, as seen at high concentrations by TEM. At
higher concentrations of MnEDTA (not shown), there is an overall loss of signal at all sites. The
PRE experiment was repeated for PrP(106-126)AVG
fibrils, with only V121 showing significant
signal reduction.
91
Figure 4-10 - Mn2+
paramagnetic relaxation enhancement effects in 13
C cross-polarization spectra of PrP(106-
126)AVG2
fibrils and oligomers.
13C CP spectra obtained at 10 kHz MAS for hydrated PrP(106-126)
AVG2 fibrils (A, B) and non-fibrillar oligomers (C,
D). Spectra for fibrils in buffer containing MnEDTA (B) show a marked decrease in peak intensity for the G126Cα
resonance relative to samples lacking MnEDTA (A). Similar spectra obtained from non-fibrillar oligomers are
shown for samples with (D) and without (C) MnEDTA. In all cases, a 5:1 ratio of peptide:MnEDTA was used. The
G126Cα peaks are indicated by vertical lines, and a horizontal line is set to the amplitude of this peak in the Mn2+
free spectrum of each pair. Quantitative analysis (E) showing signal loss for all probed sites.
92
PRE experiments were also performed on non-fibrillar oligomers formed from PrP(106-126)AVG
and PrP(106-126)AVG2
, using the same experimental conditions. A 40-45% loss of signal is
observed for the G126 13
C resonances in the non-fibrillar oligomers, with smaller (10-20%)
changes in intensity at all other sites probed. This suggests that while the loss of signal intensity
at most sites was slightly larger than for the fibrils, most of the β-sheet core is somewhat
shielded from the MnEDTA. The signal loss at G126 is suggestive of a peptide arrangement in
which approximately half of the C-termini are exposed to the bulk solvent.
4.4.5 Proposed Structural Model for Non-fibrillar Oligomers of PrP(106-126)
Using the torsion angles predicted by TALOS for residues 107-125, based on the combined
MAS and solution NMR data sets, a set of peptide chains were constructed using CHIMERA and
energy minimized. Due to the high degree of similarity between the monomer and oligomer
secondary structures, both datasets were combined to produce a single set of structural models.
Single chains with representative structures including either extended or kinked torsion angles at
H111/M112 are shown in Figure 4-11A, and are otherwise in a predominantly -sheet
conformation. Using the intermolecular distances measured from the PITHIRDS and RAD
dipolar recoupling data, a tetrameric assembly containing two parallel β-sheets arranged with
antiparallel face-to-face packing between the sheets was created (Figure 4-11B). The NOE and
RAD constraints defining the extended β-strands in the monomers and the antiparallel
arrangement of sheets in the oligomers are summarized in Figure 4-12. Since we cannot exclude
the possibility of poor TALOS predictions for H111 and M112 torsion angles, an alternate
arrangement with one pair of strands containing an extended N-terminus is shown in Figure 4-
11C. A tetramer was used in each case as the minimum assembly that can satisfy the dipolar
recoupling data, although the presence of more extended fibril-like structures cannot be
excluded.
93
Figure 4-11 - Structural models for non-fibrillar oligomers formed by PrP(106-126).
(A) Example structural models for an individual peptide chain, based on backbone torsion angle predictions for the
structured PrP(106-126) monomer. Several residues are labeled on the extended monomer structure. (B) An energy-
minimized model for a fibril-like subunit consistent with the intermolecular contacts obtained from analysis of
dipolar recoupling experiments. A similar model in which the N-terminal segments from one parallel pair of strands
are extended is also shown in (C). (D) A schematic of a putative spherical assembly of fibril-like subunits, in which
a hollow shell of radially aligned peptide chains form a micelle-like structure. Blue dots indicate the presence of
water in the center of the sphere.
94
Figure 4-12 - Schematic representation of intramolecular and intermolecular restraints used in structure
calculations and model building.
Two copies of the PrP(106-126) amino acid sequence are shown, with intramolecular constraints indicated on the
lower copy. 13
C and 15
N chemical shifts were obtained by solid state NMR for all boxed residues. Unambiguous 1H
and 13
C shifts, and intraresidue NOEs from solution NMR are available for green residues, while ambiguous
assignments exist for blue residues. Intramolecular Hαi – HNi+1 NOEs are indicated by dashed green arrows, and an
ambiguous Hαi - HNi+2 or HNi-2 NOE is indicated by blue dashed arrows. Intermolecular connectivities obtained
from dipolar recoupling under MAS are indicated by black arrows. In each case, at least two crosspeaks define the
interaction between residues on adjacent protein chains. Amino acids whose proximity in parallel in-register -
strands has been established by PITHIRDS recoupling curves are underlined in the upper sequence.
Our TEM, DLS and AFM measurements (discussed in chapter 3) indicate that the oligomers are
spherical objects, at least 20-30 nm in diameter. The NMR data presented here indicate a
predominantly extended peptide structure with a local interchain packing reminiscent of amyloid
fibrils with approximately 50% surface exposure of the C-termini. Additionally, the relatively
narrow linewidths observed for both 13
C under MAS, and for 1H and
13C under solution NMR
conditions are suggestive of single conformations/environments for residues in the core, with the
potential for some heterogeneity and mobility at the N- and C-termini. The model that best fits
all of these requirements is a variation on the micelle-like arrangement of peptide chains
previously proposed for spherical oligomers of Aβ (Laurents et al. 2005; Chimon et al. 2007).
Such an arrangement is depicted in Figure 4-11D, in which the tetramers from 4-11C are aligned
radially within a hollow, water-filled sphere with a diameter consistent with DLS and TEM
measurements.
95
4.5 Discussion
While the formation of non-fibrillar oligomers has been proposed as a common element of the
aggregation pathway of amyloid peptides, and an important element of amyloid cytotoxicity, few
structural details of these assemblies have been reported. In chapter 3, we described optimized
solution conditions for forming stable oligomers of PrP(106-126) in preparations that are free
from fibrils, and which are therefore suitable for structural studies. Under these conditions,
similar to those used by Kayed et al. for cytotoxicity studies, spherical oligomers are formed
with an apparent hydrodynamic radius of approximately 30 nm. Preliminary solid state NMR and
CD measurements indicated a predominantly β-sheet secondary structure. Consistent with
previous reports of membrane disruption and cytotoxicity of non-fibrillar amyloid oligomers
(Kayed et al. 2003; Kayed et al. 2004; Chimon et al. 2007), PrP(106-126) oligomers exhibited a
potent ability to cause liposome leakage, while monomers and fibrils formed by this peptide are
relatively inert. Thus we expect that these represent the cytotoxic form of PrP(106-126).
Our results provide local structural constraints which define the presence of parallel in-register
-sheets, packed in an antiparallel arrangement. This arrangement of chains is present in both
lyophilized and rehydrated PrP(106-126) oligomers, and is essentially indistinguishable in our
experiments from the interchain packing previously observed for amyloid fibrils of this peptide
as discussed in chapter 2. We propose, based on NMR and biophysical data, that a hollow,
water-filled micelle-like assembly is the most likely internal structure for the large spherical
oligomers of PrP(106-126). A similar micellar arrangement of fibril-like subunits has recently
been proposed for large oligomers formed by A (Chimon et al. 2007), based on the presence of
parallel in-register -strands in lyophilized oligomers formed by that protein. One distinct
difference between the oligomers formed by these two different peptides, however, is the lack of
thioflavin T (ThT) binding to PrP(106-126) oligomers, while the non-fibrillar oligomers of A
bind ThT. This suggests either differing amounts of cross- structure, or poor accessibility to
dye in the case of PrP(106-126). For instance, in a micelle-like structure, the more ordered β-
sheet core may be shielded from large solutes by the disordered N-termini on the surface of the
sphere.
96
It is important to note that while the overall similarity of our proposed model for PrP(106-126)
oligomers and previous models of Aβ may argue in favor of a common structure for non-fibrillar
amyloid oligomers, it is likely that many local conformations and quaternary structures are
accessible to amyloid proteins during the misfolding process. This is clearly illustrated when our
results and those of Chimon et al (Chimon et al. 2007) are compared to recent solution NMR
studies of the local structure of small (16-64 kDa) A(1-42) oligomers (Yu et al. 2009). In that
study, SDS stabilized small oligomers were shown to form extended -strands, but to have a
combination of fibril-like and non-fibrillar contacts between strands. Similarly, pentamers of
Aβ1-42 contain shorter versions of the β-turn-β confirmation seen in Aβ fibrils (Ahmed et al.
2010). Likewise, recent solid state NMR data reported for non-fibrillar oligomers of α-synuclein,
seem to indicate a significantly different secondary structure for oligomers relative to fibrillar
protein, and may support significantly decreased order in the oligomers (Kim, Hai-Young et al.
2009). Furthermore, cylindrin structures described by Laganowsky et al contain steric zipper
motifs within their β-barrel like structures (Laganowsky et al. 2012). Additionally, some
amyloidogenic sequences are able to induce membrane fusion or induce negative curvature upon
binding to membranes in apparently monomeric disordered or helical conformations, based on
solution NMR studies of their structure when associated with micelles (Brender et al. 2008;
Brender et al. 2009; Nanga et al. 2009). Taken together, this suggests that there may be several
distinct modes of amyloid activity at membranes.
Despite the relatively small changes in secondary and quaternary structure for PrP(106-126)
oligomers relative to amyloid fibrils, we do see strong indications of decreased order and
increased local mobility or conformational heterogeneity in the oligomers. This is highly
dependent on the hydration state of the sample, which contrasts with the relative insensitivity of
most amyloid fibrils to the presence of bulk water (Paravastu et al. 2006; Petkova et al. 2006).
Upon complete hydration of the oligomers with an equal mass of water, but in the absence of
bulk water, we observed relatively small local changes in secondary and quaternary structure for
residues 113-126 using MAS NMR experiments, and only small increases in the 13
C linewidths
for sites within the hydrophobic core of the peptide. This is supported by relatively small
changes in the 13
C T1 and T2 NMR relaxation times (Figure 4-13).
97
Figure 4-13 - 13
C spin relaxation times obtained under MAS for non-fibrillar oligomers and amyloid fibrils
formed by PrP(106-126).
Longitudinal (T1) and transverse (T2) spin relaxation rates are shown in (A) and (B), respectively. Data are shown
for specific Cα and Cβ resonances in fibrils and non-fibrillar oligomers, using both hydrated and lyophilized
samples as indicated. In each case the reported value was obtained by fitting relaxation data to a single exponential
decay.
98
In the presence of bulk water, our solution NMR experiments indicate the presence of a small
population of partially structured monomers in exchange with the large spherical oligomers, as
supported by NMR diffusion measurements. No other set of resonances is observed in the
solution spectra, demonstrating the absence of any significant amount of unstructured monomer
in the oligomer preparations. The presence of an equilibrium between the monomer and oligomer
states is supported by the fact that extensive dialysis does not change the intensity of the NMR
signals from the monomer. Despite containing significant secondary structure, the absence of
long range NOEs, the presence of significant line broadening at several sites and the
identification of multiple resonances for other residues, suggests that the monomeric peptide
remains poorly ordered and likely samples multiple conformations when released from
oligomers. The appearance of intraresidue NOE signals with negative intensity also supports the
presence of rapid local motions in the monomeric peptide, providing further evidence of a
partially structured but poorly ordered entity. The broad resonances observed within the
GAAAAG palindromic sequence may alternately suggest that this is an important site of
interaction with the larger assembly, with line broadening resulting from transient associations.
In terms of the biological activity of PrP(106-126) assemblies, the model proposed in Figure 4-
11D suggests some potential mechanisms for the membrane disruption cytotoxicity attributed to
large non-fibrillar oligomers. It is known that hydrophobic interfaces and surfaces can catalyze
fibril formation, or that they may increase the rate of fibril growth, possibly through increased
local concentration and organization of monomers in a two-dimensional environment. Under
these conditions, the dissociation of oligomers into small ‘seeds’ that can subsequently nucleate
fibril formation at a membrane surface might readily occur, creating a loss of bilayer integrity
and resulting in cell death. This possibility is supported by previous reports that Aβ fibrillization
can cause defects in supported planar bilayers (Yip et al. 2002). An alternate hypothesis is that
upon association of oligomers with membranes, short fibril-like segments are able to insert into
the bilayer, forming a barrel-stave or toroidal pore. Pore formation by amyloid peptides has been
suggested as a likely mechanism or membrane disruption and cell death by amyloid peptides,
based on reports of single-channel conductance induced by amyloid peptides and by a number of
molecular modeling studies (Kayed et al. 2004; Jang et al. 2008). Based on recent reports that
fragmented amyloid fibrils exhibit significantly increased cytotoxicity relative to intact fibrils
(Xue et al. 2009), it is likely that the large non-fibrillar oligomers act as reservoirs of small fibril-
99
like segments capable of exerting a similar effect on cell membranes. Such a possibility may in
part explain the common action observed for large oligomers formed by several different
amyloid peptides (Kayed et al. 2003).
Based on the importance of non-fibrillar oligomers in the pathogenesis of amyloid diseases, it is
essential to develop a detailed understanding of their relationship to the amyloid fibrils that are
the hallmark of these diseases. While there has been evidence presented suggesting that soluble
oligomers may represent misfolding intermediates on the pathway to fibril formation (Auer et al.
2008; Frare et al. 2009), other studies suggest that they form via an off-pathway misfolding event
(Necula et al. 2007; Glabe, C. 2008). The presence of fibril-like structures in non-fibrillar
oligomers of PrP(106-126), as well as in large oligomers of Aβ, seems to suggest that structural
rearrangement of oligomers into mature amyloid fibrils may be possible in these systems,
although the inherent stability of the non-fibrillar PrP(106-126) assemblies implies a barrier to
this change in assembly. Thus, while it is clear that these two different oligomeric states are
surprisingly similar in local structure, the relationship between them remains to be determined.
Using well-defined systems such as PrP(106-126), in which stable fibrils and oligomers can be
prepared under varying solution conditions, detailed examination of amyloid misfolding
pathways should be possible.
100
5 Membrane Interactions of PrP(106-126) Oligomers
Solid state NMR and TEM experiments in this chapter were conducted by P.Walsh.
AFM/TIRF was performed by Gill Vanderlee in the laboratory of Dr. Chris Yip at the University
of Toronto, Department of Chemical Engineering and Applied Chemistry.
Cell toxicity studies were conducted by Jason Yau in collaboration with Valerie Sim at the
University of Alberta’s Centre for Prions and Protein Folding Diseases
101
5.1 Abstract
The formation of fibrillar aggregates has long been associated with neurodegenerative disorders
such as Alzheimer's and Parkinson's diseases. While fibrils are still considered important to the
pathology of these disorders, it is now widely understood that smaller amyloid oligomers are the
toxic entities along the misfolding pathway. One common characteristic between amyloid
systems is the ability of amyloid oligomers to disrupt membranes; a commonality proposed to be
responsible for their toxicity. This chapter describes the membrane interactions and toxicity of a
model amyloid peptide – PrP(106-126). This peptide forms amyloid fibrils, as well as non-
fibrillar oligomers which interact with model membranes causing vesicles to be removed from
simple, anionic containing lipid mixtures and cause a loss of lipid raft motifs in cholesterol
containing mixtures. Furthermore, we show that these oligomers are toxic to numerous cell lines
as well as rat cerebellar slices.
