A STUDY OF EARLY STAGES IN PROTEIN AGGREGATION
By
Srabasti Acharya

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Physics - Doctor of Philosophy
Quantitative Biology - Dual Major
2014

ABSTRACT
A STUDY OF EARLY STAGES IN PROTEIN AGGREGATION
By
Srabasti Acharya
Protein aggregation has been widely associated with neurodegenerative diseases like
Alzheimer’s and Parkinson’s. One of the challenges associated with treating these diseases is that
by the time people go to the doctor, the aggregation process has already spread all over their
brain and killed a big chunk of the brain cells. The biochemical changes start happening at least a
decade before the actual symptoms occur. In order to treat these diseases, it is essential to catch
the aggregation process very early. This thesis presents a novel biophysical approach of attacking
the aggregation problem, which is aimed at early detection of neurodegenerative diseases.
My findings suggest that monomer reconfiguration controls early steps in protein
aggregation. When reconfiguration is fast, bimolecular association is not stable and can be
disrupted easily, but as reconfiguration slows, association is more stable and the likelihood of
aggregation increases. This hypothesis is tested in both Aβ peptide and α-synuclein.
Furthermore, we find that mutations that have been associated with these diseases promote
aggregation by slowing down reconfiguration. On the other hand, good mutations that inhibit
aggregation speed up reconfiguration and thus reduce the chances of the protein from getting
trapped into making stable intermolecular interactions. This suggests that speeding up
reconfiguration can be used as an effective strategy of rescuing proteins from aggregation. Using
this technique, we have identified an inhibitor molecule which can become a potential drug
candidate for fighting neurodegeneration.

ACKNOWLEDGEMENTS

The central factor that kept me from losing my sanity during the course of my graduate
years was my advisor Dr. Lisa Lapidus. When I joined her lab, I had absolutely no prior
experience in conducting independent research. She walked me in baby steps, ushering me to a
point where I can design, conduct and analyze experiments independently. Her perpetual
tolerance towards my ignorance has encouraged me to shed all inhibitions in raising questions
and clarifying daily doubts. The long open discussions I had with her often transcended beyond
research, triggering a bond of camaraderie and comfort. Without her constant guidance, support
and motivation, this research journey would not have been possible.
I am thankful to my committee members for providing me with constructive feedback on
my research projects and presentations. I am especially grateful to Dr. Mahanti, who has been a
father figure to me from the day I started at MSU. He has been a very kind mentor and has
always addressed my concerns regarding courses, research and even future plans.
I would like to thank all my lab members, past and present, for creating such an amicable
work environment. I have been surrounded by exceptionally talented colleagues, always willing
to share their scientific wisdom. I am especially grateful to Kinshuk for his invaluable research
tips and suggestions; the frequent scientific discussions with him have been very fulfilling and
inspired me to read more papers. I would also like to thank Aimee and Dena for being such great
friends and giving me splendid memories of working together.

iii

My mother shielded me through thick and thin, and was always just one phone call away.
Without the financial support of my father, I would not have made it to grad school. I am
obligated to both of them for never giving up on me.
I am endlessly indebted to my husband, Ritesh, for supporting me through moments of
extreme hopelessness and desperation. I would have quit grad school long ago, hadn’t it been for
his undying conviction in my potential. He gifted me with a daily dose of fun and laughter,
sometimes in person and sometimes through skype. His occasional trips to East Lansing
rejuvenated my spirits and motivated me to invest more time into research; that turned out to be
the driving force behind my accomplishments.

iv

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................................ vii
LIST OF FIGURES ..................................................................................................................... viii
Chapter 1 ..........................................................................................................................................1
Introduction to Protein Aggregation ................................................................................................1
1.1 Protein Folding...........................................................................................................................1
1.2 Protein Misfolding and Aggregation .........................................................................................2
1.3 Reconfiguration and Aggregation ..............................................................................................6
Chapter 2 ........................................................................................................................................11
Measuring Intramolecular Contact Formation ...............................................................................11
2.1 Method .....................................................................................................................................11
2.2 Data Analysis Using SSS theory..............................................................................................16
2.3 Boltzmann Energy Reweighting ..............................................................................................21
2.4 The Charge Effect ....................................................................................................................24
Chapter 3 ........................................................................................................................................26
Protein Aggregation in Alzheimer’s disease..................................................................................26
3.1 Aβ Production ..........................................................................................................................26
3.2 Intracellular Aβ Accumulation ................................................................................................27
3.3 Aβ Associated with Alzheimer’s disease .................................................................................28
3.4 Neurotoxic Effects of Aβ .........................................................................................................29
3.5 Structure of Aβ Monomer ........................................................................................................30
3.6 Dimers and Trimers .................................................................................................................31
3.7 Higher Order Aggregates .........................................................................................................31
3.8 Comparison of Aβ42 and Aβ42 ..................................................................................................34
3.9 Expression and Purification Protocol.......................................................................................35
3.10 Thioflavin T Assay ................................................................................................................38
3.11 Trp Cys Quenching Measurement .........................................................................................39
3.12 Mutation Sites ........................................................................................................................40
3.13 Measuring Reconfiguration ...................................................................................................44
Chapter 4 ........................................................................................................................................56
Protein Aggregation in Parkinson’s disease ..................................................................................56
4.1 Introduction ..............................................................................................................................56
4.2 Physiological Function.............................................................................................................58
4.3 Effects of Oligomers on Neurotoxicity ....................................................................................59
4.4 Structure ...................................................................................................................................61
4.5 Mutations Associated with PD.................................................................................................64
4.6 Expression and Purification Protocol.......................................................................................68
4.7 Contact Quenching Experiment ...............................................................................................69
v

4.8 Effect of Good and Bad Mutations on Reconfiguration ..........................................................69
4.9 Effect of Inhibitors on Reconfiguration ...................................................................................79
4.10 Molecule that Slows Down Reconfiguration .........................................................................81
4.11 Molecule Tweezer ..................................................................................................................84
Chapter 5 ........................................................................................................................................95
Conclusion .....................................................................................................................................95
BIBLIOGRAPHY ..........................................................................................................................98

vi

LIST OF TABLES

Table 3.1 Parameters for two different polymeric models that best fit the experimental data
……………………………………………………………………………………………………55
Table 4.1 Parameters obtained from Gaussian analysis for aggregation prone
mutations…………………………………………………………………………………………93

vii

LIST OF FIGURES

Figure 1.1 Two amino acids forming a peptide bond by removal of a water molecule
(http://molecularsciences.org/files/images/peptidebond.jpg)…………………………..................9
Figure 1.2 Energy landscape connecting protein folding and aggregation [11]
(http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2706318/#!po=45.1220)....................................9
Figure 1.3 Kinetic scheme describing the early phases of aggregation [17]
(http://pubs.rsc.org/en/Content/ArticleLanding/2013/MB/c2mb25334h#!divAbstract)...............10
Figure 1.4 Oligomer formation rate with time at different reconfiguration rates [17]
(http://pubs.rsc.org/en/Content/ArticleLanding/2013/MB/c2mb25334h#!divAbstract)...............10
Figure 2.1 Determination of the rate of contact formation between the probe, tryptophan (W),
and the quencher, cysteine (C), at the ends of a flexible peptide. Pulsed optical excitation leads
to population of the lowest excited triplet state of tryptophan. Tryptophan contacts cysteine in a
diffusion-limited process with rate kD+, and then either diffuses away or is quenched by the
cysteine…………………………………………………………………………………………..13
Figure 2.2 A typical Gaussian polymer chain made of N rigid rods connected together by flexible
pivot points……………………………………………………………………………................17
Figure 2.3 A schematic representation of worm like chain model showing the constraint
parameters parameter length
and excluded volume diameter ……………………………..20
Figure 2.4 Experimental setup for measuring contact quenching……………………………......25
Figure 3.1 Sequence of Aβ42 WT with arrows showing the positions of Trp and Cys mutations
for the 2 loops created…………………………………………………………………………...42
Figure 3.2 Thioflavin T fluorescence spectra of Aβ42 WT along with the 2 Trp/Cys containing
mutants…………………………………………………………………………………………...42
Figure 3.3 The CD spectra of Aβ42 WT, and the 2 loops containing Trp/Cys, both for Aβ42 and
Aβ40………………………………………………………………………………………………43
Figure 3.4 Decay of Trp triplet states W19C35 at pH 7.5, 0 C as monitored by the decay of
absorption of 445 nm light and captured through oscilloscope………………………………….50
viii

Figure 3.5 The fast lifetimes of Trp triplet at pH 7.5 for various viscosities and temperatures as
marked. a) Aβ42 M35C, F19W loop, pH 7.5 b) Aβ42 F4C, F19W loop, pH 7.5 c) Aβ40 M35C,
F19W loop, pH 7.5 d) Aβ40 F4C, F19W loop, pH 7.5 e) Aβ42 M35C, F19W loop, pH 10 f) Aβ42
M35C, F19W loop, pH 7.5: Curcumin…………………………………………………………..51
Figure 3.6 Comparison of intramolecular diffusion of the 2 loops of Aβ40 and Aβ42 (peptide
concentration =30 µM) a) Measured reaction-limited rates, as determined from the intercept of
plots such as Fig. 9 (a, b, c, d) versus temperature b) Measured diffusion-limited rates as
determined from the slope of plots such as Fig. 9 a, b, c, d normalized to the viscosity of water at
each temperature. Errors from the linear fits are shown as error bars in (a) and (b). c) Average
Trp-Cys distance calculated from the reweighted worm-like chain (WLC) model as described
above d) Intramolecular diffusion coefficients determined using the WLC model……………..52
Figure 3.7 Modulation of reconfiguration of Aβ42 by altering solution conditions and adding an
inhibitor a) Measured intramolecular diffusion coefficients at pH 7.5 and pH 10 for Aβ42 M35C,
F19W loop b) Average Trp-Cys distance at pH 7.5 and pH 10 for Aβ42 M35C, F19W loop
determined using the WLC model from measured reaction and diffusion-limited rates obtained
from plots shown in Fig. 9e c) Measured intramolecular diffusion coefficients for Aβ42 M35C,
F19W loop, with and without inhibitor Curcumin (peptide: Curcumin=1:1) d) Average Trp-Cys
distance for Aβ42 M35C, F19W loop, with and without curcumin, determined using the WLC
model from measured reaction and diffusion-limited rates obtained from plots shown in Fig.
9f…………………………………………………………………………………………………53
Figure 3.8 Formation of oligomer [O] versus time for Aβ42 (black) and Aβ40 (red) which was
calculated using previously reported kinetic model that correlates reconfiguration with oligomer
formation…………………………………………………………………………………………54
Figure 4.1 (
) versus viscosity at pH 7.5, at various temperatures as marked, for αsynuclein a) Y39W-A69C loop in the wild type sequence b) A53T mutant c) V74E mutant.
Protein concentration=30 µM……………………………………………………………………76
Figure 4.2 Effects of Aggregation-Enhancing, PD-Causing, Familial Mutations (A53T, E46K,
and A30P). a) Reaction-limited rates. b) Diffusion-limited rates, normalized to the viscosity at
water at that temperature. c) Average root mean square distance between W39 and C69
determined from reaction-limited rates d) Intramolecular diffusion coefficients determined for
each diffusion-limited rate……………………………………………………………………….77
Figure 4.3 Effects of Aggregation Reducing Mutations (V74E and T72P). a) Reaction-limited
rates. b) Diffusion-limited rates, normalized to the viscosity of water at that temperature. c)
Average root mean square distance between W39 and C69 determined for each reaction-limited
rate d) Intramolecular diffusion coefficients determined for each diffusion-limited
rate………………………………………………………………………………………………..78
ix

Figure 4.4 (
) versus viscosity at pH 7.5, at various temperatures for inhibitor with αsynuclein Y39W-A69C loop a) Molecular tweezer, CLR01 b) inhibitor Baicalein c) inhibitor
EGCG. Protein concentration=30 µM…………………………………………………………...88
Figure 4.5 Effects of Natural Inhibitors on α-synuclein. a) Reaction-limited rates. b) Diffusionlimited rates, normalized to the viscosity of water at that temperature. c) Average root mean
square distance between W94 and C69, determined from the reaction-limited rates d)
Intramolecular diffusion coefficients determined from diffusion-limited rates…………………89
Figure 4.6 Effect of EGCG on α-synuclein. a) Reaction-limited rates. b) Diffusion-limited rates,
normalized to the viscosity of water at that temperature. c) Average root mean square distance
between W94 and C69, determined from the reaction-limited rates d) Intramolecular diffusion
coefficients determined from diffusion-limited rates…………………………………………....90
Figure 4.7 Concentration dependant effect of Molecular Tweezer on α-synuclein. a) Reactionlimited rates. b) Diffusion-limited rates, normalized to the viscosity of water at that temperature.
c) Average root mean square distance between W94 and C69, determined from the reactionlimited rates d) Intramolecular diffusion coefficients determined from diffusion-limited
rates………………………………………………………………………………………………91
Figure 4.8 Molecular Tweezer CLR01 molecule………………………………………………..92
Figure 4.9 Effect of binding of CLR01 on Trp94 fluorescence of α-synuclein. Protein
concentration was fixed at 1µM and the CLR01 concentration was gradually
increased………………………………………………………………………………………....92
Figure 4.10 Calculated effect of CLR01 binding on α-synuclein conformation. a) Re-weighted
probability distributions, Z(r), for = 0, = 20 for α-synuclein alone (blue), α-synuclein with
one CLR01 bound at Lys10 (cyan), one CLR01 bound at Lys12 (magenta), and α-synuclein
bound to two CLR01 molecules in positions 10 and 12 (red). In each case, the bound CLR01 are
approximated by Lys→Glu substitutions. b) Calculated (red triangles) reaction-limited rates
using Eq. 9 and the probability distributions calculated in (a). The measured reaction-limited
rates are plotted as black circles………………………………………………………………….93
Figure 4.11 Intramolecular contact plots a) E46K and b) V74E. Each plot is the difference
between the mutant and the wildtype sequence. The average distance is calculated from the
reweighted ensembles using Eqs. 6 and 7 for γ= 9……………………………………………...94