102
5.2 Introduction
The study of amyloids and their related toxicity is an important consideration in the growing
field of neurodegenerative disease. Disorders attributed to the accumulation of misfolded
fibrillar material include Alzheimer’s, Parkinson’s, and Prion diseases. Until recently, the
pathology of disease has been associated with the discovery of amyloid deposits within brain
tissue; however, amyloid oligomers are now considered to be the toxic entity present in most
neurodegenerative diseases associated with misfolded peptides or proteins. One commonly
proposed mechanism for the toxicity of these small, reactive amyloids is disruption of the plasma
membrane. There are a number of ways in which amyloid peptides can interact with the
membrane. One possible mechanism of disruption is pore formation. Pores were first thought to
be formed in membranes by Aβ peptide and have since been observed for a number of other
amyloid systems. Of these pores, it is proposed that peptides can either form a barrel-stave or a
torroidal pore in the membrane. The end effect of pore formation is loss of membrane potential,
as observed by single channel measurements as well as loss of solutes resulting in cell death.
Interestingly, annular protofibrils formed by Aβ1-42 are able to bind α-hemolysin antibodies
indicating that the β-barrel protein may contain a common structural element (Kayed et al. 2003).
It is also possible for peptides to disrupt membranes by behaving as a detergent. In this case,
peptides remove lipid molecules directly resulting in bilayer destabilization. Detergent-like
action by a peptide has been seen with human IAPP20-29 where small vesicle formation was
observed for a range of peptide concentrations (Brender et al. 2012). Similar to detergent action
of membrane destabilizing peptides is the carpet model of disruption whereby peptides aggregate
on the surface of the bilayer and cause general destabilization. Both carpeting as well as
detergent-like action are commonly seen with antimicrobial peptides which could share a general
mode of action with amyloid oligomers. Finally, it has been proposed that peptide fibrillization
on the surface of the membrane can lead to the formation of peptide raft-like structures inside the
bilayer, causing destabilization. The role of lipid rafts has been explored for some systems and
has yielded results which indicate that raft forming components such as cholesterol can protect
against membrane disruption or can lead to increased binding. It is well known that maintaining
lipid rafts, and the membrane proteins associated with them, is very important for in vivo cell
survival.
103
PrP(106-126), derived from the unstructured N-terminus of the full-length prion protein, is a
good model amyloid peptide. PrP(106-126) forms amyloid fibrils and oligomers, the latter of
which are toxic. This peptide has been shown to be toxic both in the presence and absence of
full-length PrP. It has been shown to interact with and disrupt model membranes, including the
binding and aggregation of GM1 ganglioside-containing membranes. Specific to membrane
disruption by PrP(106-126) oligomers, it was shown previously that these structures directly
cause membrane permeabilization. This peptide has also been shown to interact with L-type
voltage sensitive calcium channels (Thellung et al. 2000), cause changes in membrane viscosity
(Salmona et al. 1997), activate JNK-c-Jun pathway (Carimalo et al. 2005), forms channel pores
(Lin, M. C. et al. 1997; Kourie, J. I. et al. 2001) and is toxic to neuroblastoma cells (Ettaiche et
al. 2000).
104
5.3 Materials and Methods
5.3.1 Preparation of PrP(106-126) Non-fibrillar Oligomers
Non-fibrillar oligomers of PrP(106-126) were prepared as described in chapter 3. Fibrils were
prepared as described in chapter 2. Scrambled PrP(106-126) was obtained from the Advanced
Protein Technology Centre at the Hospital for Sick Children and was dissolved at a concentration
of 1 mg/ml in 10 mM acetate buffer pH 4.6.
5.3.2 Formation of Large Unilamellar liposomes
To form LUVs of POPC or 3:1POPC:POPG, appropriate amounts of 25 mg/ml lipid stocks in
chloroform (Avanti polar lipids) were measured and dried to a film under a stream of nitrogen.
This film was then taken up in water at a concentration of 25 mg/ml and lyophilized. The freeze-
dried lipids were then resuspended in 20mM HEPES buffer pH 7.4, freeze-thawed ten times.
Samples were then either used for AFM studies or extruded through a 0.4 µm filter membrane
for analysis by solid state NMR. For the formation of 1:1:1 DOPC:DSPC:cholesterol liposomes,
appropriate amounts of DOPC and DSPC in cholesterol were mixed with cholesterol and dried to
a film. This film was then resuspended as described above. In order to extrude the lipid
suspension, the mixture was heated to 70 °C for 20 minutes then allowed to return to room
temperature where they were either analyzed by AFM or extruded for solid state NMR analysis.
5.3.3 Transmission Electron Microscopy
For transmission electron microscopy, 25 mg/ml suspension of freshly extruded lipids was
subjected to either 120 µM peptide oligomers or 10mM sodium acetate pH 4.6. These samples
were then diluted 500 times, 4 µl of which was spotted on 400 mesh continuous carbon grids
which were previously glow discharged for 15 s at 30 mA negative discharge. Samples were
adsorbed for 2 minutes before blotting, rinsing twice with water and a final staining with 2%
uranyl for 15 s. Images were acquired using a Jeol 1011 microscope operating at a voltage of 80
kV.
5.3.4 Atomic Force Microscopy
Images were acquired in fluid tapping mode with a Digital Instruments (Veeco, Santa Barbara,
California, USA) Nanoscope IIIa Multimode AFM equipped with an “E” scanner (maximum
105
lateral scan area 14.6 μm x 14.6 μm), using SNL-10 short, thin tips (Veeco Probes, Camarillo,
CA and Bruker AFM Probes, Camarillo, CA). A contact/tapping mode glass fluid cell was sealed
against a freshly cleaved muscovite mica substrate with a silicone O-ring. The fluid cell, having
a volume of approximately 200 μL, was fitted with separate inlet and outlet tubing to allow for
the exchange of fluid during imaging. All images were collected at a resolution of 512 x 512
pixels at scan rates between 1-3 Hz using tip oscillation frequencies of approximately 8 or 25
kHz. Image analysis was performed with Nanoscope software (version 5.12r3, Digital
Instruments). Images were typically subjected to zero order flattening and second order plane fit
(x-axis) filters.
Mica surfaces were pretreated by filling the fluid cell with 10 mM HEPES containing 150 mM
NaCl at pH 7.4. Supported planar bilayers were formed by injecting approximately 300 μL of a
lipid vesicle suspension (typically composed of 100 μL 1 mM lipid stock and 300 μL 10 mM
HEPES containing 150 mM NaCl at pH 7.4).
PrP(106-126) oligomers were diluted in acetate buffer at pH 4. Approximately 300 μL was
injected into the fluid cell, enough to completely replace the fluid volume of the cell. AFM
images were collected until no significant changes in the bilayer were detected (approximately 1
hour). The fluid cell was flushed with at least 300 μL 10 mM HEPES containing 150 mM NaCl
at pH 7.4 and imaged for at least 30 minutes. Addition of peptide and subsequent wash could
then be repeated at higher peptide concentrations.
5.3.5 AFM-TIRF
A thin slice of V1 grade muscovite mica was cleaved from 2.5 cm round discs. Mica was secured
to the bottom of a glass Wilco-dish with UV-curable adhesive (Norland Optical Adhesive 63,
Norland Products, Cranbury, NJ).
For lipid mixtures containing anionic lipids, dishes were pretreated with 2 mL of 20 mM CaCl2
for several minutes and subsequently removed. 100 μL of 1mM stock lipid, 1mol% Dil
(fluorescent probe) and 1900 μL 10 mM HEPES containing 150 mM NaCl at pH 7.4 were added
to dishes. For lipid mixtures containing lipids with transition temperatures above room
temperature, dishes were incubated at approximately 70°C for 20 minutes and subsequently
allowed to cool back to room temperature. If excessive amounts of vesicles were observed under
106
TIRF illumination, 1 mL aliquots of buffer were exchanged from the dish as needed. If a
continuous bilayer was not present, 100 μL aliquots of 1 mM lipid stock were exchanged for 100
μL buffer from the dish as needed.
Images were acquired in fluid tapping mode with a Digital Instruments (Veeco, Santa Barbara,
California, USA) Nanoscope IIIa Bioscope AFM using SNL-10 short, thin tips (Veeco Probes,
Camarillo, CA and Bruker AFM Probes, Camarillo, CA). The AFM head was positioned in a
vertical slot above an objective based TIRF system. AFM image analysis was performed with
Nanoscope software (version 5.30r3, Digital Instruments). AFM images were typically subjected
to zero order flattening and second order plane fit (x-axis) filters.
A modified commercial Olympus Fluoview 500 (FV500) microscope that accommodates
multiple excitation laser lines was utilized. The bottom of the dish was brought into focus under
ambient lighting under oil immersion using a 60X TIRF objective lens. Appropriate filters were
then inserted and a region of the supported bilayer was brought into focus under TIRF
illumination. Images were captured with an Evolve 512 EMCCD camera (Photometrics, Tucson,
AZ) controlled by Micro-Manager (Vale Lab, USFS, CA). Fluorescent probes were excited by
parallel (s) or perpendicular (p) polarized light through the rotation of a half wave plate.
5.3.6 Solid State NMR of Liposomes
Solid state NMR experiments were carried out using a Varian VNMRS spectrometer operating at
a 1H frequency of 499.76MHz. Static
31P NMR analysis was done using a 4mm T3 static probe
with a one-pulse 50kHz 31
P field and 50kHz 1H decoupling. Spectra were processed using
NMRPipe and visualized using nmrDraw. For each static spectrum 200Hz of Gaussian line
broadening was applied.
5.3.7 Brain Slice and Cell Culture
Cerebellar slice cultures were prepared from 10-12 day old C57Bl6 mice, as described
previously (Falsig and Aguzzi 2008). Slices were cultured for 14 days prior to treatment, to
allow cultures to stabilize. A single insert with 2-3 cerebellar slices was treated apically with 200
µl of warm slice medium containing 10 µg/ml propidium iodide (PI) (Invitrogen), for 15 min in a
standard cell incubator (37 °C, 5% CO2 and 95% humidity). Inserts were removed and placed
into a new 24 well plate and washed three times both apically and basolaterally with 500 µl room
107
temp PBS pH 7.4 to remove residual PI. Inserts were fixed using fresh 4% PFA (Invitrogen) in
PBS pH 7.4 for 15 min in the dark. Inserts were washed three times both apically and
basolaterally with 500 µl room temp PBS pH 7.4 to remove fixative. Inserts were permeabilized
with 0.25% tritonX-100 in PBS pH 7.4 for 15 min in the dark. Inserts were washed three times
both apically and basolaterally with 500 µl room temperature PBS pH 7.4. Inserts were blocked
using 1% goat serum, 1% BSA in PBS pH 7.4 for 1 hr in the dark. Blocking buffer was removed
and inserts were treated with 1:4000 dilution of anti-mouse calbindin (AB Cam) in blocking
buffer for 1hr in the dark. Inserts were washed three times both apically and basolateraly with
500 µl room temp PBS pH 7.4. Inserts were treated with 1:4000 Anti-goat alexafluor 488
secondary antibody (Invitrogen) for 30 min in the dark. Inserts were washed three times both
apically and basolateraly with 500 µl of room temperature PBS pH 7.4. The membrane was
removed from the insert support and placed on a slide with the apical surface of the tissue up. 3
drops of Prolong gold with DAPI (Invitrogen) were placed on the membrane insert and a
coverslip was affixed to the slide. Slides were cured minimum 24 hrs.
PC12 (rat adrenal pheochromocytoma) cells were cultured in F12K nutrient media with 10%
horse serum, 5% FBS, penicillin, streptomycin, glucose, and sodium pyruvate; N2a (mouse
neuroblastoma) and SHSY-5Y (human neuroblastoma) cells were cultured in DMEM high
glucose supplemented with 5% FBS penicillin, streptomycin, glucose, and sodium pyruvate.
Prior to treatment, cells were changed to low serum media, containing 1% total serum (FBS) and
plated at 20-30% confluence per well in a collagen-coated 96 well plate.
Determination of slice viability by propidium iodide staining was done from images taken on a
Zeiss LSM 700. Images were deconvoluted using ImageQuant X, and analysis was completed
using Imaris 7.1.1 software.
108
Results
5.3.8 PrP(106-126) Oligomers Disrupt Anionic Lipid Bilayers
Planar 3:1 POPC:POPG bilayers were imaged using AFM in tapping mode, in solution. These
bilayers appear as a single phase and are shown in Figure 5-1. Upon addition of peptide, we see
the formation of raised structures at a height of approximately 4.5 nm above the bilayer surface.
Figure 5-1 – AFM of 3:1 POPC:POPG Supported Bilayers
AFM of 3:1 POPC:POPG + PrP(106 -126). A) Original bilayer is relatively flat. B) After addition of 7.8 µM peptide
solution. Peptide appears to adhere to defects in the bilayer. After subsequent washing with HEPES pH 7.4 buffer,
pits appear in the bilayer as shown in C. C) After addition of 15.6 µM peptide solution, peptide again adheres to
bilayer as well as small fibrillar deposits beginning to form on the mica surface in the previous formed pits.
The appearance of these deformations on the bilayer is consistent with the formation of small
vesicles being released from the bilayer, a phenomenon previously seen by the interaction of
IAPP with liposomes (Brender et al. 2012) and Aβ1-40 with supported bilayers (Yip and
McLaurin 2001). Vesicle release is confirmed as, after washing, the bilayer is depleted at
locations previously corresponding to increased height. These areas are lower by approximately
6 nm relative to the bilayer. Vesicle release was also confirmed by TEM, shown in Figure 5-2,
where round vesicles can be seen blebbing from the surface of LUVs.
109
Figure 5-2 – Negative Stain TEM of Large Unilamellar Vesicles
Negative stain TEM of 100 µm extruded liposomes. A) Untreated LUVs appear as spherical shapes with round
edges. B) After addition of 120 µM solution of PrP(106-126) oligomers surface blebbing can be seen. C) An
additional view shows multiple LUVs with surface disruptions, suggesting the release of small vesicles from the
surface of the LUV. Scale bar is equivalent to 100 nm.
To examine the changes in the bilayer on a molecular level, we utilized 31
P static solid state
NMR experiments of 400 nm large unilamellar vesicles (LUVs). The static spectra in Figure 5-3
show a broad powder pattern for 3:1POPC:POPG liposomes alone. Upon addition of 130 µM
PrP(106-126) oligomers, there is a marked reduction in line width with the narrower peak
approaching the 31
P isotropic chemical shift. The result of reduced line width in static spectra
can be attributed to the breakdown of the membrane or release of small, fast tumbling vesicles
from the larger 400 nm vesicles. The narrowing of the 31
P line indicates an averaging or
reduction in the 31
P CSA due to the fast-tumbling small vesicles. This release of vesicles and
subsequent narrowing of 31
P powder patter has been previously seen for amyloids, most notably
IAPP (Brender et al. 2012).
110
Figure 5-3 – 31
P Static NMR spectra of anionic large unilamellar vesicles
Static NMR spectra showing 3:1POPC:POPG LUVs in the absence (black) and post-treatment with 130 µM
PrP(106-126). Upon exposure to PrP(106-126) oligomers, the breadth of the powder pattern (width at the widest
point) is reduced from 53.4ppm to 26.4ppm.
5.3.9 PrP(106-126) Causes Loss of Lipid Domain Order in Cholesterol-Containing Bilayers
To determine the effect of cholesterol on the disruptive effect of PrP(106-126) oligomers, we
used supported bilayers comprised of 1:1:1 DOPC:DSPC:cholesterol. Figure 5-4 shows AFM of
a cholesterol-containing bilayer showing higher cholesterol domains. The fluorescent molecule
Dil partitions into the more ordered cholesterol-containing domain of the supported bilayers
(Spink et al. 1990) where alignment with ordered cholesterol and DSPC causes the probe to
become fluorescent under polarized light (Korlach et al. 1999). Upon addition of peptide, there
is a loss of this height difference, indicating a loss of domain structure. Furthermore, fibrilization
can be seen on and around areas where cholesterol domains are disappearing.
111
Figure 5-4 - AFM of 1:1:1 DSPC:DOPC:Cholesterol Supported Bilayers
A) Original bilayer is phase separated with phases typically having 1-2 nm difference in height. B) After addition of
7.8µM peptide solution. Phase separation has radically disappeared or changed all together and PrP(106-126) fibrils
appear. C) After addition of 20.8µM peptide oligomers fibers have become more dense and grown in length.