x

xi

Chapter 1
Introduction to Protein Aggregation
1.1 Protein Folding
Proteins are the major components of the living cell and play an important role in
maintaining life. They are produced inside our cells by the process of translation, and comprise
of amino acids linked by peptide bonds. There are 20 naturally occurring amino acids, and each
comprise of a backbone that includes a carboxyl and an amino group and a distinctive side chain
(R-group) bonded to the Cα atom (shown in Figure 1.1). Two amino acids may link together such
that the OH of the carboxyl group of one amino acid and H of the amino group of another may
be removed in the form of a water molecule giving rise to a peptide bond connecting the CO and
NH2 groups. This dipeptide may be joined by more amino acids forming a polypeptide chain,
which is also called a protein. Each protein will contain one free amino group which is known as
the N-terminus, one free carboxyl group known as the C-terminus, R groups referred to as
residues, and the linked carbon, nitrogen and oxygen atoms in between C and N-terminus (except
the R-group) referred to as the main chain or backbone. The R-group imparts unique
physiochemical properties to each amino acid. They can be broadly classified as being aliphatic
(alanine, glycine, isoleucine, leucine, proline, valine), aromatic (phenylalanine, tryptophan,
tyrosine), acidic (aspartic acid, glutamic acid), basic (arginine, histidine, lysine), hydroxylic
(serine, threonine) sulphur-containing (cysteine, methionine), amidic (asparagines, glutamine).
However, they may also be classified based on their polarity, size, charge or hydrophobicity.
Most proteins fold into a unique 3-dimensional structure called the native state that is
essential for its proper functioning. The hydrophobic collapse has been widely described as one
1

of the key factors that drive protein folding, apart from other factors like Vander Waals forces,
electrostatic Coulomb interactions, hydrogen bonds, disulfide bridges, etc. Anfinsen in his
famous experiment showed that the folding and unfolding of proteins is reversible and proved
that a protein would spontaneously fold into the most stable state, which suggests that the
information needed to determine the final native conformation is located in the sequence itself
[1]. However, Levinthal’s paradox states that it is impossible for proteins to sample all possible
conformations in physiologically relevant time to attain its native state, and the fact that protein
folding happens in millisecond timescale suggests that there must be a specific folding pathway
guided by native-like interactions already present within the unfolded state. NMR and X-ray
scattering studies have detected the presence of local and long-range contacts, native-like and
non-native interactions within the unfolded

states referred to as residual structures [2].

Experimental and computational studies have further shown that intramolecular diffusion of
unfolded states of well behaved proteins is relatively slow, suggesting a rugged energy
landscape, since a smooth landscape would enable unfolded conformations to reconfigure rapidly
by diffusion without having to slow down [3][4][5].
1.2 Protein Misfolding and Aggregation
Protein aggregation is the process in which non-native conformations of two or more
proteins interact, leading to their self-assembly into a variety of structures called aggregates.
Protein-protein interactions leading to aggregation is unwanted because it often prevents the
protein from adopting its functional state and detrimental because the resulting protein
aggregates often have toxic properties and can lead to neurodegeneration[6]. Evidence collected
over the years suggest that protein aggregation can initiate from different conformations ranging
from a fully unfolded monomer to misfolded states or a partially folded conformation [7]. In all
2

the above cases, aggregation-prone segments that are normally buried in the fully folded state
become exposed to the solvent, initiating intermolecular interactions which can lead to the
formation of amyloid-like oligomers and fibrils. What makes one protein more aggregation
prone than another? Sequence definitely plays a key role and generally sequences that have high
hydrophobicity, low net charge and high β-sheet forming propensity are more likely to
aggregate.
One of the most widely accepted mechanisms for assembly of monomers into aggregates
is the previously reported ‘nucleation growth’ mechanism. According to this mechanism, the
nucleus which could very well be a monomer, acts as a template for growth of an aggregate
through the association of further monomers. Evidence have been found of proteins populating
amyloid conformations at the monomeric level which were described as ‘β-like’ structures that
seemed increase with aggregation promoting conditions [8]. Study of insulin aggregation has
revealed that monomers without undergoing any structural organization directly assemble at the
end of a growing aggregate and lead to elongation of an amyloid like fibril [9]. However, in
some cases evidence were found suggesting that native monomers initially convert into
misfolded conformations, which initiate self-assembly through a template-independent
mechanism with formation of an amyloid-like oligomer that acts as a nucleus, leading to higher
order oligomer formation and then finally fibrils. This mechanism is referred to as ‘nucleated
conformational conversion’ mechanism [7]. It was recently shown that the Aβ40 monomers can
rapidly form small spherical oligomers in vitro that are kinetically competent to slowly convert
to amyloid fibrils following the nucleated conformational conversion [10]. The two models differ
in many ways. In the former model, it is believed that the monomer may directly interact with
the nucleus, the nucleus may be as small as a monomer and the rate-limiting step is considered to

3

be the formation of the nucleus. The latter model proposes that both the building block as well as
the nucleus is the oligomeric species, and the rate limiting step is the conversion of the nonamyloid oligomer into an amyloid-competent oligomer; once that happen amyloid formation
advances rapidly. One view that could possibly unify the two models is that the first step of
aggregation requires the monomer conformational ensemble to be in a state which exposes their
hydrophobic patches long enough to the solvent, which can initiate bimolecular association in
between two proteins.
Funnel-shaped free energy landscapes provide a useful description of protein folding in
isolation, but the behavior of proteins in the living cells are much more complex and competition
between intramolecular and intermolecular interactions will result in increasing ruggedness to
the energy landscape. Figure 1.2 shows a previously reported schematic representation of an
energy landscape that takes into account both protein folding and aggregation [11]. It shows that
a wide range of different conformational states and multitude of pathways may be available to
each protein as it navigates through the landscape. Inside living cells, the relative depth of each
well will depend upon the protein sequence, solution conditions, temperature and also protein
concentration. Energy minima on the aggregation side of the energy landscape are also poorly
defined. Sometimes, under same set of solution conditions, a multitude of fibrillar or oligomeric
morphologies may be formed, which adds further complexity and multiplicity to the aggregation
pathway. However one simple fact that emerges out of such complexity is that protein folding
and aggregation are intimately connected, and somewhere along the folding pathway the
sequence may take a wrong turn and aggregation happens as an off-pathway event or that some
sequences favor the aggregation well over the folding well. Cells have sophisticated quality
control mechanisms to ensure that a newly synthesized protein is properly folded before being

4

released. When misfolded proteins accumulate within cells, organelles like the endoplasmic
reticulum (ER) activates a complex signaling network called unfolded protein response (UPR) to
cope with the stress [12] [13]. UPR increases expression of chaperone proteins which help to
fold the proteins properly and also directs the damaged proteins towards degradation via either
ubiquitin–proteasome system or lysosomal autophagy. When such attempts to re-establish ER
homeostasis fail misfolded proteins assemble into toxic aggregates, causing cellular dysfunction
and death.
There is another interesting class of proteins called the intrinsically disordered proteins
(IDPs) which challenges the widely accepted structure-function paradigm. Such IDPs lack stable
secondary and or tertiary structures, yet fulfill regular biological functions. They could either
exist in state of extended disorder due to repulsion from a net charge, which would resemble an
ideal random coil, or a semi-collapsed state containing transient secondary structures referred to
as pre-molten globule. Such natively unfolded proteins are typically characterized by a low
overall hydrophobicity and large net charge [14]. In comparison with structured proteins, IDPs
mostly lack structure-promoting residues (like cysteine, tryptophan, tyrosine, isoleucine,
phenylalanine, valine, leucine, histidine, threonine, asparagine) and are relatively enriched in the
disorder-promoting residues (aspartic acid, methionine, lysine, arginine, serine, glutamine,
proline, and glutamic acid) [15]. Due to lack of a folded structure, the hydration of IDPs is
significantly higher than that of globular proteins of similar size, suggesting that the chain has
largely exposed interaction surfaces [16].

5

1.3 Reconfiguration and Aggregation
Aggregation may be defined as the physical process in which two or more proteins may
bump into each other and end up clumping together. Within the crowded environment of the
cells, the proteins have great chances of bumping across each other, so that raises the very
important question as to why they don’t always aggregate. Therefore there must be an escape
route which needs to be taken into account while describing aggregation. One aggregation
model that accommodates this idea has been reported recently [17]. This model talks about the
very first step in aggregation from the view of monomer dynamics. A monomeric protein at any
time may be sampling different kinds of conformations, a process called protein reconfiguration.
Some of those conformations may be aggregation competent monomers (M*), which may have
hydrophobic patches sticking out at the surface; while some of them may be aggregation
incompetent monomers (M). Assuming that there is no significant energy barrier between the
states, at any time M and M* may be inter-convertible by diffusive reconfiguration, the rate of
which is given by

. The reconfiguration rate can be estimated by assuming that it is the

rate to diffuse from one point on the chain across the chain diameter, given by
where D is the intramolecular diffusion coefficient and

is the radius of gyration of the chain.

For the aggregation process to initiate, two M* conformers have to first come together by
a bimolecular diffusion controlled rate

and form an encounter complex [M*M*]. While both

monomers are in the encounter complex two things may happen; they may make stabilizing
bimolecular interactions that lead to an oligomeric state O or one or both the monomers may
escape by reconfiguration to the innocuous conformation M which makes stabilizing bimolecular
interactions difficult and the complex comes apart. Thus there are three possibilities:

6

1) When

, the reconfiguration is really fast and before stable associations happen,

one or both M* conformers reconfigures back to an innocuous conformation M.
2) When

the reconfiguration happens really slowly, because of which M* takes a

long time to form and so the likelihood of two M* conformers coming together is really
unlikely. Hence the encounter complex is rarely made.
3) When

the reconfiguration happens with an optimum speed such that two M*

can react to form a stable encounter complex. When that happens, the encounter complex
can proceed to form an oligomeric conformation O.
The previously reported kinetic scheme shown in Figure 1.3 takes into account these
three possibilities [17]. It used the bimolecular association rate

105 s-1 calculated for a

solution of 100µM protein of average molecular weight in water and a range of

values to

solve the model. The model assumes that the rate of forming O is much slower compared to rate
of formation of the encounter complex, because it would require further structural reorganization
and may even need the addition of a third aggregation competent monomer M*. Under this
assumption

was chosen to be 100s-1, an arbitrary rate that is slower than the bimolecular

association rate and fast enough to allow solving the model within reasonable computation time.
The kinetic equations governing the scheme were solved in Matlab to calculate the oligomer O
formation rate with time. Figure 1.4 shows the plot of different oligomer formation rates at
different

values, ranging from 103 to 105s-1. It has been previously reported that the

reconfiguration rate of α-synuclein is

105s-1 at 40°C [18]. Interestingly, Figure 4 shows that

the oligomer formation rate is most rapid when bimolecular association rate matches the
105s-1. The physical significance of

reconfiguration rate of α-synuclein such that

this is that aggregation results from kinetic competition between reconfiguration and bimolecular
7

association. When reconfiguration happens at an optimum speed, the aggregation competent
monomers get enough time to interact and stick, thus forming stable bimolecular associations in
the form of aggregates.
The rate of forming O is assumed to be slower than reconfiguration, but an increased
does not affect the overall trend of maximal oligomerization occurring when
escape rate

may be different from the diffusive reconfiguration rate

possibility into account the model was also solved in the scenario such that

. Also the
. To take that
where

is some fraction determined by the equilibrium constant for M and M*. That increased the
oligomer formation rates, but the overall trend of maximal
that

remained unchanged, suggesting

assumption is not unreasonable. Clearly this aggregation model is quite simple

and the real process could be more complicated. However, since it has been able to predict the
aggregation behavior of α-synuclein correctly and also it is the only model in current literature
that explains aggregation from the monomer point of view, I will use this model in good
confidence to explain my experimental measurements in subsequent chapters.
In this thesis, I will describe the experimental technique to measure protein
reconfiguration in chapter 2 and then go on to test the above mentioned aggregation hypothesis
using two sequences, namely Aβ peptide and α-synuculein protein. In chapter 3, I will talk about
Aβ peptide aggregation in connection with Alzheimer’s disease and using aggregation model, I
will compare the early aggregation propensity of Aβ peptide based on its length, pH and addition
of an aggregation inhibitor. In chapter 4, I will talk about α-synuculein protein in connection
with Parkinson’s disease and then show that its reconfiguration rates can be moved around using
inhibitors or point mutations.

8

Figure 1.1 Two amino acids forming a peptide bond by removal of a water molecule
(http://molecularsciences.org/files/images/peptidebond.jpg)

Figure 1.2 Energy landscape connecting protein folding and aggregation [11].
(http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2706318/#!po=45.1220)
9

Figure 1.3 Kinetic scheme describing the early phases of aggregation [17].
(http://pubs.rsc.org/en/Content/ArticleLanding/2013/MB/c2mb25334h#!divAbstract)

Figure 1.4 Oligomer formation rate with time at different reconfiguration rates [17].
(http://pubs.rsc.org/en/Content/ArticleLanding/2013/MB/c2mb25334h#!divAbstract)

10

Chapter 2
Measuring Intramolecular Contact Formation
In this chapter I will describe the experimental technique for measuring protein reconfiguration
and the computation methods of analyzing the data.
2.1 Method
Intramolecular loop formation is one of the very early stages in protein folding. It can be
probed by using a pump probe laser spectroscopy method that has been described previously
[19]. The mechanism involves introducing two probes into the sequence and then monitoring
contact formation between the two probes. We have used two naturally occurring amino acids,
namely tryptophan (Trp) and cysteine (Cys) as the two probes, which were incorporated into the
protein sequence in a manner which caused negligible perturbation to the original structure and
aggregation propensity of the sequence. Sequences should ideally contain only one Trp and Cys
so that the measured quenching rate reflects only one intrachain contact formation. Normally
multiple Trp-Cys loops are created in different mutants to cover the entire protein length, and
also to account for positional dependence of the probes, if any.
Tryptophan is an ideal probe in this case, since it can be excited to long-lived triplet
energy states upon absorption of 289 nm UV light. Reportedly, tryptophan triplet life time is
about 40µs in water and even longer in the hydrophobic core of the folded protein [20]. Cysteine
is a very efficient quencher (400 fold faster than other amino acids) and can quench the excited
state of tryptophan when nearby, thus cutting short the triplet state life time. The mechanism of
quenching reportedly occurs by electron transfer and has an exponential dependence[21] on the
intramolecular Trp-Cys distance and is given by the following equation:11

Where
and

is the intramolecular distance,

is the distance of closest approach (4Å),

=4.2ns-1

= 40nm-1.
The triplet state of tryptophan absorbs visible light at 442nm wavelength, which can be

used as a probe to monitor the population of the triplet state. The observed decay rate is a
combination of several decay rates:

Where

is the decay rate in the absence of any quencher,

and

are the unimolecular

and bimolecular quenching rates respectively. Since the bimolecular rate (2 x 108 M-1s-1) is
concentration dependent, by keeping the concentration low (around 30 µM) we ensure that the
contribution of the bimolecular rate is negligible[19]. Secondly, the Trp and Cys have to be
engineered close enough in sequence so that the unimolecular contact rate is much faster than
(2 x 104s-1) and dominates.

12

C
h

kD+

q

*
kD*
W
oo
Figure 2.1 Determination oo
of the rate of contact formation between the probe, tryptophan (W),
and the quencher, cysteineoo
(C), at the ends of a flexible peptide. Pulsed optical excitation leads
to population of the lowestoo
excited triplet state of tryptophan. Tryptophan contacts cysteine in a

diffusion-limited process with rate kD+, and then either diffuses away or is quenched by the
cysteine.