Unlike with anionic membranes, we do not see any vesicle formation at the surface of the
membrane or associated pitting of the membrane. This is also shown in the 31
P static spectra in
Figure 5-5. The similar powder pattern breadth indicates that there are no changes in the size of
the LUVs. Furthermore, Figure 5-6 shows TIRF images and the resulting changes in order
parameter of cholesterol containing bilayers. The loss of domain structure can be seen both in the
AFM as well as fluorescence of the fluorescent probe. Prior to peptide addition, 2 distinct
domains can be seen; after peptide addition, there are no domain structures visible.
112
Figure 5-5 – Static 31
P Spectra of cholesterol-containing LUVs
Static NMR spectra of 1:1:1 DOPC:DSPC:cholesterol LUVs showing a similar breadth of powder pattern between
untreated and LUVs exposed to 120 µM PrP(106-126) oligomers.
113
Figure 5-6 – Polarized TIRF and AFM Images of 1:1:1 DOPC:DSPC:Cholesterol
Shown are parallel excitation( Fs) and perpendicular excitation (Fp) of the fluorescent probe dil and AFM images
prior to and after the addition of PrP(106-126) to a final concentration of 4.1µM on 1:1:1 DSPC:DOPC:Chol. A
bilayer with distinct lipid domains is easily visualized in the AFM and parallel excitation images before addition of
peptide. After addition, domain structure is disrupted.
5.3.10 PrP(106-126) Oligomers are Cytotoxic to Cultured Cells
To assess the reported toxicity of PrP(106-126) oligomers, we treated a number of different cell
lines with various concentrations of PrP(106-126) oligomers, fibrils and scrambled peptide.
Figure 5-7 shows the toxicity of PrP(106-126) oligomers using the toxilight assay after 24hrs of
incubation with N2a Tim (Fig 5-7A), N2a C16 (Fig 5-7B), PC-12 (Fig 5-7C) and SH-SY5Y (Fig
5-7D) cells. The toxilight assay colorometrically measures the release of adenylate kinase from
cells as an indication of cell death (Miret et al. 2006). In all cases, 100 µM PrP(106-126)
oligomers are sufficient to cause cell death at a similar level to that of the positive control,
staurospaurine.
114
Figure 5-7 – Toxilight cell-death assay
Toxilight cell death assay results are shown for A) N2A Tim, B) PC12, C) SHSY-5Y and D) N2A C16 Cells. In all
cases, 100 µM and 50µM PrP(106-126) was able to produce statistically significant cell death versus buffer alone
(*** corresponding to 99% confidence, **97%, * 95% confidence).
To further confirm toxicity, we employed the use of an MTS reduction assay. After 48hrs of
exposure to PrP(106-126), cell viability is reduced to levels corresponding to the positive control
at 100 µM PrP(106-126). N2a Tim (Figure 5-8A), N2a C16 (Figure 5-8B), PC-12 (Figure 5-8C)
and SH-SY5Y (Figure 5-8D) cells are all sensitive to PrP(106-126) oligomers. This assay
measures a cell’s ability to reduce the tetrazolium salt MTS into a water-soluble formazan
product which is measured spectrophotometrcially (Buttke et al. 1993). Decreasing
concentrations of PrP(106-126) causes an increase in cell viability, with scrambled PrP(106-126)
as well as fibrils not causing significant decreases in the cells’ ability to reduce MTS.
115
Figure 5-8 – MTS Reduction Assay
The ability of the 4 different cell lines are quantified in panels A-D. A decrease in a cells ability to reduce MTS is
indicative of cell death and is represented as a percentage reduction/cell survival. In each cell line, 100 µM and 50
µM PrP(106-126) oligomers are sufficient to cause cell death at the same level as the staurosporine control.
116
5.3.11 Oligomers of PrP(106-126) are Toxic to Rat Cerebellar Brain Slices
The introduction of PrP(106-126) oligomers to cultured rat brain slices caused significant cell
death. Figure 5-9 shows the effect of PrP(106-126) oligomers on rat brain slices where no
peptide reveals neuronal cells with calbindin (a protein responsible for calcium binding and
release for proper function in the cerebellum) and nuclei present (Figure 5-9A). After 24hrs
treatment with PrP(106-126), there is propidium idodide staining present, indicating release of
nuclear DNA in 55% of cells. After 48hrs treatment with PrP(106-126) 61% of rat cerebellar
cells have died.
Figure 5-9 – Exposure of rat cerebellar slices to PrP(106-126) oligomers
Confocal microscope images of rat cerebellar slices in A) the absence of PrP(106-126) oligomers B) 24hrs after
treatment with 100 µM PrP(106-126) and C) 48hrs after treatment. Green stain represents the protein calbindin, blue
is the nuclear stain DAPI. Priopidium iodide staining is indicated by white dots, revealing the release of nuclear
DNA.
117
5.4 Discussion
The formation of non-fibrillar oligomers is an important aspect in the toxicity of amyloid
diseases. These spherical structures are considered to be the toxic species along the misfolding
pathways associated with amyloid diseases while mature fibrils are considered to be somewhat
benign. In chapters 3 and 4, the molecular structure of amyloid oligomers formed by PrP(106-
126) was examined. This was then compared to the structures examined in chapter 2 and it was
determined that these two structures share a similar subunit comprised of parallel β-sheet stacked
in an antiparallel arrangement. Also in chapter 3, it was shown that PrP(106-126) oligomers have
the ability to disrupt anionic lipid bilayers as determined by a liposome dye-release assay.
In an effort to explore the ability of PrP(106-126) oligomers to disrupt membranes, we examined
their interactions with both anionic and zwitterionic cholesterol-containing bilayers using solid
state NMR as well as AFM and TIRF. We then correlated these interactions with toxicity studies
of cultured mammalian cells. Our results show different modes of membrane disruption for the
two lipid mixtures used to conduct biophysical experiments. In the first case, we used an anion
containing lipid mixture of 3:1POPC:POPG and observed the conversion of large vesicles into
much smaller ones indicating the ability of PrP(106-126) oligomers to act in a detergent-like
manner. This work is the first detailed report of membrane disruption by the peptide PrP(106-
126) and follows a similar report of IAPP membrane disruption using solid state NMR (Brender
et al. 2012). The overall determination that small vesicles are being produced from larger ones
was found by visualization through both AFM and TEM. This correlates well with an overall
reduced size of vesicles as seen by a narrowing in the NMR powder patter for 31
P which has been
previously reported for the peptide IAPP when disrupting anionic membranes (Brender et al.
2008). The disruption of anionic membranes is a common mechanism of antimicrobial peptides,
some of which have been shown to act as a detergent to solubilize membranes.
While using a more mammalian-like lipid mixture containing equal parts DOPC, DSPC and
cholesterol we observed that the bilayer remained intact, however, there was a distinct loss of
domain structure. We then showed that PrP(106-126) oligomers are toxic to a number of
different mammalian cell lines including rat cerebellar slices. The loss of cholesterol domain
structure along with the ability of PrP(106-126) oligomers to kill cells points to the possibility of
a cell-surface cholesterol domain reorganization as the cause for cell death.
118
The ability of PrP(106-126) oligomers to cause a loss of domain structure in cholesterol-
containing membranes is the first report of such activity by an amyloid peptide. The loss of
cholesterol domains can be seen through both AFM as well as TIRF, where addition of the
peptide is either causing the formation of a single domain or causing the sequestering of
cholesterol domains. It has been previously reported that cholesterol may be protective against
disruption by amyloid oligomers of Aβ1-42 by preventing the oligomers from sequestering
ganglioside GM1 at the cell surface (Cecchi et al. 2009). Similar results were demonstrated for
the Sup35 system where fibrils binding to GM1 on the surface of live cells were able to induce
caspase activity due to the rearrangement of cell-surface Fas (Bucciantini et al. 2005) While we
have not yet examined the interaction of PrP(106-126) oligomers with GM1, this peptide is
clearly able to interact with and disrupt lipid rafts in a non-ganglioside associated manner.
PrP(106-126) has previously been shown to be toxic in a number of ways such as by increasing
membrane microviscosity (Salmona et al. 1997) and causing apoptosis (Florio et al. 1998; Silei
et al. 1999; Thellung et al. 2000). While we are not the first to describe the toxicity of the peptide
PrP(106-126), this work is the first comprehensive study to show the specific toxicity of
PrP(106-126) non-fibrillar oligomers versus fibrils since 2003 (Kayed et al. 2003). Other groups,
including those describing the apoptotic-activating ability of PrP(106-126) do not specifically
describe the aggregation state of the peptide, which could be one of the factors leading to
conflicting reports of cell toxicity. The hypothesis that the disruption of cholesterol domains as
the cause of toxicity of PrP(106-126) comes from the direct observation of the loss of cholesterol
domains by biophysical methods. However, given that PrP(106-126) has previously been shown
to interact with the ganglioside GM1 (Kurganov et al. 2004), we cannot rule out this interaction
as a cause for toxicity as it is with Aβ peptides (Choo-Smith and Surewicz 1997). Differing
modes of membrane disruption based on lipid composition raises the issue that disruption in
amyloid systems can be greatly affected by the lipid mixtures used in experiments. For example,
POPC (Comellas et al. 1021), DOPG (Lu et al. 2011), DMPG:DMPC (Smith et al. 2009) are all
examples of lipid compositions used to study membrane interactions with amyloid proteins and
peptides. In the case where total membrane disruption is not observed, such as the cholesterol
containing bilayers presented here, one must still be on the lookout for other factors pertaining to
the membrane that can affect cellular toxicity.
119
6 Summary and Future Directions
6.1 Summary
The overall goal of this thesis work is to understand the molecular basis for amyloid oligomer
versus fibril toxicity for the peptide PrP(106-126). In this thesis I describe structural aspects of
the model amyloid peptide PrP(106-126), the first being the molecular model of amyloid fibrils.
The second aspect was the structure of amyloid oligomers formed by the peptide, and how they
related to the fibrils. Finally, I set out to determine the interaction of PrP(106-126) oligomers
with model membranes and how these interactions related to toxicity.
In chapter 2, the structure of the core of amyloid fibrils formed by PrP(106-126) was determined
using a combination of solid state NMR, TEM, CD and AFM. Specifically, I showed, through
the use of 13
C-13
C dipolar recoupling experiments, that fibrils of this peptide are arranged in
parallel β-sheets. Utilizing qualitative dipolar recoupling, I showed that these parallel β-sheets
are stacked in an anti-parallel arrangement – an example of a class-I steric zipper. This work
was one of the first structural models of a fibril core published.
In chapter 3, the morphology and secondary structure of stable β-oligomers of PrP(106-126) was
shown. The overall shape of these non-fibrillar oligomers is approximately 30 nm by 12-15 nm
which was determined by double-carbon TEM as well as dynamic light scattering. Both CD and
NMR chemical shift analysis were used to determine the secondary structure of these oligomers,
each showing that the oligomers contain significant β-sheet content. Finally, the ability of these
peptide oligomers to act on model membranes was examined in chapter 3; a consideration which
is very important when considering amyloid systems as most amyloid peptides are membrane
active. It was demonstrated that PrP(106-126) oligomers have very potent membrane disrupting
ability while α-helical monomers, random coil monomers and mature fibrils were benign. These
results lend support to the idea that amyloid oligomers and not fibrils are the toxic entity.
Chapter 4 shows the relationship between fibrils and oligomers through the examination of the
structures formed by PrP(106-126) oligomers. Through the use of solid state NMR, I showed that
the basic subunit of these large oligomers is a parallel β-strand, stacked antiparallel with itself;
meaning that the basic subunits of the fibril and oligomer are identical. Molecular insights of
structured, monomeric peptide in exchange with large oligomers of PrP(106-126) were also
120
gained using solution NMR with the chemical shifts determined using solution and solid state
NMR both being in good agreement. This is the most detailed structural characterization of a
large non-fibrillar oligomer complex to date. The monomers, oligomers and fibrils having the
same subunit provide a basis for the conversion of oligomers to fibrils directly or by the
formation of structured monomers to fibrils.
Finally, chapter 5 is a collaborative effort to explain the toxicity of PrP(106-126) oligomers
through the use of model membranes and biophysical techniques. Through collaboration with the
Sim Lab at the University of Alberta, it was demonstrated that oligomers of PrP(106-126) are
toxic to four different cell lines – N2A Tim, N2A C16, PC-12 and SHSY-5Y cells. In order to
gain insight into a more relevant mammalian model membrane, we employed the use of a
cholesterol-containing composition. In this case, the membrane did not break down into vesicles
as shown by solid state NMR; however, AFM shows the loss of cholesterol domain structure
along with fibrilization at the surface of the membrane. This loss of domain structure is one
potential explanation for the cytotoxicity of the peptide and could also explain a lack of total
membrane disruption by amyloid peptides on bilayers containing cholesterol.
As a whole, the results presented in this thesis provide the structure of amyloid fibrils and
oligomers of a model amyloid system. The structural similarity in the basic subunit shared
between the fibril and oligomer is contrasted by the two entity’s differing toxicity. In line with
the accepted hypothesis that amyloid oligomers are toxic while fibrils are not, this work provides
a basis for the determination of toxicity based on membrane interactions. Furthermore, it brings
the importance of membrane composition into the forefront of the discussion on membrane
disruption as changing the type of lipids used has been shown to have drastic effects on the mode
of disruption.
121
6.2 Future Directions
6.2.1 Continuing Studies on PrP(106-126) Oligomer Interactions with Lipid Bilayers
In order to best compare cellular toxicity with biophysical membrane studies, increasingly
complex mixtures of lipids should be examined in order to eventually achieve a system that more
closely resembles neuronal cells. In previous studies of the amyloid peptides Aβ1-40, Aβ1-42,
amylin and PrP(106-126), GM1 has been shown to play an important role in peptide aggregation
and toxicity (Kurganov et al. 2004). Therefore, studies should be employed on cholesterol-
containing bilayers with the addition of the ganglioside GM1 utilizing AFM and solid state
NMR. Specifically, one should look for changes in the ganglioside containing bilayer after the
addition of peptide with respect to the non-ganglioside bilayer described in chapter 5. Since it
was demonstrated that GM1 clusters allow for increased aggregation of Aβ (Ikeda et al. 2011), it
will be interesting to know whether GM1 will provide the same aggregation center for PrP(106-
126). This means that fibrilization events must be monitored at the surface of the membrane,
which can be achieved using Thioflavin-T detected by TIRF-AFM.
In order to gain additional information about the effect of PrP(106-126) oligomers on lipid
bilayers, additional solid state NMR experiments can be performed. Specifically, the focus on
position-specific phospholipid acyl chain order parameters can be examined using 2H solid state
NMR. In this set of experiments, lipid molecules deuterated at each acyl chain position (in this
case perdeuterated DSPC) are used to make extruded liposomes of 1:1:1
DOPC:DSPC:cholesterol. The powder pattern that results from deuterated lipids is a symmetric
set of peaks, separated by the quadrupolar coupling constant between adjacent deuterons. Static
spectra are taken before and after the addition of the peptide and the resulting pattern are
deconvoluted by de-Pake-ing (Schäfer et al. 1995). The analysis of each peak from the deuterons
allows for the determination of each quadrupolar coupling constant which is affected by the
molecular motions of the lipid acyl chain and therefore reflecting the order parameter associated
with each acyl position. Furthermore, 2H NMR can be used to assess the change in order of the
cholesterol in liposomes directly by utilizing cholesterol containing a single deuterated site. As
with chain deuterated DSPC the change in order parameter of a single deuteron on cholesterol
can be directly observed by comparing spectra before and after the addition of peptide. While the
122
signal-to-noise will be low for a 33% cholesterol sample containing a single 2H labeled site, the
experiment will directly correlate any change in order upon addition of peptide oligomers. By
measuring the change in order parameters of cholesterol, it can be determined whether the
cholesterol is becoming more or less ordered upon addition of PrP(106-126) oligomers. This will
allow for the determination of whether cholesterol is forming a mixed phase with DOPC or is
being sequestered into one large, cholesterol domain.