The microsecond-nanosecond intramolecular loop formation can be explained well by the
simple 2-step kinetic model that has been illustrated in Figure2.1. Tryptophan is first excited to
its triplet state by UV light; since the sequence diffuses freely the tryptophan and cysteine may
come in close contact towards each other at a rate
at a quenching rate

or diffuses away at rate

. Cysteine either quenches the tryptophan
.

Assuming the following equation approximates the above model:-

Steady state approximation leads to the following equations:-

The observed rate

is given by:-

13

Three situations may arise:(1)

(diffusion-limited) then

(2)

(reaction limited) then

(3)

(experimentally feasible):-

(diffusion limited rate)
(reaction limited rate)

In the ideal case of a diffusion process, as shown by Smoluchowski, the reaction occurs
only when the reacting pair comes within a center-to-center distance

(0.4nm) For distances less

than or equal to a, reaction is instantaneous and irreversible, while at distances larger than a,
reaction rate is assumed to be zero. The Smoluchowski rate is given by

, where

is the sum of diffusion coefficients of the reactants. In real systems, there is distance
dependence to this rate, as exhibited by equation (1) and hence
same way as before. Also since the quenching rate

cannot be approximated in the

is comparable to

, there is a good

chance for tryptophan to escape quenching due to the encounter complex breaking apart. Hence
the measured rate would be a mixture of reaction-limited and diffusion-limited rate and does not
solely reflect diffusion dynamics. The former depends only on temperature (as temperature
affects the chemical reaction speed, viscosity does not) while the latter depends on both
temperature and viscosity of the solution. This gives us a chance to segregate the two rates by
monitoring the observed rate as a function of various temperature and viscosities. Based on
equation (7) we assume that

versus viscosity at a constant temperature exhibits a linear
14

relationship. Hence by plotting
the reaction limited rate

versus

and fitting to a straight line at each temperature,

=1/intercept and diffusion limited rate

=1/ ( slope). The linear

versus viscosity relationship may break down in some cases, and power law dependence
of the form
long loops

needs to be applied [22], but for systems comprising of 5-20 residue
has been estimated to be very close to 1.

Figure 2.4 demonstrates the optical setup for measuring the transient absorption of the
tryptophan triplet states using a pump-probe spectroscopy. Two different laser beams are
involved in this collinear geometry setup. The Nd:YAG pulse laser (Continuum Surelite, pulse
duration: 8 ns, interval time: 10 ms) is employed to pump the tryptophan to the triplet states. The
wavelength of the pulse beam is shifted from 266 nm to 289 nm using a Raman cell (filled with
methane at 250 psi) to minimize photodamage [23]. Between each pulse, a continuous wave
(CW) 441 nm diode laser beam (add) is used to monitor the tryptophan triplet states. It is split
into two parts: a reference beam and a sample probe beam. The probe beam is collimated with
respect to the pulse beam so that it passes through the sample that contains the tryptophan
molecules that have been excited by the pulse beam. The intensities of the probe beam and
reference beam are measured using two silicon photodiodes (New-Focus) and LeCroy DA
1886A differential and SR445A, SRS Inc four channel preamplifier with 5 x gain at each stage
(which adds two extra stages of amplification). It is then recorded with digital oscilloscopes
(Tektronix TDS), subtracted and then stored in a computer using GPIB interface. The absorption
decay trace reflects the lifetime of tryptophan triplet states which can range from nanosecond to
millisecond. Two oscilloscopes are used to cover the time range; one with a 10µs window and
one with a 10ms window. To eliminate any pulse leakage and high frequency cable noise, a

15

background with the pump but without the probe beam is recorded each time at the beginning of
each experiment and thereafter subtracted from all other recorded decay traces.
The UV pulse also tends to generate hydrated electrons and neutral radicals that absorb
light near 450 nm and decay within 3 µs [19]. Furthermore, O2 is a good quencher of the
tryptophan triplet state. Therefore, before each experiment, the free electrons are scavenged and
oxygen molecules are removed by degassing the sample solution with N2O, a free electron
scavenger, for atleast one hour to eliminate unwanted contribution from such photo-effects.
Another undesired photo-effect is thermal lensing, in which the heat induced by pulse laser
changes the refractive index of the solution, causing the probe beam to slightly refract and move
across the detector. Thermal lensing may result in a pseudo-decay trace on the millisecond
timescale. Ways to diminish the effect of thermal lensing include decreasing the power of the
pulse laser, realigning optics of the pulse and probe beams and adjusting detector positions.
The sample cuvette is placed in a Peltier temperature controlled sample holder (Quantum
Northwest) for the purpose of measuring sample at different temperatures. N-acetyl-Ltryptophanamide (NATA), which is an uncharged analogue of tryptophan, is used to optimize the
alignment of the instrument. NATA has a 40 µs decay time in the presence of deionized water, in
the absence of any additional quencher [20].

2.2 Data Analysis Using SSS theory
The polymer theory of Szabo, Schulten and Schulten (SSS) is the model used to explain
the experimental results. SSS theory models intrachain diffusion dynamics as a motion on a 1dimensional potential of mean force which depends on the probability distribution of intrachain
distances [24].

16

Where

is the Boltzmann constant and

is the equilibrium Trp/Cys distance distribution,

the observed rate of bringing two ends of the chain in close contact can be expressed as [25]:-

Where

is the closest contact distance,

is the contour length of the loop, D is the

intramolecular diffusion constant. The reaction-limited and diffusion-limited rate is given by the
following equations:

It is much easier to numerically solve the above equations instead of analytically to obtain the
diffusion coefficient D. In order to do that, the only unknown in the equation 10 and 11 is the
Trp/Cys distance distribution

.

Figure 2.2 A typical Gaussian polymer chain made of N rigid rods connected together by flexible pivot
points.

17

The simplest distribution that can be chosen is the freely jointed/Gaussian chain as
depicted in Figure 2.2. The pivot points connecting the rigid rods are absolutely flexible, and
have no orientational correlation whatsoever, and all configurations are possible. The typical
is of the following form:-

Substituting equation 12 in equations 10 and 11 yields the following:-

Equations 13 and 14 lead to two very important results that tell us that the reaction-limited rate is
inversely proportional to the chain volume and also the diffusion limited rate is directly
proportional to diffusion coefficient D and inversely proportional to the chain volume.

Though a Gaussian chain is a very simple model, these qualitative relationships between rates
and chain properties hold for even the sophisticated models of intramolecular diffusion studies.
A more sophisticated model that has been widely used for data analysis is the worm like
chain (WLC) model with excluded volume illustrated in Figure 2.3. It is more realistic as it
18

imposes two important constraints on the simplified Gaussian chain model. Firstly, the real chain
stiffness due to chemical bonds is taken into account by introduction of a new parameter called
the persistence length

which is defined as the distance over which the polymer begins to lose

memory of direction (as opposed to a Gaussian chain which has no orientational correlation) and
ensures that the chain does not fold into itself over short distances. In the WLC model the
peptide backbone is assumed to have a uniform rigidity

with persistence length

.

The second constraint imposed is the excluded volume effect by introduction of a hard sphere
shell of diameter

to the end of each virtual peptide bond. Excluded volume of one molecule is

defined as the volume that is inaccessible to other molecules due to the presence of the first
molecule. A criteria to prevent clashes has been imposed that requires that the distance between
two non-neighboring spheres must exceed

.