In order to determine how PrP(106-126) oligomers interact with phospholipid bilayers, it will be
important to continue studies on the phospholipid head groups of lipid molecules in bilayers of
various compositions. The first of these continued experiments will be to perform 31
P MAS
studies to gain information on local chemical environment changes of the phosphorus in the lipid
head group. The chemical shift of 31
P in lipid head groups has previously been shown to be
highly sensitive to changes in local environment, especially through electrostatic interactions
(Auger, 2010). Samples will be analyzed under MAS, with 31
P chemical shift being measured
and compared with addition of peptide. MAS studies can also be utilized to determine the effect
of the lipid chain by utilizing 13
C chemical shift analysis. Changes in chemical environment of
the lipids associated with the addition of peptide oligomers can be monitored by measuring the
chemical shifts with and without peptide (Hong 2006). This, in conjunction with 2H studies of
perdeuterated lipids and cholesterol, should provide details into any changes in order, motions
and chemical environment of lipid, the lipid acyls chains, and cholesterol molecules in the
bilayer.
Finally, the diffusion of lipid molecules through the liposome can be determined by utilizing
specialized solid state NMR experiments. Through the use of the CODEX (centre-band-only-
detection-of-exchange), it has been demonstrated that the lateral diffusion of lipid molecules can
be directly observed (Saleem et al. 2012). In order to determine the diffusion of a lipid molecule
through the liposome, the size distribution of the liposome must be known, as well as the
diffusion of liposomes through the solution. To do this, DLS and pulse-field gradient diffusion
measurements are used. Afterwards, the signal decay (due to recoupling of 31
P CSA) is measured
and the rates of lateral diffusion are calculated utilizing the liposome size and diffusion through
the solution (Saleem et al. 2012). This will give insights into the change in the average lateral
diffusion after addition of PrP(106-126) oligomers allowing for the determination of whether the
123
phospholipids have increased or decreased mobility. Given that cholesterol domains are
relatively rigid, a decrease in order will appear as a change in mobility.
6.2.2 PrP(106-126) Structures Formed in the Presence of the Bilayer
Since PrP(106-126) oligomers are cytotoxic and have membrane-changing behavior it would be
interesting to examine what structural conformers are adopted by this peptide after interaction
with the lipid bilayer. For instance, it can be determined if there are any preferential interactions
with lipid head groups and peptide residues by conducting 31
P-13
C transfer experiments. The
most important experiment to conduct would be to examine GAVL labeled peptide after
liposome treatment to determine if the stacked parallel β-sheet contacts are maintained during the
liposome disruption process. These experiments will be the same as those conducted in chapters
2 and 4 including 13
C RAD and PITHIRDS and could also include the use of 13
C-15
N dipolar
mediated experiments such as rotational-echo double-resonance (REDOR) (Gullion and Schaefer
1989) and transferred-echo double-resonance (TEDOR) (Hing et al. 1992). These 2 additional
experiments allow for through-space N-C assignments for structural restraints. In the case of the
detergent-like membrane disruption model, the simplest structures formed in the peptide-lipid
micelle would be the basic oligomer/fibril subunit. A smaller oligomer formed by parallel β-
sheets stacked antiparallel could form the base where hydrophobic interactions between the β-
sheet core could be made with the lipid acyl chains. Anionic head groups could also interact
electrostatically with the charged residues of the peptide. Since there is evidence that cholesterol-
containing bilayers cause fibrilization events, it would be necessary to examine if these fibrils
contain the same structural elements as both the previously structured fibrils and oligomers.
While the fibrils may have the same structure, since they would be likely propagated from seeds
containing that structure (oligomer fragments), other groups have shown that different conditions
can cause various fibril morphologies and structures (Paravastu et al. 2009).
In both the case for anionic as well as cholesterol containing lipids, it is important to examine
solvent exposure of residues after addition to liposomes. The use of paramagnetic ions such as
Mn2+
can yield information on the solvent exposure of residues. This technique has previously
been used to examine depth of peptide insertion into the bilayer (Su et al. 2008) and can also be
used to determine which residues are exposed to bulk solution in either the detergent model of
disruption (as with the addition of peptide oligomers to 3:1 POPC:POPG) or if fibrils
124
preferentially bury specific residues into the bilayer in a cholesterol containing disruption model.
In the experiment, labeled residues exposed to Mn2+
experience relaxation enhancement causing
their signal to be reduced as a result of their shortened transverse relaxation. Therefore, we can
compare the 1-dimensional or 2-dimensional spectra of labeled peptide once added to the bilayer
in the presence and absence of Mn2+
; any loss in signal is an indication of the exposure of the
corresponding residue to the bulk solvent.
6.2.3 Additional Toxicity Studies
While the use of rat cerebellar slices is of particular interest since it maintains synaptic
connections, exploring the effect PrP(106-126) oligomers have on cells is of great use in
determining its exact mode of toxicity. Since it was determined that cholesterol-containing
domains are lost or rearranged, it would be of interest to track cholesterol rafts by confocal
microscopy in live cells which has previously been used to track the domain structures that form
with differing lipid compositions (Korlach et al. 1999). Applying this technique, whereby dye is
incorporated into lipid raft motifs and imaged using a confocal fluorescence microscopy, should
give information on how oligomeric PrP(106-126) is affecting lipid raft integrity and
morphology in live cells. Since it has been shown that proteins contained in lipid rafts play an
important role in apoptosis, it should be determined whether the addition of PrP(106-126)
oligomers causes a change in membrane protein localization. For example, the protein Fas has
been shown to cause the activation of apoptotic pathways when it is removed from cholesterol-
rafts (Cahuzac et al. 2006; Gajate et al. 2009; Castro et al. 2011). To examine the effect of
PrP(106-126) oligomer-induced lipid domain loss, one could expose cells to PrP(106-126)
oligomers then track the cell surface Fas using anti-FAS antibodies and fluorescence confocal
microscopy. If Fas is changing location from an associated domain or self-associating, one
would have the ability to directly visualize this. One of the downstream effects of Fas triggered
cell-death is the activation of caspases 8 and 10 (Ashkenazi 2008). The expression levels of these
proapoptotic markers could be monitored simply by Western blot. While this idea of membrane
bound Fas moving into another lipid phase may be more complicated than this summary
suggests, it provides a good starting point for studies associated with determining the method of
toxicity of cells containing cholesterol. By looking for cell surface apoptosis activation, we can
correlate the observed biophysical effect with the cytotoxic properties of PrP(106-126)
oligomers. To look for additional changes in overall protein expression, the use of mass
125
spectrometry-based proteomics can be utilized. Control and oligomer-exposed cells can be
exposed to proteolytic digestion followed by ultra-high performance liquid chromatography
(UHPLC) with subsequent mass spectrometry for identification of peptide fragments (Altelaar et
al. 2013). New advances in quantitative mass spectrometry allow for the identification of
increases or decreases in peptide fragment levels which is proportional to protein expression
levels. Proteins showing changes in expression can then be validated by Western blot and
confocal microscopy with antibody tagging. Given the common structures and proposed modes
of cytotoxicity observed across amyloid proteins, the use of the model PrP(106-126) peptide
provides a good step toward understanding the role of amyloid oligomers in neurodegenerative
disease. If we can determine one or more of the causes of cytotoxicity, we could have the ability
to relate these to other amyloid diseases such as Alzheimer’s or Parkinson’s.
126
6.3 Final Conclusions
The results presented in this thesis show that cytotoxic amyloid oligomers of the peptide
PrP(106-126) share a common structure with that of the non-toxic fibrils. The structural
similarities between fibrils, oligomers and structured monomers in exchange with oligomers
provide the basis for describing how these structures may form along the amyloid misfolding
pathway. The membrane disrupting abilities of PrP(106-126) non-fibrillar oligomers through two
distinct methods provides a starting point to relate the demonstrated toxicity to mammalian cells,
starting with the loss of cholesterol domains. This work only breaks the surface of the problem
that is amyloid misfolding and toxicity, but lays the groundwork for the continued study of the
PrP(106-126) peptide system as well as other, as-yet-to-be characterized systems.
127
References
Ahmed, M., Davis, J., Aucoin, D., Sato, T., Ahuja, S., Aimoto, S., Elliott, J. I., Van Nostrand, W.
E. and Smith, S. O. (2010). Structural conversion of neurotoxic amyloid-beta(1-42)
oligomers to fibrils. Nature Structural and Molecular Biology, Vol. 17, No. 5, pp. 561-7.
Altelaar, A. F., Munoz, J. and Heck, A. J. (2013). Next-generation proteomics: towards an
integrative view of proteome dynamics. Nat Rev Genet, Vol. 14, No. 1, pp. 35-48.
Altieri, A. S., Hinton, D. P. and Byrd, R. A. (1995). Association of Biomolecular Systems via
Pulsed Field Gradient NMR Self-Diffusion Measurements. Journal of the American
Chemical Society, Vol. 117, No. 28, pp. 7566-7567.
Ashkenazi, A. (2008). Directing cancer cells to self-destruct with pro-apoptotic receptor
agonists. Nat Rev Drug Discov, Vol. 7, No. 12, pp. 1001-12.
Auer, S., Meersman, F., Dobson, C. M. and Vendruscolo, M. (2008). A generic mechanism of
emergence of amyloid protofilaments from disordered oligomeric aggregates. PLoS
Comput Biol, Vol. 4, No. 11, pp. e1000222.
Auger, M. (2010). Structural and Dynamics Studies of Lipids by Solid-State NMR. Solid State
NMR Studies of Biopolymers. A. McDermott and T. Polenova, Wiley.
Baglioni, S., Casamenti, F., Bucciantini, M., Luheshi, L. M., Taddei, N., Chiti, F., Dobson, C. M.
and Stefani, M. (2006). Prefibrillar amyloid aggregates could be generic toxins in higher
organisms. Journal of Neuroscience, Vol. 26, No. 31, pp. 8160-7.
Balbach, J. J., Ishii, Y., Antzutkin, O. N., Leapman, R. D., Rizzo, N. W., Dyda, F., Reed, J. and
Tycko, R. (2000). Amyloid fibril formation by A beta 16-22, a seven-residue fragment of
the Alzheimer's beta-amyloid peptide, and structural characterization by solid state NMR.
Biochemistry, Vol. 39, No. 45, pp. 13748-59.
Balbach, J. J., Petkova, A. T., Oyler, N. A., Antzutkin, O. N., Gordon, D. J., Meredith, S. C. and
Tycko, R. (2002). Supramolecular structure in full-length Alzheimer's beta-amyloid
fibrils: evidence for a parallel beta-sheet organization from solid-state nuclear magnetic
resonance. Biophysical Journal, Vol. 83, No. 2, pp. 1205-16.
Baxa, U., Wickner, R. B., Steven, A. C., Anderson, D. E., Marekov, L. N., Yau, W. M. and
Tycko, R. (2007). Characterization of beta-sheet structure in Ure2p1-89 yeast prion
fibrils by solid-state nuclear magnetic resonance. Biochemistry, Vol. 46, No. 45, pp.
13149-13162.
Beck, J. A., Poulter, M., Campbell, T. A., Adamson, G., Uphill, J. B., Guerreiro, R., Jackson, G.
S., Stevens, J. C., Manji, H., Collinge, J. and Mead, S. (2010). PRNP allelic series from
19 years of prion protein gene sequencing at the MRC Prion Unit. Hum Mutat, Vol. 31,
No. 7, pp. E1551-63.
128
Bennett, A. E., Rienstra, C. M., Auger, M., Lakshmi, K. V. and Griffin, R. G. (1995).
Heteronuclear decoupling in rotating solids. Journal of Chemical Physics, Vol. 103, No.
16, pp. 6951-6958.
Bergstrom, A. L., Chabry, J., Bastholm, L. and Heegaard, P. M. (2007). Oxidation reduces the
fibrillation but not the neurotoxicity of the prion peptide PrP106-126. Biochim Biophys
Acta, Vol. 1774, No. 9, pp. 1118-27.
Berthelot, K., Cullin, C. and Lecomte, S. (2013). What does make an amyloid toxic:
Morphology, structure or interaction with membrane? Biochimie, Vol. 95, No. 1, pp. 12-
9.
Bhatia, R., Lin, H. and Lal, R. (2000). Fresh and globular amyloid beta protein (1-42) induces
rapid cellular degeneration: evidence for AbetaP channel-mediated cellular toxicity.
FASEB Journal, Vol. 14, No. 9, pp. 1233-43.
Bodenhausen, G. and Ruben, D. J. (1980). Natural abundance N-15 NMR by enhanced
heteronuclear spectroscopy. Chemical Physics Letters, Vol. 69, pp. 185-189.
Brender, J. R., Hartman, K., Gottler, L. M., Cavitt, M. E., Youngstrom, D. W. and
Ramamoorthy, A. (2009). Helical conformation of the SEVI precursor peptide PAP248-
286, a dramatic enhancer of HIV infectivity, promotes lipid aggregation and fusion.
Biophys J, Vol. 97, No. 9, pp. 2474-83.
Brender, J. R., Lee, E. L., Cavitt, M. A., Gafni, A., Steel, D. G. and Ramamoorthy, A. (2008).
Amyloid fiber formation and membrane disruption are separate processes localized in
two distinct regions of IAPP, the type-2-diabetes-related peptide. J Am Chem Soc, Vol.
130, No. 20, pp. 6424-9.
Brender, J. R., Salamekh, S. and Ramamoorthy, A. (2012). Membrane disruption and early
events in the aggregation of the diabetes related peptide IAPP from a molecular
perspective. Acc Chem Res, Vol. 45, No. 3, pp. 454-62.
Brown, D. R. (2000). Prion protein peptides: optimal toxicity and peptide blockade of toxicity.
Mol Cell Neurosci, Vol. 15, No. 1, pp. 66-78.
Bucciantini, M., Giannoni, E., Chiti, F., Baroni, F., Formigli, L., Zurdo, J., Taddei, N., Ramponi,
G., Dobson, C. M. and Stefani, M. (2002). Inherent toxicity of aggregates implies a
common mechanism for protein misfolding diseases. Nature, Vol. 416, No. 6880, pp.
507-11.
Bucciantini, M., Nosi, D., Forzan, M., Russo, E., Calamai, M., Pieri, L., Formigli, L., Quercioli,
F., Soria, S., Pavone, F., Savistchenko, J., Melki, R. and Stefani, M. (2005). Toxic effects
of amyloid fibrils on cell membranes: the importance of ganglioside GM1. Faseb J, Vol.
26, No. 2, pp. 818-31.
Burdick, D., Soreghan, B., Kwon, M., Kosmoski, J., Knauer, M., Henschen, A., Yates, J.,
Cotman, C. and Glabe, C. (1992). Assembly and aggregation properties of synthetic
129
Alzheimer's A4/beta amyloid peptide analogs. Journal of Biological Chemistry, Vol. 267,
No. 1, pp. 546-54.
Buttke, T. M., McCubrey, J. A. and Owen, T. C. (1993). Use of an aqueous soluble
tetrazolium/formazan assay to measure viability and proliferation of lymphokine-
dependent cell lines. J Immunol Methods, Vol. 157, No. 1-2, pp. 233-40.
Cahuzac, N., Baum, W., Kirkin, V., Conchonaud, F., Wawrezinieck, L., Marguet, D., Janssen,
O., Zornig, M. and Hueber, A. O. (2006). Fas ligand is localized to membrane rafts,
where it displays increased cell death-inducing activity. Blood, Vol. 107, No. 6, pp. 2384-
91.