Worm like chains are computationally generated using the method described previously
[25][26]. Typically 2-10 million chains with a persistence length of 4 Å and an excluded volume of 4
Å are generated and the position of every 10th link (corresponding to Cα) is calculated. A

histogram of distances between two points within the sequence that represent the location of Trp
and Cys distances is normalized to give probability distribution

where

where

is the contour length of the chain. Chains are grown from the N-terminus by vectorially adding
links to the chain that are 1/10th the length of a peptide bond (10 segments per peptide bond) with
a random azimuthal angle

between 0 and 2

radians and an axial angle . A Monte Carlo

algorithm is applied in the construction of each chain, and every time a clash is detected (less
than

distance between 2 non-neighboring residues) after the addition of 10 links, the chain is

truncated 3 persistence lengths before the clash and the chain is regenerated. Otherwise
additional links are added until the full length of the protein is reached.
19

Figure 2.3 A schematic representation of worm like chain model showing the constraint
parameters parameter length
and excluded volume diameter .

There is considerable evidence that highly denatured states act globally like random, noninteracting chains [27] and a simple WLC model with SSS theory should be sufficient to
describe these states. However experimental data suggest that under physiological conditions
(mimicked by low denaturant concentrations) the polypeptide chain is much more compact (as
evidenced by higher values of
lowest possible values of

). A compact conformation can be physically reproduced by the

, but it was found in some sequences that even

compacting the chain enough to yield experimentally observed values of

=0 was not

. This problem could

occur due to of our initial assumption that denatured states can be described by random coil
structures, and we were generating absolutely random worm-like chains in that regard. A
growing number of studies of unfolded proteins that have shown residual structures under
folding conditions such as low denaturants, contradicts our initial assumption [28]. These
20

observations suggest more complicated unfolded states with residual structures marked by
transient intrachain interactions and secondary structure formation which may bias the unfolded
states towards the native conformation. This leads to the conclusion that a homogenous worm
like chain model is probably too simple a model to describe unfolded proteins under
physiological conditions. MD simulations have been able to map long range interactions in
sequences [29] but since our experimental method is much more sensitive to the short range
interactions (tail of the distribution) we require much more than 10000 conformations that were
generated and used by previous methods.

2.3 Boltzmann Energy Reweighting
To account for the above details, the WLC model was modified to reweight and bias the
worm like chains generated, based on favorable intramolecular interactions. The worm like
chains are generated the same way as before but then a bias is assigned to those chains which
have residues of similar hydrophobocity near each other. This upgraded model is called the
Energy-Reweighted WLC model, and has been able to reproduce experimental contact
quenching rates measured under folding conditions in a computationally inexpensive manner
[28].
To perform re-weighting, a new parameter
of pair-wise interactions between Cα atoms:-

21

is defined as a total energy that is a sum

Where i and j are the indices of two non-adjacent residues
two residues which must be larger than
hydrophobic associations [28];

and

is the distance between these

and less than 6.5Å, which is the cutoff distance for
are the normalized hydrophobicities of the two residues

as determined by a standard hydrophobic scale.

is the strength of the hydrophobic potential

between two residues which equal zero if one is hydrophobic and one hydrophilic, and equals ,
if both are hydrophobic or hydrophilic.

is a tunable parameter that exhibits the strength of the

hydrophobic interaction and depends upon various factors, such as temperature and solvent
conditions, etc. The cutoff 0.3 is chosen based on the spread of normalized hydrophobicity
distribution of different amino acids.
Using
distribution

and

as independent reaction coordinates, we can define a 2-D probability

such that:-

And the 1-d Trp-Cys distance distribution is defined as
Let

represent the energy re-weighted distribution of distance

22

.
between the Trp and Cys.

N is the normalization constant.
Substituting

for

in the SSS theory equations 10 and 11 we obtain the following:-

When this model was being developed, different forms of potentials were used to
calculate hydrophobicity energy, and total energy terms with varied forms of r dependence
(namely

distance dependence, Leonard-Jones distance dependence, no distance dependence,

etc) were tested and it was found that the overall trend remained unaltered in different sequences
at different denaturant concentrations, though the

values differed from one model to another.

However if the energy function was not cut off at 6.5Å, the convergence of

between sequences

was seen to fall apart. This suggested that inclusion of long range interactions may introduce
sequence-dependant complexity to the re-weighting functions. Also, another possibility was
considered where only the hydrophobic residues lying near each other were given the bias
(assuming that hydrophobic residues drive the collapse) but the convergence of
among the tested sequences.

23

was not found

2.4 The Charge Effect
In order to account for the charge effects, the total energy term needs to include the electrostatic
energy term in addition to hydrophobic energy. The total energy term becomes:-

Where

and

distance between them.

are the charges of two non-adjacent residues and

is the free space

is the tunable parameter, and can be varied usually varied by holding

constant or vice versa, to predict the measured quenching rates.

24

Raman

Sample

Cell

Diode laser

Nd:YAG
266 nm

cw 442 nm

UV
Holder
filter

289 nm

Oscilloscopes
Silicon Diode

Figure 2.4 Experimental setup for measuring contact quenching

25

Chapter 3
Protein Aggregation in Alzheimer’s disease
In this chapter, I will talk about Aβ peptide aggregation in connection with Alzheimer’s disease
and then compare the early aggregation propensity of Aβ peptide based on its length, pH and
addition of an aggregation inhibitor. I will show that the aggregation model described in chapter
1 is commensurate with the data.
3.1 Aβ Production
APP (Amyloid beta precursor protein) is a transmembrane protein of approximately
120kDa ubiquitously expressed in mammalian cells. It may be attached to neuronal cell
membranes with a transmembrane segment and may exist in different isoforms ranging from 695
to 770 residues, of which the 695, 751 and 770-residue isoforms predominate in the brain
[30][31]. The Aβ region of APP is located partially within the ectodomain and partially within
the transmembrane domain of APP. Though the biological function of APP is not entirely
known, there is some evidence suggesting the involvement of APP in G-coupled cell signaling
and being a regulator protein of cell-trafficking[32][33]. APP has selective metal binding sites in
the N-terminal region for copper, and there have been studies showing that lowering the copper
levels, lowered the gene expression of APP and also depletion of metals in drinking water gave
lower Aβ fibril formation among rabbits in vivo [34]. This has led to suggestions that APP may
be involved in regulation of copper levels and also in maintaining metal homeostasis.
The proteolytic cleavage of APP can follow either a non-amyloidogenic or an
amyloidogenic pathway. In the non-amyloidogenic pathway, APP is cleaved by α-secretase at a
position 83 amino acids from the C-terminus, producing a large N-terminal ectodomain (sAPPα)
26

which may be secreted into the extracellular medium. The remaining C83 fragment is retained in
the membrane and cleaved by ɤ-secretase, producing two benign fragments P3 and AICD
(Amyloid Precursor Protein Intracellular Domain). While during the amyloidogenic pathway, the
intial proteolysis is mediated by β-secretase at a location 99 amino acids from C-terminus. This
cut releases sAPPα into extracellular space, leaving the C99 fragment within the membrane
whose first amino acid corresponds to Aβ. The next step is cleavage of C99 fragment by ɤsecretase at positions anywhere between 38 to 43, which releases different isoforms of the Aβ
peptide[35][36]. Out of the different isoforms the most abundant is the Aβ40 peptide, whereas a
small proportion is Aβ42 isoform (approximately 10%)

3.2 Intracellular Aβ Accumulation
The Aβ peptide was first identified as a component of extracellular plaques in the mid1980s [37] and also there was evidence suggesting the presence of Aβ as a normal product of
APP metabolism throughout life, which could be measured circulating in extracellular fluids like
cerebrospinal fluid and plasma [35]. However, soon studies showing the presence of
intraneuronal Aβ started pouring in, elucidating roles of intracellular abeta in the AD pathology
[36] [38][39]. Cell biology studies have demonstrated the presence of parent protein APP in
plasma membrane, trans-Golgi network, endoplasmic reticulum (ER), endosomal, lysosomal and
mitochondrial membranes. The liberation of Aβ reportedly occurs wherever APP and the β and
ɤ-secretases are localized; the APP cleavage may happen within different cellular compartments
thus giving rise to intracellular Aβ, or it may happen at the plasma membrane or in a secretory
pathway, which would result in Aβ being released extracellularly[36][39]. There have been

27

studies showing that in addition to Aβ being produced intracellularly, Aβ secreted extracellularly
can be taken up by cells and internalized into intracellular pools through receptors and
transporters[40]. Exciting observations from mouse model studies have shown that fluorescently
labeled Aβ injected into tail vein can cross defective blood brain barriers and accumulate
intracellularly in the neurons of the cerebral cortex [41]. This has led to conclusions that neurons
can take up extracellular Aβ thus contributing to the brain Aβ load and leading to pathogenicity
in AD models. In the interstitial fluid of human brain, Aβ concentration is regulated by its rate of
production from parent protein APP, influx into the blood brain barrier via various receptors, and
its rapid clearance by enzymatic degradation [42]. Aβ accumulation in the brain beyond its
critical concentration in the micromolar range favors aggregation [43] and continuous removal of
Aβ species from brain by transport across blood brain barrier and metabolism is essential to
prevent its reported neurotoxic effects[44].

3.3 Aβ Associated with Alzheimer’s disease
Alzheimer’s disease (AD) is a progressive and irreversible brain disorder that is known to
destroy memory and cognitive skills. In 1907, German psychiatrist and neuro-pathologist Alois
Alzheimer first described two striking pathological alterations in the brain of one of his female
patients who died of unusual mental illness and severe dementia. Post mortem studies of her
brain tissues revealed 1) peculiar extracellular deposits in specific brain regions, which is now
referred to as senile plaques (SP’s) and 2) neurofibrillary tangles (NFT’s) occurring
intraneuronally. These two observations turned out to be hallmark features of the disease and
their presence during postmortem examination is still required to confirm AD diagnosis. During

28

mid-1980s biochemists discovered that the SP’s comprised of aggregates of a small peptide Aβ
[37] and that NFT’s are composed of aggregates of the tau protein, which becomes abnormally
phosphorylated [45]. Apart from SP’s and NFT’s, there are also significant other pathological
changes that reportedly occur in the AD brain, including inflammatory responses and oxidative
stress, all of which combined could lead to severe neuronal and synaptic dysfunction and loss.
Extensive research on Aβ peptide has generated considerable evidence connecting it to
the AD pathology.
1) Various missense mutations have been identified in the three genes APP, PS1 and PS2all of which have lead to increased Aβ production and accumulation, thus leading to early
onset familial AD. For example, most of the APP mutations cluster at or near the
cleavage sites within APP, thus favoring proteolytic processing of APP by β and ɤsecretases [46] [35]. Other familial mutations such as Arctic mutation (E22G) reportedly
increase the aggregation propensity of Aβ, thus causing genetically inherited early and
aggressive onset of the disease [47].
2) Direct incubation of neuronal cells with Aβ or overexpression of APP results in cell death
[43] [48][49].
3) Anti-Abeta antibodies have been able to reverse the histological and cognitive
impairments in mice models, that overexpress Aβ or APP [50][51].

3.4 Neurotoxic Effects of Aβ
There are various ways in which Aβ has been reported to exert neurotoxic effects like
oxidative stress, mitochondrial dysfunction, formation of ion-channels and membrane disruption,
29

synaptic dysfunction, activation of cellular processes such as apoptosis and inflammation
[52][53][54][55]. Initially the ‘amyloid cascade’ hypothesis attributed dementia to nerve cell
death caused due to toxicity of large insoluble amyloid fibrils. However research evidence in
recent years suggests that it is actually the lower-order Aβ assemblies that are the neurotoxic
species. To begin with, relatively weak correlations exist between fibrillar plaque density and
severity of dementia among AD patients, whereas correlations between soluble Aβ levels and the
extent of synaptic loss and cognitive impairment are reportedly stronger[56] [57]. Formation of
soluble oligomers precedes AD symptoms and initial AD symptoms usually occurs much before
plaque accumulation [53][58]. The concentration of soluble lower order Aβ oligomers are
elevated in the human AD inflicted brain tissues [59] and oligomeric species as small as dimers
and trimers have reportedly exhibited neurotoxic characteristics in mouse models and cell
cultures [60][61][62]. There is also evidence showing that microinjections of as low as 1pM of
cytosolic Aβ42 induced cell death of cultured human neurons, suggesting that intracellular Aβ
may be neurotoxic in nature [63].
3.5 Structure of Aβ Monomer
The Aβ sequence has an amphipathic character, as its N-terminal segment is hydrophilic,
and 12-14 residues in the C-terminal are extremely hydrophobic. In aqueous solution, a random
coil structure of Aβ40 has been observed [64]. The Aβ42 assumes a beta sheet structure rapidly at
physiological pH, in contrast to less hydrophobic specie Aβ40 which preserves the random-coil
conformation over longer periods of time, before assembling into β-sheet-rich oligomers [53].

30

3.6 Dimers and Trimers
SDS-Stable Aβ dimers have been detected both in vitro as well as in vivo [53] and those
isolated from human brain have a hydrophobic core diameter of about 3-4 nm [65]. The
dimerization process begins intracellularly as reported by studies on cells obtained from human
brains [66]. Soluble Aβ dimers extracted directly from the cerebral cortex of human AD inflicted
brains inhibited long term potentiation (LTP), enhanced long term depression (LTD) and reduced
dendritic spine density in normal rodent hippocampus and disrupted memory of a learned
behavior in rats, whereas neither insoluble plaque cores nor Aβ monomers significantly altered
synaptic plasticity [67]. Another study has reported the effect of soluble oligomers secreted by
cultured cells expressing a mutant form of APP that causes aggressive forms of familial AD and
have found strong evidence suggesting that trimers may be selective inhibitors of certain forms
of hippocampus LTP in mice [68]. The above findings support the hypothesis that soluble
oligomers as small as dimers or trimers could be the potential synapto-toxic species in course of
AD pathogenesis.

3.7 Higher Order Aggregates
Different oligomers (globulomers, spheroids, annular) comprising of 3-50 Aβ monomer
subunits have been observed in vivo AD patients, mouse models, Aβ secreting cell cultures as
well as in vitro using AFM and SEC. They exhibit varying degrees of toxicity[53] [69]. An
interesting study has shown that extra-cellularly accumulated soluble Aβ dodecamers of 56kDa
(Aβ*56) isolated from the brains of middle aged mice caused memory deficits in them and even
disrupted memory when administered to young rats, independently of plaques or neuronal loss
[62]. Observations and characterization of annular Aβ oligomers in vitro and cell cultures led to
31

the channel hypothesis, according to which the annular oligomers cause membrane-disruption
and may lead to Calcium influx into cells [53].