Carimalo, J., Cronier, S., Petit, G., Peyrin, J. M., Boukhtouche, F., Arbez, N., Lemaigre-
Dubreuil, Y., Brugg, B. and Miquel, M. C. (2005). Activation of the JNK-c-Jun pathway
during the early phase of neuronal apoptosis induced by PrP106-126 and prion infection.
European Journal of Neuroscience, Vol. 21, No. 9, pp. 2311-9.
Carr, H. Y. and Purcell, E. M. (1954). Effects of Diffusion on Free Precession in Nuclear
Magnetic Resonance Experiments. Physical Review, Vol. 94, No. 3, pp. 630-638.
Castro, B. M., de Almeida, R. F., Goormaghtigh, E., Fedorov, A. and Prieto, M. (2011).
Organization and dynamics of Fas transmembrane domain in raft membranes and
modulation by ceramide. Biophys J, Vol. 101, No. 7, pp. 1632-41.
Caughey, B. and Lansbury, P. T. (2003). Protofibrils, pores, fibrils, and neurodegeneration:
separating the responsible protein aggregates from the innocent bystanders. Annual
Review of Neuroscience, Vol. 26, pp. 267-98.
Cecchi, C., Nichino, D., Zampagni, M., Bernacchioni, C., Evangelisti, E., Pensalfini, A., Liguri,
G., Gliozzi, A., Stefani, M. and Relini, A. (2009). A protective role for lipid raft
cholesterol against amyloid-induced membrane damage in human neuroblastoma cells.
Biochim Biophys Acta, Vol. 1788, No. 10, pp. 2204-16.
Chan, J. C. (2011). Solid-State NMR Techniques for the Structural Determination of Amyloid
Fibrils. Topics in Current Chemistry, Vol. 2011, pp. 1.
Cheng, Y., Wolf, E., Larvie, M., Zak, O., Aisen, P., Grigorieff, N., Harrison, S. C. and Walz, T.
(2006). Single Particle Reconstructions of the Transferrin-Transferrin Receptor Complex
Obtained with Different Specimen Preparation Techniques. Journal of Molecular
Biology, Vol. 355, No. 5, pp. 1048-1065.
Cheon, M., Chang, I., Mohanty, S., Luheshi, L. M., Dobson, C. M., Vendruscolo, M. and Favrin,
G. (2007). Structural reorganisation and potential toxicity of oligomeric species formed
during the assembly of amyloid fibrils. PLoS Computational Biology, Vol. 3, No. 9, pp.
1727-38.
Chimon, S. and Ishii, Y. (2005). Capturing intermediate structures of Alzheimer's beta-amyloid,
Abeta(1-40), by solid-state NMR spectroscopy. Journal of the American Chemical
Society, Vol. 127, No. 39, pp. 13472-3.
130
Chimon, S., Shaibat, M. A., Jones, C. R., Calero, D. C., Aizezi, B. and Ishii, Y. (2007). Evidence
of fibril-like beta-sheet structures in a neurotoxic amyloid intermediate of Alzheimer's
beta-amyloid. Nature Structural and Molecular Biology, Vol. 14, No. 12, pp. 1157-1164.
Cho, C. H., Urquidi, J., Singh, S. and Robinson, G. W. (1999). Thermal Offset Viscosities of
Liquid H2O, D2O, and T2O. The Journal of Physical Chemistry B, Vol. 103, No. 11, pp.
1991-1994.
Chong, Y. H., Shin, Y. J., Lee, E. O., Kayed, R., Glabe, C. G. and Tenner, A. J. (2006). ERK1/2
activation mediates Abeta oligomer-induced neurotoxicity via caspase-3 activation and
tau cleavage in rat organotypic hippocampal slice cultures. J Biol Chem, Vol. 281, No.
29, pp. 20315-25.
Choo-Smith, L. P. and Surewicz, W. K. (1997). The interaction between Alzheimer amyloid
beta(1-40) peptide and ganglioside GM1-containing membranes. FEBS Lett, Vol. 402,
No. 2-3, pp. 95-8.
Choy, W. Y., Mulder, F. A., Crowhurst, K. A., Muhandiram, D. R., Millett, I. S., Doniach, S.,
Forman-Kay, J. D. and Kay, L. E. (2002). Distribution of molecular size within an
unfolded state ensemble using small-angle X-ray scattering and pulse field gradient NMR
techniques. Journal of Molecular Biology, Vol. 316, No. 1, pp. 101-12.
Cobb, N. J., Sonnichsen, F. D., McHaourab, H. and Surewicz, W. K. (2007). Molecular
architecture of human prion protein amyloid: a parallel, in-register beta-structure.
Proceedings of the National Academy of Sciences of the United States of America, Vol.
104, No. 48, pp. 18946-51.
Colby, D. W. and Prusiner, S. B. (2011). Prions. Cold Spring Harb Perspect Biol, Vol. 3, No. 1,
pp. a006833.
Comellas, G., Lemkau, L. R., Zhou, D. H., George, J. M. and Rienstra, C. M. (1021). Structural
intermediates during alpha-synuclein fibrillogenesis on phospholipid vesicles. J Am
Chem Soc, Vol. 134, No. 11, pp. 5090-9.
Conway, K. A., Harper, J. D. and Lansbury, P. T., Jr. (2000). Fibrils formed in vitro from alpha-
synuclein and two mutant forms linked to Parkinson's disease are typical amyloid.
Biochemistry, Vol. 39, No. 10, pp. 2552-63.
Conway, K. A., Lee, S. J., Rochet, J. C., Ding, T. T., Williamson, R. E. and Lansbury, P. T., Jr.
(2000). Acceleration of oligomerization, not fibrillization, is a shared property of both
alpha-synuclein mutations linked to early-onset Parkinson's disease: implications for
pathogenesis and therapy. Proceedings of the National Academy of Sciences of the United
States of America, Vol. 97, No. 2, pp. 571-6.
Cooper, G. J., Willis, A. C., Clark, A., Turner, R. C., Sim, R. B. and Reid, K. B. (1987).
Purification and characterization of a peptide from amyloid-rich pancreases of type 2
diabetic patients. Proceedings of the National Academy of Sciences of the United States
of America, Vol. 84, No. 23, pp. 8628-32.
131
Cornilescu, G., Delaglio, F. and Bax, A. (1999). Protein backbone angle restraints from
searching a database for chemical shift and sequence homology. Journal of Biomolecular
NMR, Vol. 13, No. 3, pp. 289-302.
Coustou, V., Deleu, C., Saupe, S. and Begueret, J. (1997). The protein product of the het-s
heterokaryon incompatibility gene of the fungus Podospora anserina behaves as a prion
analog. Proc Natl Acad Sci U S A, Vol. 94, No. 18, pp. 9773-8.
Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J. and Bax, A. (1995). NMRPipe: a
multidimensional spectral processing system based on UNIX pipes. Journal of
Biomolecular NMR, Vol. 6, No. 3, pp. 277-93.
Deli, M. A., Sakaguchi, S., Nakaoke, R., Abraham, C. S., Takahata, H., Kopacek, J., Shigematsu,
K., Katamine, S. and Niwa, M. (2000). PrP fragment 106-126 is toxic to cerebral
endothelial cells expressing PrP(C). Neuroreport, Vol. 11, No. 17, pp. 3931-6.
Demuro, A., Mina, E., Kayed, R., Milton, S. C., Parker, I. and Glabe, C. G. (2005). Calcium
dysregulation and membrane disruption as a ubiquitous neurotoxic mechanism of soluble
amyloid oligomers. Journal of Biological Chemistry, Vol. 280, No. 17, pp. 17294-300.
Dobson, C. M. (2003). Protein folding and misfolding. Nature, Vol. 426, No. 6968, pp. 884-90.
Duer, M. J. (2007). Dipolar Coupling: Its Measurement and Uses. Solid-State NMR
Spectroscopy Principles and Applications, Blackwell Science Ltd: 111-178.
Duer, M. J. (2007). Shielding and Chemical shift. Solid-State NMR Spectroscopy Principles and
Applications, Blackwell Science Ltd: 216-236.
Dufourc, E. J., Smith, I. C. P. and Dufourcq, J. (1986). Molecular details of melittin-induced
lysis of phospholipid membranes as revealed by deuterium and phosphorus NMR.
Biochemistry, Vol. 25, No. 21, pp. 6448-6455.
Dupiereux, I., Zorzi, W., Lins, L., Brasseur, R., Colson, P., Heinen, E. and Elmoualij, B. (2005).
Interaction of the 106-126 prion peptide with lipid membranes and potential implication
for neurotoxicity. Biochem. Biophys. Res. Commun., Vol. 331, No. 3, pp. 894-901.
Eanes, E. D. and Glenner, G. G. (1968). X-ray diffraction studies on amyloid filaments. Journal
of Histochemistry & Cytochemistry, Vol. 16, No. 11, pp. 673-7.
Ebenezer, P. J., Weidner, A. M., LeVine, H., 3rd, Markesbery, W. R., Murphy, M. P., Zhang, L.,
Dasuri, K., Fernandez-Kim, S. O., Bruce-Keller, A. J., Gavilan, E. and Keller, J. N.
Neuron specific toxicity of oligomeric amyloid-beta: role for JUN-kinase and oxidative
stress. J Alzheimers Dis, Vol. 22, No. 3, pp. 839-48.
Eliezer, D. (2006). Amyloid ion channels: a porous argument or a thin excuse? The Journal of
General Physiology, Vol. 128, No. 6, pp. 631-3.
132
Ettaiche, M., Pichot, R., Vincent, J. P. and Chabry, J. (2000). In vivo cytotoxicity of the prion
protein fragment 106-126. Journal of Biological Chemistry, Vol. 275, No. 47, pp. 36487-
90.
Falsig, J. and Aguzzi, A. (2008). The prion organotypic slice culture assay--POSCA. Nat Protoc,
Vol. 3, No. 4, pp. 555-62.
Fandrich, M., Meinhardt, J. and Grigorieff, N. (2009). Structural polymorphism of Alzheimer
Abeta and other amyloid fibrils. Prion, Vol. 3, No. 2, pp. 89-93.
Ferreiro, E., Oliveira, C. R. and Pereira, C. M. (2008). The release of calcium from the
endoplasmic reticulum induced by amyloid-beta and prion peptides activates the
mitochondrial apoptotic pathway. Neurobiology of Disease, Vol. 30, No. 3, pp. 331-42.
Florio, T., Paludi, D., Villa, V., Principe, D. R., Corsaro, A., Millo, E., Damonte, G., D'Arrigo,
C., Russo, C., Schettini, G. and Aceto, A. (2003). Contribution of two conserved glycine
residues to fibrillogenesis of the 106-126 prion protein fragment. Evidence that a souble
variant of the 106-126 peptide is neurotoxic. Journal of Neurochemistry, Vol. 85, No. 1,
pp. 62-72.
Florio, T., Thellung, S., Amico, C., Robello, M., Salmona, M., Bugiani, O., Tagliavini, F.,
Forloni, G. and Schettini, G. (1998). Prion protein fragment 106-126 induces apoptotic
cell death and impairment of L-type voltage-sensitive calcium channel activity in the
GH3 cell line. Journal of Neuroscience Research, Vol. 54, No. 3, pp. 341-52.
Forloni, G., Angeretti, N., Chiesa, R., Monzani, E., Salmona, M., Bugiani, O. and Tagliavini, F.
(1993). Neurotoxicity of a prion protein fragment. Nature, Vol. 362, No. 6420, pp. 543-6.
Forloni, G., Salmona, M., Bugiani, O. and Tagliavini, F. (2000). Comment on: Neurotoxicity of
prion peptide 106-126 not confirmed, by Beat Kunz, Erika Sandmeier, Philipp Christen.
FEBS Letters 485 (1999) 65-68. FEBS Letters, Vol. 466, No. 1, pp. 205-6; discussion
207.
Frare, E., Mossuto, M. F., de Laureto, P. P., Tolin, S., Menzer, L., Dumoulin, M., Dobson, C. M.
and Fontana, A. (2009). Characterization of oligomeric species on the aggregation
pathway of human lysozyme. Journal of Molecular Biology, Vol. 387, No. 1, pp. 17-27.
Gajate, C., Gonzalez-Camacho, F. and Mollinedo, F. (2009). Lipid raft connection between
extrinsic and intrinsic apoptotic pathways. Biochem Biophys Res Commun, Vol. 380, No.
4, pp. 780-4.
Geddes, A. J., Parker, K. D., Atkins, E. D. and Beighton, E. (1968). "Cross-beta" conformation
in proteins. Journal of Molecular Biology, Vol. 32, No. 2, pp. 343-58.
Glabe, C. (2008). Structural classification of toxic amyloid oligomers. Journal of Biological
Chemistry, Vol. 283, No. 44, pp. 29639-29643.
133
Glenner, G. G. and Wong, C. W. (1984). Alzheimer's disease: initial report of the purification
and characterization of a novel cerebrovascular amyloid protein. Biochemical and
Biophysical Research Communications, Vol. 120, No. 3, pp. 885-90.
Govaerts, C., Wille, H., Prusiner, S. B. and Cohen, F. E. (2004). Evidence for assembly of prions
with left-handed beta-helices into trimers. Proceedings of the National Academy of
Sciences of the United States of America, Vol. 101, No. 22, pp. 8342-7.
Gu, Y., Fujioka, H., Mishra, R. S., Li, R. and Singh, N. (2002). Prion peptide 106-126 modulates
the aggregation of cellular prion protein and induces the synthesis of potentially
neurotoxic transmembrane PrP. Journal of Biological Chemistry, Vol. 277, No. 3, pp.
2275-86.
Gullion, T. and Schaefer, J. (1989). Rotational-echo double-resonance NMR. Journal of
Magnetic Resonance (1969), Vol. 81, No. 1, pp. 196-200.
Haass, C. and Selkoe, D. J. (2007). Soluble protein oligomers in neurodegeneration: lessons from
the Alzheimer's amyloid beta-peptide. Nature Reviews Molecular Cell Biology, Vol. 8,
No. 2, pp. 101-12.
Heise, H., Hoyer, W., Becker, S., Andronesi, O. C., Riedel, D. and Baldus, M. (2005).
Molecular-level secondary structure, polymorphism, and dynamics of full-length alpha-
synuclein fibrils studied by solid-state NMR. Proceedings of the National Academy of
Sciences of the United States of America, Vol. 102, No. 44, pp. 15871-6.
Hing, A. W., Vega, S. and Schaefer, J. (1992). Transferred-echo double-resonance NMR.
Journal of Magnetic Resonance (1969), Vol. 96, No. 1, pp. 205-209.
Holscher, C., Delius, H. and Burkle, A. (1998). Overexpression of nonconvertible PrPc delta114-
121 in scrapie-infected mouse neuroblastoma cells leads to trans-dominant inhibition of
wild-type PrP(Sc) accumulation. J Virol, Vol. 72, No. 2, pp. 1153-9.
Hong, M. (2006). Oligomeric structure, dynamics, and orientation of membrane proteins from
solid-state NMR. Structure, Vol. 14, No. 12, pp. 1731-40.
Huang, H., Milojevic, J. and Melacini, G. (2008). Analysis and optimization of saturation
transfer difference NMR experiments designed to map early self-association events in
amyloidogenic peptides. J Phys Chem B, Vol. 112, No. 18, pp. 5795-802.
Huang, T. H., Yang, D. S., Plaskos, N. P., Go, S., Yip, C. M., Fraser, P. E. and Chakrabartty, A.
(2000). Structural studies of soluble oligomers of the Alzheimer beta-amyloid peptide.
Journal of Molecular Biology, Vol. 297, No. 1, pp. 73-87.
Ikeda, K., Yamaguchi, T., Fukunaga, S., Hoshino, M. and Matsuzaki, K. (2011). Mechanism of
amyloid beta-protein aggregation mediated by GM1 ganglioside clusters. Biochemistry,
Vol. 50, No. 29, pp. 6433-40.
Iwahara, J. and Clore, G. M. (2006). Detecting transient intermediates in macromolecular
binding by paramagnetic NMR. Nature, Vol. 440, No. 7088, pp. 1227-30.
134
Jahn, T. R., Makin, O. S., Morris, K. L., Marshall, K. E., Tian, P., Sikorski, P. and Serpell, L. C.
(2009). The common architecture of cross-beta amyloid. Journal of Molecular Biology,
Vol. 395, No. 4, pp. 717-27.
Jang, H., Zheng, J., Lal, R. and Nussinov, R. (2008). New structures help the modeling of toxic
amyloidbeta ion channels. Trends in Biochemical Sciences, Vol. 33, No. 2, pp. 91-100.
Jang, H., Zheng, J. and Nussinov, R. (2007). Models of beta-amyloid ion channels in the
membrane suggest that channel formation in the bilayer is a dynamic process.
Biophysical Journal, Vol. 93, No. 6, pp. 1938-49.
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, Vol. 48, No. 3, pp. 491-8.
Jaroniec, C. P., MacPhee, C. E., Astrof, N. S., Dobson, C. M. and Griffin, R. G. (2002).
Molecular conformation of a peptide fragment of transthyretin in an amyloid fibril.
Proceedings of the National Academy of Sciences of the United States of America, Vol.
99, No. 26, pp. 16748-53.
Jaroniec, C. P., MacPhee, C. E., Bajaj, V. S., McMahon, M. T., Dobson, C. M. and Griffin, R. G.
(2004). High-resolution molecular structure of a peptide in an amyloid fibril determined
by magic angle spinning NMR spectroscopy. Proceedings of the National Academy of
Sciences of the United States of America, Vol. 101, No. 3, pp. 711-6.
Jarrett, J. T., Berger, E. P. and Lansbury, P. T., Jr. (1993). The carboxy terminus of the beta
amyloid protein is critical for the seeding of amyloid formation: implications for the
pathogenesis of Alzheimer's disease. Biochemistry, Vol. 32, No. 18, pp. 4693-7.
Jeener, J., Meier, B. H., Bachmann, P. and Ernst, R. R. (1979). Investigation of exchange
processes by two-dimensional NMR spectroscopy. Journal of Chemical Physics, Vol. 71,
pp. 4546-4553.
Jobling, M. F., Stewart, L. R., White, A. R., McLean, C., Friedhuber, A., Maher, F., Beyreuther,
K., Masters, C. L., Barrow, C. J., Collins, S. J. and Cappai, R. (1999). The hydrophobic
core sequence modulates the neurotoxic and secondary structure properties of the prion
peptide 106-126. Journal of Neurochemistry, Vol. 73, No. 4, pp. 1557-1565.
Johnson, B. A. and Blevins, R. A. (1994). NMRView: a computer program for the visualization
and analysis of NMR data. Journal of Biomolecular NMR, Vol. 4, pp. 603-614.
Jones, M., Wight, D., McLoughlin, V., Norrby, K., Ironside, J. W., Connolly, J. G., Farquhar, C.
F., Macgregor, I. R. and Head, M. W. (2008). An Antibody to the Aggregated Synthetic
Prion Protein Peptide (PrP106-126) Selectively Recognizes Disease-Associated Prion
Protein (PrP(Sc)) from Human Brain Specimens. Brain Pathol.
Kaneko, K., Peretz, D., Pan, K. M., Blochberger, T. C., Wille, H., Gabizon, R., Griffith, O. H.,
Cohen, F. E., Baldwin, M. A. and Prusiner, S. B. (1995). Prion protein (PrP) synthetic
135
peptides induce cellular PrP to acquire properties of the scrapie isoform. Proc. Natl.
Acad. Sci. USA, Vol. 92, No. 24, pp. 11160-4.
Kay, L. E., Keifer, P. and Saarinen, T. (1992). Pure Absorption Gradient Enhanced
Heteronuclear Single Quantum Correlation Spectroscopy with Improved Sensitivity.
Journal of the American Chemical Society, Vol. 114, pp. 10663-10665.
Kayalar, C. and Duzgunes, N. (1986). Membrane action of colicin E1: detection by the release of
carboxyfluorescein and calcein from liposomes. Biochim Biophys Acta, Vol. 860, No. 1,
pp. 51-6.
Kayed, R., Bernhagen, J., Greenfield, N., Sweimeh, K., Brunner, H., Voelter, W. and
Kapurniotu, A. (1999). Conformational transitions of islet amyloid polypeptide (IAPP) in
amyloid formation in vitro. Journal of Molecular Biology, Vol. 287, No. 4, pp. 781-96.
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, Vol. 300, No. 5618, pp. 486-9.
Kayed, R., Pensalfini, A., Margol, L., Sokolov, Y., Sarsoza, F., Head, E., Hall, J. and Glabe, C.
(2009). Annular protofibrils are a structurally and functionally distinct type of amyloid
oligomer. Journal of Biological Chemistry, Vol. 284, No. 7, pp. 4230-7.
Kayed, R., Sokolov, Y., Edmonds, B., McIntire, T. M., Milton, S. C., Hall, J. E. and Glabe, C. G.
(2004). Permeabilization of lipid bilayers is a common conformation-dependent activity
of soluble amyloid oligomers in protein misfolding diseases. Journal of Biological
Chemistry, Vol. 279, No. 45, pp. 46363-6.
Kim, H.-Y., Cho, M.-K., Kumar, A., Maier, E., Siebenhaar, C., Becker, S., Fernandez, C. O.,
Lashuel, H. A., Benz, R., Lange, A. and Zweckstetter, M. (2009). Structural Properties of
Pore-Forming Oligomers of α-Synuclein. Journal of the American Chemical Society, Vol.
131, No. 47, pp. 17482-17489.
Koradi, R., Billeter, M. and Wuthrich, K. (1996). MOLMOL: a program for display and analysis
of macromolecular structures. J Mol Graph, Vol. 14, No. 1, pp. 51-5, 29-32.
Korlach, J., Schwille, P., Webb, W. W. and Feigenson, G. W. (1999). Characterization of lipid
bilayer phases by confocal microscopy and fluorescence correlation spectroscopy. Proc
Natl Acad Sci U S A, Vol. 96, No. 15, pp. 8461-6.
Kostka, M., Hogen, T., Danzer, K. M., Levin, J., Habeck, M., Wirth, A., Wagner, R., Glabe, C.
G., Finger, S., Heinzelmann, U., Garidel, P., Duan, W., Ross, C. A., Kretzschmar, H. and
Giese, A. (2008). Single particle characterization of iron-induced pore-forming alpha-
synuclein oligomers. J Biol Chem, Vol. 283, No. 16, pp. 10992-1003.
Kourie, J. I. and Culverson, A. L. (2000). Prion peptide fragment PrP[106-126] forms distinct
cation channel types. Journal of Neuroscience Research, Vol. 62, pp. 120-133.
136
Kourie, J. I., Farrelly, P. V. and Henry, C. L. (2001). Channel activity of deamidated isoforms of
prion protein fragment 106-126 in planar lipid bilayers. Journal of Neuroscience
Research, Vol. 66, No. 2, pp. 214-20.
Kumar-Singh, S., Theuns, J., Van Broeck, B., Pirici, D., Vennekens, K., Corsmit, E., Cruts, M.,
Dermaut, B., Wang, R. and Van Broeckhoven, C. (2006). Mean age-of-onset of familial
alzheimer disease caused by presenilin mutations correlates with both increased Abeta42
and decreased Abeta40. Human Mutation, Vol. 27, No. 7, pp. 686-95.
Kurganov, B., Doh, M. and Arispe, N. (2004). Aggregation of liposomes induced by the toxic
peptides Alzheimer's As, human amylin and prion(106-126): facilitation by membrane-
bound GM1 ganglioside. Peptides, Vol. 25, No. 2, pp. 217-232.
Kuwata, K., Matumoto, T., Cheng, H., Nagayama, K., James, T. L. and Roder, H. (2003). NMR-
detected hydrogen exchange and molecular dynamics simulations provide structural
insight into fibril formation of prion protein fragment 106-126. Proceedings of the
National Academy of Sciences of the United States of America, Vol. 100, No. 25, pp.
14790-14795.
Laganowsky, A., Liu, C., Sawaya, M. R., Whitelegge, J. P., Park, J., Zhao, M., Pensalfini, A.,
Soriaga, A. B., Landau, M., Teng, P. K., Cascio, D., Glabe, C. and Eisenberg, D. (2012).
Atomic view of a toxic amyloid small oligomer. Science, Vol. 335, No. 6073, pp. 1228-
31.
Lambert, M. P., Barlow, A. K., Chromy, B. A., Edwards, C., Freed, R., Liosatos, M., Morgan, T.
E., Rozovsky, I., Trommer, B., Viola, K. L., Wals, P., Zhang, C., Finch, C. E., Krafft, G.
A. and Klein, W. L. (1998). Diffusible, nonfibrillar ligands derived from Abeta1-42 are
potent central nervous system neurotoxins. Proceedings of the National Academy of
Sciences of the United States of America, Vol. 95, No. 11, pp. 6448-53.
Lashuel, H. A., Hartley, D., Petre, B. M., Walz, T. and Lansbury, P. T., Jr. (2002).
Neurodegenerative disease: amyloid pores from pathogenic mutations. Nature, Vol. 418,
No. 6895, pp. 291.
Laurents, D. V., Gorman, P. M., Guo, M., Rico, M., Chakrabartty, A. and Bruix, M. (2005).
Alzheimer's Abeta40 studied by NMR at low pH reveals that sodium 4,4-dimethyl-4-
silapentane-1-sulfonate (DSS) binds and promotes beta-ball oligomerization. Journal of
Biological Chemistry, Vol. 280, No. 5, pp. 3675-85.
Lee, S., Mou, Y., Lin, S.-Y., Chou, F.-C., Tseng, W.-H., Chen, C.-h., Lu, C.-Y. D., Yu, S. S. F.
and Chan, J. C. C. (2008). Steric Zipper of the Amyloid Fibrils Formed by Residues 109-
122 of the Syrian Hamster Prion Protein. Journal of Molecular Biology, Vol. 378, No. 5,
pp. 1142-1154.
Lin, H., Bhatia, R. and Lal, R. (2001). Amyloid beta protein forms ion channels: implications for
Alzheimer's disease pathophysiology. FASEB Journal, Vol. 15, No. 13, pp. 2433-44.
Lin, M. C., Mirzabekov, T. and Kagan, B. L. (1997). Channel formation by a neurotoxic prion
protein fragment. Journal of Biological Chemistry, Vol. 272, No. 1, pp. 44-7.
137
Losic, D., Martin, L. L., Mechler, A., Aguilar, M. I. and Small, D. H. (2006). High resolution
scanning tunnelling microscopy of the beta-amyloid protein (Abeta1-40) of Alzheimer's
disease suggests a novel mechanism of oligomer assembly. Journal of Structural Biology,
Vol. 155, No. 1, pp. 104-10.
Lu, J. X., Yau, W. M. and Tycko, R. (2011). Evidence from solid-state NMR for nonhelical
conformations in the transmembrane domain of the amyloid precursor protein. Biophys J,
Vol. 100, No. 3, pp. 711-9.
Luca, S., Yau, W.-M., Leapman, R. and Tycko, R. (2007). Peptide conformation and
supramolecular organization in amylin fibrils: Constraints from solid-state NMR.
Biochemistry, Vol. 46, pp. 13505-13522.
Luhrs, T., Ritter, C., Adrian, M., Riek-Loher, D., Bohrmann, B., Dobeli, H., Schubert, D. and
Riek, R. (2005). 3D structure of Alzheimer's amyloid-beta(1-42) fibrils. Proceedings of
the National Academy of Sciences of the United States of America, Vol. 102, No. 48, pp.
17342-7.
Madine, J., Jack, E., Stockley, P. G., Radford, S. E., Serpell, L. C. and Middleton, D. A. (2008).
Structural insights into the polymorphism of amyloid-like fibrils formed by region 20-29
of amylin revealed by solid-state NMR and X-ray fiber diffraction. Journal of the
American Chemical Society, Vol. 130, No. 45, pp. 14990-5001.
Marion, D., Ikura, M., Tschundin, R. and Bax, A. (1989). Rapid Recording of 2D NMR Spectra
without Phase Cycling. Application to the Study of Hydrogen Exchange in Proteins.
Journal of Magnetic Resonance, Vol. 85, pp. 393-399.
Markley, J. L., Bax, A., Arata, Y., Hilbers, C. W., Kaptein, R., Sykes, B. D., Wright, P. E. and
Wuthrich, K. (1998). Recommendations for the Presentation of NMR Structures of
Proteins and Nucleic Acids. Pure and Applied Chemistry, Vol. 70, pp. 117-142.
McLaurin, J. and Chakrabartty, A. (1996). Membrane disruption by Alzheimer beta-amyloid
peptides mediated through specific binding to either phospholipids or gangliosides.
Implications for neurotoxicity. Journal of Biological Chemistry, Vol. 271, No. 43, pp.
26482-9.
McLaurin, J. and Chakrabartty, A. (1997). Characterization of the interactions of Alzheimer
beta-amyloid peptides with phospholipid membranes. European Journal of Biochemistry,
Vol. 245, No. 2, pp. 355-63.
Meiboom, S. and Gill, D. (1958). Modified spin-echo method for measuring nuclear relaxation
times. Review of Scientific Instruments, Vol. 29, No. 8, pp. 688-691.
Miller, Y., Ma, B., Tsai, C. J. and Nussinov, R. (2010). Hollow core of Alzheimer's Abeta42
amyloid observed by cryoEM is relevant at physiological pH. Proceedings of the
National Academy of Sciences of the United States of America, Vol. 107, No. 32, pp.
14128-33.
138
Miret, S., De Groene, E. M. and Klaffke, W. (2006). Comparison of in vitro assays of cellular
toxicity in the human hepatic cell line HepG2. J Biomol Screen, Vol. 11, No. 2, pp. 184-
93.
Mizuno, N., Baxa, U. and Steven, A. C. (2012). Structural dependence of HET-s amyloid fibril
infectivity assessed by cryoelectron microscopy. Proc Natl Acad Sci U S A, Vol. 108, No.
8, pp. 3252-7.
Morcombe, C. R., Gaponenko, V., Byrd, R. A. and Zilm, K. W. (2004). Diluting abundant spins
by isotope edited radio frequency field assisted diffusion. Journal of the American
Chemical Society, Vol. 126, No. 23, pp. 7196-7.
Morcombe, C. R. and Zilm, K. W. (2003). Chemical shift referencing in MAS solid state NMR.
Journal of Magnetic Resonance, Vol. 162, No. 2, pp. 479-86.
Naiki, H., Higuchi, K., Hosokawa, M. and Takeda, T. (1989). Fluorometric determination of
amyloid fibrils in vitro using the fluorescent dye, thioflavin T1. Anal Biochem, Vol. 177,
No. 2, pp. 244-9.
Naito, A., Kamihira, M., Inoue, R. and Saito, H. (2004). Structural diversity of amyloid fibril
formed in human calcitonin as revealed by site-directed 13C solid-state NMR
spectroscopy. Magnetic Resonance in Chemistry, Vol. 42, No. 2, pp. 247-57.
Nanga, R. P., Brender, J. R., Vivekanandan, S., Popovych, N. and Ramamoorthy, A. (2009).
NMR structure in a membrane environment reveals putative amyloidogenic regions of
the SEVI precursor peptide PAP(248-286). J Am Chem Soc, Vol. 131, No. 49, pp. 17972-
9.
Necula, M., Kayed, R., Milton, S. and Glabe, C. G. (2007). Small molecule inhibitors of
aggregation indicate that amyloid beta oligomerization and fibrillization pathways are
independent and distinct. Journal of Biological Chemistry, Vol. 282, No. 14, pp. 10311-
24.
Nielsen, J. T., Bjerring, M., Jeppesen, M. D., Pedersen, R. O., Pedersen, J. M., Hein, K. L.,
Vosegaard, T., Skrydstrup, T., Otzen, D. E. and Nielsen, N. C. (2009). Unique
identification of supramolecular structures in amyloid fibrils by solid-state NMR
spectroscopy. Angewandte Chemie International Edition, Vol. 48, No. 12, pp. 2118-21.
O'Donovan, C. N., Tobin, D. and Cotter, T. G. (2001). Prion protein fragment PrP-(106-126)
induces apoptosis via mitochondrial disruption in human neuronal SH-SY5Y cells.
Journal of Biological Chemistry, Vol. 276, No. 47, pp. 43516-23.
Olofsson, A., Lindhagen-Persson, M., Sauer-Eriksson, A. E. and Ohman, A. (2007). Amide
solvent protection analysis demonstrates that amyloid-beta(1-40) and amyloid-beta(1-42)
form different fibrillar structures under identical conditions. Biochemical Journal, Vol.
404, No. 1, pp. 63-70.
139
Ono, K., Condron, M. M. and Teplow, D. B. (2009). Structure-neurotoxicity relationships of
amyloid beta-protein oligomers. Proceedings of the National Academy of Sciences of the
United States of America, Vol. 106, No. 35, pp. 14745-50.
Pan, J., Han, J., Borchers, C. H. and Konermann, L. (2011). Conformer-specific hydrogen
exchange analysis of abeta(1-42) oligomers by top-down electron capture dissociation
mass spectrometry. Analytical Chemistry, Vol. 83, No. 13, pp. 5386-93.
Paravastu, A. K., Petkova, A. T. and Tycko, R. (2006). Polymorphic fibril formation by residues
10-40 of the Alzheimer's beta-amyloid peptide. Biophysical Journal, Vol. 90, No. 12, pp.
4618-29.
Paravastu, A. K., Qahwash, I., Leapman, R. D., Meredith, S. C. and Tycko, R. (2009). Seeded
growth of beta-amyloid fibrils from Alzheimer's brain-derived fibrils produces a distinct
fibril structure. Proc Natl Acad Sci U S A, Vol. 106, No. 18, pp. 7443-8.
Pastor, M. T., Kummerer, N., Schubert, V., Esteras-Chopo, A., Dotti, C. G., Lopez de la Paz, M.
and Serrano, L. (2008). Amyloid toxicity is independent of polypeptide sequence, length
and chirality. Journal of Molecular Biology, Vol. 375, No. 3, pp. 695-707.
Petkova, A. T., Ishii, Y., Balbach, J. J., Antzutkin, O. N., Leapman, R. D., Delaglio, F. and
Tycko, R. (2002). A structural model for Alzheimer's beta -amyloid fibrils based on
experimental constraints from solid state NMR. Proceedings of the National Academy of
Sciences of the United States of America, Vol. 99, No. 26, pp. 16742-7.
Petkova, A. T., Leapman, R. D., Guo, Z., Yau, W. M., Mattson, M. P. and Tycko, R. (2005).
Self-propagating, molecular-level polymorphism in Alzheimer's beta-amyloid fibrils.
Science, Vol. 307, No. 5707, pp. 262-5.
Petkova, A. T., Yau, W. M. and Tycko, R. (2006). Experimental constraints on quaternary
structure in Alzheimer's beta-amyloid fibrils. Biochemistry, Vol. 45, No. 2, pp. 498-512.
Piotto, M., Saudek, V. and Sklenar, V. (1992). Gradient-tailored excitation for single-quantum
NMR spectroscopy of aqueous solutions. Journal of Biomolecular NMR, Vol. 2, pp. 661-
665.
Pollard, H. B., Arispe, N. and Rojas, E. (1995). Ion channel hypothesis for Alzheimer amyloid
peptide neurotoxicity. Cell. Mol. Neurobiol., Vol. 15, No. 5, pp. 513-26.
Prusiner, S. B. (1982). Novel proteinaceous infectious particles cause scrapie. Science, Vol. 216,
No. 4542, pp. 136-44.
Prusiner, S. B. (1996). Molecular biology and pathogenesis of prion diseases. Trends in
Biochemical Sciences, Vol. 21, No. 12, pp. 482-7.
Prusiner, S. B., Scott, M. R., DeArmond, S. J. and Cohen, F. E. (1998). Prion protein biology.
Cell, Vol. 93, No. 3, pp. 337-48.
140
Quist, A., Doudevski, I., Lin, H., Azimova, R., Ng, D., Frangione, B., Kagan, B., Ghiso, J. and
Lal, R. (2005). Amyloid ion channels: a common structural link for protein-misfolding
disease. Proceedings of the National Academy of Sciences of the United States of
America, Vol. 102, No. 30, pp. 10427-32.
Raleigh, D. P., Levitt, M. H. and Griffin, R. G. (1988). Rotational resonance in solid state NMR.
Chemical Physics Letters, Vol. 146, No. 1-2, pp. 71-76.
Reisner, D. (2003). Biochemistry and structure of PrPC and PrP
SC. British Medical Bulletin, Vol.
66, pp. 22-33.
Robustelli, P., Cavalli, A. and Vendruscolo, M. (2008). Determination of protein structures in the
solid state from NMR chemical shifts. Structure, Vol. 16, No. 12, pp. 1764-9.
Rucker, S. P. and Shaka, A. J. (1989). Broad-Band Homonuclear Cross Polarization in 2D NMR
Using DIPSI-2. Molecular Physics, Vol. 68, pp. 509-517.
Sachse, C., Fandrich, M. and Grigorieff, N. (2008). Paired beta-sheet structure of an Abeta(1-40)
amyloid fibril revealed by electron microscopy. Proceedings of the National Academy of
Sciences of the United States of America, Vol. 105, No. 21, pp. 7462-6.
Saito, H. (1986). Conformation-dependent 13
C chemical shifts: a new means of conformational
characterization as obtained by high-resolution solid-state 13
C NMR. Magnetic
Resonance in Chemistry, Vol. 24, pp. 835-852.
Saleem, Q., Lai, A., Morales, H. H. and Macdonald, P. M. (2012). Lateral diffusion of bilayer
lipids measured via (31)P CODEX NMR. Chem Phys Lipids, Vol. 165, No. 7, pp. 721-
730.
Salmona, M., Forloni, G., Diomede, L., Algeri, M., De Gioia, L., Angeretti, N., Giaccone, G.,
Tagliavini, F. and Bugiani, O. (1997). A neurotoxic and gliotrophic fragment of the prion
protein increases plasma membrane microviscosity. Neurobiology of Disease, Vol. 4, No.
1, pp. 47-57.
Salmona, M., Malesani, P., DeGioia, L., Gorla, S., Bruschi, M., Molinari, A., Della Vedova, F.,
Pedrotti, B., Marrari, M. A., Awan, T., Bugiani, O., Forloni, G. and Tagliavini, F. (1999).
Molecular determinants of the physicochemical properties of a critical prion protein
region comprising residues 106-126. Biochemical Journal, Vol. 342, pp. 207-214.
Sawaya, M. R., Sambashivan, S., Nelson, R., Ivanova, M. I., Sievers, S. A., Apostol, M. I.,
Thompson, M. J., Balbimie, M., Wiltzius, J. J., McFarlane, H. T., Madsen, A. O., Riekel,
C. and Eisenberg, D. (2007). Atomic structures of amyloid cross-beta spines reveal varied
steric zippers. Nature, Vol. 447, pp. 453-457.
Schäfer, H., Mädler, B. and Volke, F. (1995). De-Pake-ing of NMR Powder Spectra by
Nonnegative Least-Squares Analysis with Tikhonov Regularization. Journal of Magnetic
Resonance, Series A, Vol. 116, No. 2, pp. 145-149.
141
Schleucher, J., Sattler, M. and Griesinger, C. (1993). Coherence Selection by Gradients without
Signal Attenuation: Application to the Three-Dimensional HNCO Experiment.
Angewandte Chemie-Intl. Ed., Vol. 32, pp. 1489-1491.
Schmidt, M., Sachse, C., Richter, W., Xu, C., Fandrich, M. and Grigorieff, N. (2009).
Comparison of Alzheimer Abeta(1-40) and Abeta(1-42) amyloid fibrils reveals similar
protofilament structures. Proceedings of the National Academy of Sciences of the United
States of America, Vol. 106, No. 47, pp. 19813-8.
Seelig, J. (1978). 31P nuclear magnetic resonance and the head group structure of phospholipids
in membranes. Biochim Biophys Acta, Vol. 515, No. 2, pp. 105-40.
Selvaggini, C., De Gioia, L., Cantu, L., Ghibaudi, E., Diomede, L., Passerini, F., Forloni, G.,
Bugiani, O., Tagliavini, F. and Salmona, M. (1993). Molecular characteristics of a
protease-resistant, amyloidogenic and neurotoxic peptide homologous to residues 106-
126 of the prion protein. Biochemical and Biophysical Research Communications, Vol.
194, No. 3, pp. 1380-6.
Sharpe, S., Kessler, N., Anglister, J. A., Yau, W. M. and Tycko, R. (2004). Solid-state NMR
yields structural constraints on the V3 loop from HIV-1 Gp120 bound to the 447-52D
antibody Fv fragment. Journal of the American Chemical Society, Vol. 126, No. 15, pp.
4979-90.
Shen, Y., Vernon, R., Baker, D. and Bax, A. (2009). De novo protein structure generation from
incomplete chemical shift assignments. J Biomol NMR, Vol. 43, No. 2, pp. 63-78.
Shewmaker, F., McGlinchey, R. P., Thurber, K. R., McPhie, P., Dyda, F., Tycko, R. and
Wickner, R. B. (2009). The functional curli amyloid is not based on in-register parallel
beta-sheet structure. Journal of Biological Chemistry, Vol. 1, pp. 1.
Shewmaker, F., Wickner, R. B. and Tycko, R. (2006). Amyloid of the prion domain of Sup35p
has an in-register parallel beta-sheet structure. Proceedings of the National Academy of
Sciences of the United States of America, Vol. 103, No. 52, pp. 19754-9.
Shirahama, T. and Cohen, A. S. (1967). High-resolution electron microscopic analysis of the
amyloid fibril. J Cell Biol, Vol. 33, No. 3, pp. 679-708.
Silei, V., Fabrizi, C., Venturini, G., Salmona, M., Bugiani, O., Tagliavini, F. and Lauro, G. M.
(1999). Activation of microglial cells by PrP and beta-amyloid fragments raises
intracellular calcium through L-type voltage sensitive calcium channels. Brain Research,
Vol. 818, No. 1, pp. 168-170.
Silva, R. A. G. D., Barber-Armstrong, W. and Decatur, S. M. (2003). The Organization and
Assembly of a beta-Sheet Formed by a Prion Peptide in Solution: An Isotope-Edited
FTIR Study. Journal of the American Chemical Society, Vol. 125, No. 45, pp. 13674-
13675.
142
Silveira, J. R., Raymond, G. J., Hughson, A. G., Race, R. E., Sim, V. L., Hayes, S. F. and
Caughey, B. (2005). The most infectious prion protein particles. Nature, Vol. 437, No.
7056, pp. 257-61.
Simoneau, S., Rezaei, H., Sales, N., Kaiser-Schulz, G., Lefebvre-Roque, M., Vidal, C., Fournier,
J. G., Comte, J., Wopfner, F., Grosclaude, J., Schatzl, H. and Lasmezas, C. I. (2007). In
vitro and in vivo neurotoxicity of prion protein oligomers. PLoS Pathogens, Vol. 3, No.
8, pp. e125.
Singh, N., Gu, Y., Bose, S., Kalepu, S., Mishra, R. S. and Verghese, S. (2002). Prion peptide
106-126 as a model for prion replication and neurotoxicity. Frontiers of Bioscience, Vol.
7, pp. a60-a71.
Smith, P. E., Brender, J. R. and Ramamoorthy, A. (2009). Induction of negative curvature as a
mechanism of cell toxicity by amyloidogenic peptides: the case of islet amyloid
polypeptide. J Am Chem Soc, Vol. 131, No. 12, pp. 4470-8.
Sokolov, Y., Kozak, J. A., Kayed, R., Chanturiya, A., Glabe, C. and Hall, J. E. (2006). Soluble
amyloid oligomers increase bilayer conductance by altering dielectric structure. Journal
of General Physiology, Vol. 128, No. 6, pp. 637-47.
Spillantini, M. G., Schmidt, M. L., Lee, V. M., Trojanowski, J. Q., Jakes, R. and Goedert, M.
(1997). Alpha-synuclein in Lewy bodies. Nature, Vol. 388, No. 6645, pp. 839-40.
Spink, C. H., Yeager, M. D. and Feigenson, G. W. (1990). Partitioning behavior of
indocarbocyanine probes between coexisting gel and fluid phases in model membranes.
Biochim Biophys Acta, Vol. 1023, No. 1, pp. 25-33.
Stejskal, E. O. and Tanner, J. E. (1965). Spin Diffusion Measurments: Spin Echoes in the
Presence of a Time-Dependent Field Gradient. Journal of Chemical Physics, Vol. 42. pp.
288-292.
Stewart, J. C. (1980). Colorimetric determination of phospholipids with ammonium
ferrothiocyanate. Anal Biochem, Vol. 104, No. 1, pp. 10-4.
Su, Y., Mani, R. and Hong, M. (2008). Asymmetric insertion of membrane proteins in lipid
bilayers by solid-state NMR paramagnetic relaxation enhancement: a cell-penetrating
Peptide example. Journal of the American Chemical Society, Vol. 130, No. 27, pp. 8856-
64.
Tagliavini, F., Prelli, F., Verga, L., Giaccone, G., Sarma, R., Gorevic, P., Ghetti, B., Passerini,
F., Ghibaudi, E., Forloni, G. and et al. (1993). Synthetic peptides homologous to prion
protein residues 106-147 form amyloid-like fibrils in vitro. Proc Natl Acad Sci U S A,
Vol. 90, No. 20, pp. 9678-82.
Takegoshi, K., Nakamura, S. and Terao, T. (2001). 13C-1H dipolar-assisted rotational resonance
in magic-angle spinning NMR. Chemical Physics Letters, Vol. 344, No. 5-6, pp. 631-637.
143
Thellung, S., Florio, T., Corsaro, A., Arena, S., Merlino, M., Salmona, M., Tagliavini, F.,
Bugiani, O., Forloni, G. and Schettini, G. (2000). Intracellular mechanisms mediating the
neuronal death and astrogliosis induced by the prion protein fragment 106-126. Int J Dev
Neurosci, Vol. 18, No. 4-5, pp. 481-92.
Thellung, S., Florio, T., Villa, V., Corsaro, A., Arena, S., Amico, C., Robello, M., Salmona, M.,
Forloni, G., Bugiani, O., Tagliavini, F. and Schettini, G. (2000). Apoptotic cell death and
impairment of L-type voltage-sensitive calcium channel activity in rat cerebellar granule
cells treated with the prion protein fragment 106-126. Neurobiology of Disease, Vol. 7,
No. 4, pp. 299-309.
Torchia, D. (1978). The Measurement of Proton-Enhanced Carbon -13 T1 Values by a Method
Which Suppresses Artifacts. Journal of Magnetic Resonance, Vol. 30, pp. 613-616.
Tycko, R. (2007). Symmetry-based constant-time homonuclear dipolar recoupling in solid state
NMR. Journal of Chemical Physics, Vol. 126, No. 6, pp. 064506.
Tycko, R. (2011). Solid-state NMR studies of amyloid fibril structure. Annu Rev Phys Chem,
Vol. 62, pp. 279-99.
Tycko, R. and Ishii, Y. (2003). Constraints on supramolecular structure in amyloid fibrils from
two-dimensional solid-state NMR spectroscopy with uniform isotopic labeling. Journal
of the American Chemical Society, Vol. 125, No. 22, pp. 6606-7.
Tycko, R., Kheterpal, I. and Wetzel, R. (2006). Characterization of Amyloid Structures at the
Molecular Level by Solid State Nuclear Magnetic Resonance Spectroscopy. Methods in
Enzymology, Academic Press. Volume 413: 103-122.
Um, J. W., Nygaard, H. B., Heiss, J. K., Kostylev, M. A., Stagi, M., Vortmeyer, A., Wisniewski,
T., Gunther, E. C. and Strittmatter, S. M. (2012). Alzheimer amyloid-beta oligomer
bound to postsynaptic prion protein activates Fyn to impair neurons. Nat Neurosci, Vol.
15, No. 9, pp. 1227-35.
Uversky, V. N., Lee, H. J., Li, J., Fink, A. L. and Lee, S. J. (2001). Stabilization of partially
folded conformation during alpha-synuclein oligomerization in both purified and
cytosolic preparations. J Biol Chem, Vol. 276, No. 47, pp. 43495-8.
Vaiana, S. M., Ghirlando, R., Yau, W.-M., Eaton, W. A. and Hofrichter, J. (2008).
Sedimentation studies on human amylin fail to detect low-molecular-weight oligomer.
Biophysical Journal, Vol. 94, pp. L45-L47.
van der Wel, P. C., Lewandowski, J. R. and Griffin, R. G. (2007). Solid-state NMR study of
amyloid nanocrystals and fibrils formed by the peptide GNNQQNY from yeast prion
protein Sup35p. Journal of the American Chemical Society, Vol. 129, No. 16, pp. 5117-
30.
van der Wel, P. C., Lewandowski, J. R. and Griffin, R. G. (2007). Structural characterization of
GNNQQNY amyloid fibrils by magic angle spinning NMR. Biochemistry, Vol. 49, No.
44, pp. 9457-69.
144
Veshtort, M. and Griffin, R. G. (2006). SPINEVOLUTION: a powerful tool for the simulation of
solid and liquid state NMR experiments. Journal of Magnetic Resonance (1969), Vol.
178, No. 2, pp. 248-82.
Vilar, M., Chou, H. T., Luhrs, T., Maji, S. K., Riek-Loher, D., Verel, R., Manning, G., Stahlberg,
H. and Riek, R. (2008). The fold of alpha-synuclein fibrils. Proceedings of the National
Academy of Sciences of the United States of America, Vol. 105, No. 25, pp. 8637-42.
Vilar, M., Wang, L. and Riek, R. (2012). Structural studies of amyloids by quenched hydrogen-
deuterium exchange by NMR. Methods Mol Biol, Vol. 849, pp. 185-98.
Walsh, D. M., Klyubin, I., Fadeeva, J. V., Cullen, W. K., Anwyl, R., Wolfe, M. S., Rowan, M. J.
and Selkoe, D. J. (2002). Naturally secreted oligomers of amyloid beta protein potently
inhibit hippocampal long-term potentiation in vivo. Nature, Vol. 416, No. 6880, pp. 535-
9.
Walsh, D. M. and Selkoe, D. J. (2004). Oligomers on the brain: the emerging role of soluble
protein aggregates in neurodegeneration. Protein & Peptide Letters, Vol. 11, No. 3, pp.
213-28.
Walsh, D. M. and Selkoe, D. J. (2007). A beta oligomers - a decade of discovery. Journal of
Neurochemistry, Vol. 101, No. 5, pp. 1172-84.
Walsh, D. M., Tseng, B. P., Rydel, R. E., Podlisny, M. B. and Selkoe, D. J. (2000). The
oligomerization of amyloid beta-protein begins intracellularly in cells derived from
human brain. Biochemistry, Vol. 39, No. 35, pp. 10831-9.
Wasmer, C., Lange, A., Van Melckebeke, H., Siemer, A. B., Riek, R. and Meier, B. H. (2008).
Amyloid fibrils of the HET-s(218-289) prion form a beta solenoid with a triangular
hydrophobic core. Science, Vol. 319, No. 5869, pp. 1523-6.
Weissmann, C. (2004). The state of the prion. Nature Reviews Microbiology, Vol. 2, No. 11, pp.
861-71.
Wickner, R. B. (1994). [URE3] as an altered URE2 protein: evidence for a prion analog in
Saccharomyces cerevisiae. Science, Vol. 264, No. 5158, pp. 566-9.
Wickner, R. B., Masison, D. C. and Edskes, H. K. (1995). [PSI] and [URE3] as yeast prions.
Yeast, Vol. 11, No. 16, pp. 1671-85.
Wilkins, D. K., Grimshaw, S. B., Receveur, V., Dobson, C. M., Jones, J. A. and Smith, L. J.
(1999). Hydrodynamic Radii of Native and Denatured Proteins Measured by Pulse Field
Gradient NMR Techniques Biochemistry, Vol. 38, No. 50, pp. 16424-16431.
Williams, A. D., Sega, M., Chen, M., Kheterpal, I., Geva, M., Berthelier, V., Kaleta, D. T.,
Cook, K. D. and Wetzel, R. (2005). Structural properties of Abeta protofibrils stabilized
by a small molecule. Proceedings of the National Academy of Sciences of the United
States of America, Vol. 102, No. 20, pp. 7115-20.
145
Wiltzius, J. J., Sievers, S. A., Sawaya, M. R., Cascio, D., Popov, D., Riekel, C. and Eisenberg, D.
(2008). Atomic structure of the cross-beta spine of islet amyloid polypeptide (amylin).
Protein Science, Vol. 17, No. 9, pp. 1467-74.
Wishart, D. S. and Sykes, B. D. (1994). The 13C chemical-shift index: a simple method for the
identification of protein secondary structure using 13C chemical-shift data. Journal of
Biomolecular NMR, Vol. 4, No. 2, pp. 171-80.
Xue, W. F., Hellewell, A. L., Gosal, W. S., Homans, S. W., Hewitt, E. W. and Radford, S. E.
(2009). Fibril fragmentation enhances amyloid cytotoxicity. Journal of Biological
Chemistry, Vol. 284, No. 49, pp. 34272-82.
Yip, C. M., Darabie, A. A. and McLaurin, J. (2002). Abeta42-peptide assembly on lipid bilayers.
Journal of Molecular Biology, Vol. 318, No. 1, pp. 97-107.
Yip, C. M. and McLaurin, J. (2001). Amyloid-beta peptide assembly: a critical step in
fibrillogenesis and membrane disruption. Biophys J, Vol. 80, No. 3, pp. 1359-71.
Yu, L., Edalji, R., Harlan, J. E., Holzman, T. F., Lopez, A. P., Labkovsky, B., Hillen, H.,
Barghorn, S., Ebert, U., Richardson, P. L., Miesbauer, L., Solomon, L., Bartley, D.,
Walter, K., Johnson, R. W., Hajduk, P. J. and Olejniczak, E. T. (2009). Structural
Characterization of a Soluble Amyloid beta-Peptide Oligomer. Biochemistry, Vol. 48,
No. 9, pp. 1870-1877.
Yu, X., Wang, Q. and Zheng, J. (2010). Structural determination of Abeta25-35 micelles by
molecular dynamics simulations. Biophysical Journal, Vol. 99, No. 2, pp. 666-74.
Yu, X. and Zheng, J. (2011). Polymorphic Structures of Alzheimer's beta-Amyloid Globulomers.
PLoS One, Vol. 6, No. 6, pp. e20575.
Zahn, R., Liu, A., Luhrs, T., Riek, R., von Schroetter, C., Lopez Garcia, F., Billeter, M.,
Calzolai, L., Wider, G. and Wuthrich, K. (2000). NMR solution structure of the human
prion protein. Proceedings of the National Academy of Sciences of the United States of
America, Vol. 97, No. 1, pp. 145-50.
Zhang, A., Qi, W., Good, T. A. and Fernandez, E. J. (2009). Structural differences between
Abeta(1-40) intermediate oligomers and fibrils elucidated by proteolytic fragmentation
and hydrogen/deuterium exchange. Biophysical Journal, Vol. 96, No. 3, pp. 1091-104.
Zhang, R., Hu, X., Khant, H., Ludtke, S. J., Chiu, W., Schmid, M. F., Frieden, C. and Lee, J. M.
(2009). Interprotofilament interactions between Alzheimer's Abeta1-42 peptides in
amyloid fibrils revealed by cryoEM. Proceedings of the National Academy of Sciences of
the United States of America, Vol. 106, No. 12, pp. 4653-8.
Zheng, J., Jang, H., Ma, B. and Nussinov, R. (2008). Annular structures as intermediates in fibril
formation of Alzheimer Abeta17-42. Journal of Physical Chemistry B, Vol. 112, No. 22,
pp. 6856-65.
146
Zhu, Y. J., Lin, H. and Lal, R. (2000). Fresh and nonfibrillar amyloid beta protein(1-40) induces
rapid cellular degeneration in aged human fibroblasts: evidence for AbetaP-channel-
mediated cellular toxicity. FASEB Journal, Vol. 14, No. 9, pp. 1244-54.
147
Appendices Appendix 1 – Additional NMR data for PrP(106-126) Fibrils
N CO Cα Cβ Cγ
A113 121.1 173.3 49 20.7
G114 106.7 168.8 43.6
A115 21.2
A116 121.8 172.8 49.1 21.7
A117 172.7
A118 119.1 173.7 49.4 21.8/23.7†
G119 168.8
A120 172.9
V121 118.2 170.8 57.1 34.0 19.1
V122 122.9 171.9 57.2 33.9 19.3
G123 109.5 169.6 43.7
G124 43.8
L125 (121.2) (172.5) (51.8) 44.7 23.8
G126 113.3 174.7 45.7
† 3:1
Table A-1 - 13
C and 15
N chemical shift assignments for PrP(106-126) fibrils.
Peaks with broad (> 3ppm) NMR linewidths have their assignment in parentheses. Where two peaks are observed
for a single site, the ratio of the cross peak volumes (from 13
C-13
C correlation spectra with 10 ms RAD mixing) is
indicated.
148
Residue (°) (°)
Gly114 -143 (±16) 144 (±24)
Ala115 -129 (±23) 144 (±14)
Ala116 -140 (±14) 144 (±13)
Ala117 -132 (±10) 150 (±14)
Ala118 -127 (±11) 142 (±13)
Gly119 -149 (±14) 159 (±14)
Ala120 -110 (±27) 140 (±14)
Val121 -137 (±15) 143 (±12)
Val122 -128 (±11) 142 (±18)
Gly123 -137 (±19) 148 (±23)
Gly124 -97 (±27) 142 (±28)
Leu125 -134 (±17) 140 (±18)
Table A-2 - Backbone and torsion angles predicted for the sheet-forming region of PrP(106-126) using
TALOS analysis of 13
C and 15
N chemical shifts.
149
Figure A-1 - 13
C and 15
N chemical shift correlation spectra for PrP(106-126)AVG
fibrils.
A 2D 13
C-13
C correlation spectrum obtained at 10 kHz MAS, using a 10 ms RAD mixing period is shown in the
upper panel, with a 13
C-15
N heteronuclear correlation spectrum in the lower panel. Cross peak assignments are
shown using internuclear connectivities in the 13
C-13
C correlation spectrum, while the intraresidue N-C cross peaks
are identified in the 13
C-15
N spectrum. In the direct dimension, 1024 complex points were taken with a dwell in t1 of
25 μs and 200 complex points in the indirect dimension. A 10 kHz 1H field was applied during the RAD period and
64 scans were taken per FID.
150
Figure A-2 -
13C and
15N chemical shift correlation spectra for PrP(106-126)
AVG2 fibrils
A 2D 13
C-13
C correlation spectrum obtained at 10 kHz MAS, using a 10 ms RAD mixing period is shown in the
upper panel, with a 13C-15N heteronuclear correlation spectrum in the lower panel. Cross peak assignments are
shown using internuclear connectivities in the 13
C-13
C correlation spectrum, while the intraresidue N-C cross peaks
are identified in the 13
C-15
N spectrum. In the direct dimension, 1024 complex points were taken with a dwell in t1 of
25 μs and 200 complex points in the indirect dimension. A 10 kHz 1H field was applied during the RAD period and
64 scans were taken per FID.
151
Appendix 2 – Additonal data for PrP(106-126) Oligomers
Figure A-3 – Thioflavin-T fluorescence of PrP(106-126) Oligomers Under Various Conditions
Curves represent the emission spectrum of ThT recorded with an excitation wavelength of 442 nm. Peptide
oligomer concentrations of 60 μM (red) and 520 μM are shown (turquoise) at pH4.6 and do not exhibit ThT
fluorescence at 482 nm. Likewise, no ThT emission is observed for 60 μM oligomeric PrP(106-126) at pH 8.0.
Each curve is normalized relative to a ThT containing blank, and to the fluorescence intensity observed at 482 nm
for ThT in the presence of 60 μM fibrillar PrP(106-126) at pH 8.0 (blue curve).
152
1H chemical shifts (ppm)
Residue HN Hα Hβ Hγ Hδ Hε
K106 (4.07) (1.91) (1.44) (1.70) (3.00)
T107 4.35 4.12 1.19
*N108 8.28 4.60 3.31,3.37 7.86,7.96
M109 8.40 4.42 1.93,2.04 2.47,2.54
K110 (8.20) (4.25) (1.35) (1.64) (2.95)
H111 8.29 4.55 3.09 7.02,7.00 7.36,7.62
M112 4.47 2.13,2.24 3.17,3.29
A113 8.41 4.29 1.38
G114 8.27 3.90
A115 (3.96)
A116 (3.92)
A117 (3.74)
A118 (3.96)
G119 8.28 3.90
A120 7.85 4.30 1.35
V121 8.26 4.10 2.03 0.88,0.91
V122 8.19 4.09 2.04 0.93
G123 8.02 3.58,3.35
G124 (8.02) (3.54,3.30)
L125 8.22 4.18 1.67,1.78 1.31 0.86
G126 (8.27) (3.60,3.35)
Table A-3 – 1H Chemical Shifts for PrP(106-126) Oligomers
Ambiguous assignments are listed in brackets, and shifts obtained from solid state NMR data are in italics.
153
13C chemical shifts (ppm)
Residue CO Cα Cβ Cγ Cδ Cε
K106 (55.82) (33.34) (23.97) (29.22)
T107 62.42 70.00 21.56
*N108 41.38
M109 32.81 33.75 25.04
K110 (52.52) (30.01)
H111 29.56 117.5 132.05
M112 51.28 28.84
A113 175.6 51.4 23.7 24.93
G114 170.5 46.0
A115 (48.23) 22.6
(18.63)
A116 174.6 54.5
(45.17)
23.7
(18.57)
A117 174.3 (46.09) (19.81)
A118 175.1 51.4
(48.23)
23.6
(18.63)
G119 170.3
A120 174.9 19.33
V121 173.0 59.4 35.3 20.9
21.17,20.7
V122 173.4 59.4 35.7 21.2
25.11,23.24
G123 171.5 45.7
154
45.7
G124 45.2
(41.28)
L125 174.4 54.2
56.6
45.8
33.44
26.0
24.28
G126 177.0
47.3
(41.16)
Table A-4 – 13
C chemical shifts for PrP(106-126) Oligomers
Ambiguous assignments are listed in brackets, and shifts obtained from solid state NMR data are in italics.