That can alter Calcium homeostasis by

increasing levels of cytoplasmic Calcium levels, which in turn alters neuronal excitability and
synaptic plasticity, while also exhibiting neurotoxicity [70][71].
Aβ-Derived Diffusible Ligands (ADDL’s) are oligomers ranging in size from 17 to 42
kDa and may comprise of groups of small soluble oligomers; ADDL levels in CSF of living AD
patients monitored with a nanoparticle-based detection technique, were reportedly much higher
in concentration than non-demented age matched controls [72] thus supporting the hypothesis
that soluble Aβ oligomers exhibit strong correlation with cognitive deficits of AD patients.
ADDL’s have also exhibited neurotoxic effects in vitro [53].
Protofibrils (PFs) are highly polydispersed structures in solution ranging from spheres
that are 8 nm in diameter upto curvilinear structures upto 200 nm in length, mostly seen in vitro
by EM or AFM. They are precursor of amyloid fibrils and bind to dyes such as Congo red and
Thioflavin T suggesting high beta sheet content and could be potentially neurotoxic [53][69].
Their formation is dependent on concentration, pH and ionic strength, and they could either
proceed to form fibrils due to addition of preformed fibrillar seeds or could dissociate into lower
molecular weight Aβ oligomers [69].
Amyloid fibrils found in vivo and in vitro are ‘thermodynamically stable’, highly
insoluble and structurally ordered aggregates, composed of repeating β-sheet units aligned
perpendicularly to the fibre axis [53]. They have been observed to bind well to dyes such as
Congo red and Thioflavin T and also exhibit Fourier transform infrared (FTIR) spectroscopy
peak at around 1620, indicating the presence of beta sheet. They usually have a characteristically
distinctive cross-β pattern that arises from X-ray diffraction studies. The term ‘cross- β’ was

32

based on observation of two sets of diffraction lines, one longitudinal and one transverse,
forming a characteristic ‘cross’ pattern. The two notable reflections are: a meridional at 4.7-4.8
Å that arises from the spacing between hydrogen bonded β-strands and an equatorial around 1011 Å that occurs from the distance between the β-sheets and may vary according to the
respective side-chains present. The stacks of beta sheets are usually short and traverse the
breadth of the amyloid fibrils, while the length is decided by the number of aligned β-strands
[73].
Combination of results from hydrogen-deuterium (HD) exchange, NMR, mutagenesis
and solid state NMR (ssNMR) led to the conclusion that 3D structure of Aβ42 and Aβ40 amyloid
fibrils consist of two parallel β-sheets joined by a β-turn, represented by a sheet-turn-sheet
morphology [74][75][76]. NMR results have shown that Aβ40 fibrils may have the following
key features: Residues 1-10 are disordered, 12-24 form β-strand1, 25-29 initiate β-turn, 30-40
form β-strand2 connected to β-strand1 with help of β-turn. While HD exchange, NMR and
mutagenesis results have shown that Aβ42 fibrils may have the following key features: Residues
1-17 are disordered, 18-26 form β-strand1, 27-30 initiate β-turn, 31-42 form β-strand2 connected
to β-strand1 with help of β-turn.
Each β-sheet may be stabilized as a result of different inter-molecular side chain
interactions, such as hydrophobic interactions, salt bridges, etc. At least two sheets arising from
two Aβ molecules may be required to form a protofilament which can then interact at higher
levels forming mature fibrils. The 3-D structure of Aβ42 protofilament consists of two-stacked,
parallel β-sheets that perpetuate along the fibril axis. The residues L17, F19 and A21 of β-sheet1
may mediate hydrophobic intermolecular contacts with even numbered residues of β-sheet2. The
loop region of residues 27-30 is connected to β-sheet1 by means of intermolecular salt bridge

33

D23-K28, which also forms contacts to residues I32 and L34 of β-sheet2. Cryo-EM analysis
have shown protofilament dimensions of around 25 Å thick and 45 Å long, and two
protofilaments may wind and twist around each other with a hollow core in between.
Amyloid plaques found in AD patients’ and mouse brains represent the final stage of Aβ
aggregation process in vivo. They are large extracellular Aβ deposits composed of insoluble Aβ
amyloid fibrils and could be as big as 50 um in size [53]. Seeding of Aβ has been found to be
highly specific; seeds from other amyloidogenic proteins are not able to trigger Aβ fibrillization
or plaque formation.

3.8 Comparison of Aβ42 and Aβ40
The most abundant alloforms of the peptide co-exist under normal conditions in the brain
in an Aβ42/Aβ40 ratios of 1:9 [77]. Out of the two, the Aβ42 specie have been widely observed to
be much more aggregation prone, both in vitro and in vivo. The critical concentration of fibril
formation is typically measured by determining the concentration at which the rates of fibril
formation and fibril dissolution are equal [43] and it has been observed that critical concentration
of Aβ42 is five times lower than that of Aβ40 under physiological conditions. Lag time of fibril
formation is determined by the time needed to reach to 10% of the maximum fibril growth
(which is marked by maximum rise in ThT fluorescence intensity). It has been reported that the
lag time of Aβ40 is around 2.5 hours [78] whereas lag time of Aβ42 is in the order of minutes [79]
and Aβ40 is much more soluble than Aβ42 in vitro. There have been reports showing that minor
increase in Aβ42/Aβ40 ratios have led to the formation of neurotoxic oligomers, which can
potentially lead to cell death [78] and may also lead to more aggressive forms of the disease
compared to sporadic AD cases, affecting synaptic activity, viability of neuronal cells and

34

memory development in animal models [77]. Aβ42 plays a more important role in AD
pathogenesis, since it aggregates faster and is reportedly more prone to forming neurotoxic
species than Aβ40 [75][77]. The higher aggregation propensity of Aβ42 makes purification,
handling and reproducibility a big challenge and a lot less number of studies have happened on
Aβ42 specie as compared to Aβ40[53].

3.9 Expression and Purification Protocol
Most researchers use chemically synthesized Aβ peptide for experimental studies because
recombinant purification of the peptide has several challenges. However, a recent work showed
that recombinantly produced Aβ42 aggregates much faster and is more neurotoxic than its
chemically synthesized counterpart [80]. That motivated us to purify the peptide recombinantly
for our experiments. Aβ is usually expressed with a fusion tag, which makes subsequent
purification stages much easier. The fusion construct Aβ42 plasmid was a kind gift by Dr. Rudi
Glockshuber, from ETH Zurich. The fusion construct consist mainly of a Histidine tag and a
soluble polypeptide segment comprising 19 repeats of the tetrapeptide sequence NANP which is
a previously reported solubilizing Aβ fusion partner, followed by the TEV protease recognition
sequence ENLYFQ↓ (the arrow indicates the cleavage site). The Aβ42 sequence is at the end of
the fusion tag [81][80].
Two different Trp/Cys pair-containing Aβ42 and Aβ40 mutants were created using the
QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA) and DNA sequencing
was used to ensure the presence of the desired mutations. The sequenced wild type and mutant
plasmids were then extracted from the XL1 blue competent cells and transformed into E. coli
BL21 (DE3) cells and plated out in LB-agar plates containing 100µg/ml Ampicillin.
35

E. coli BL21 (DE3) cells bearing the corresponding the plasmid for expressing the Aβ
fusion construct was first grown in 5ml tube cultures of LB media containing 100µg/ml
Ampicillin, while being shaken at 37°C at 220 rpm for 4 hours. It was then transferred to 1liter
LB media flasks (Ampicillin concentration 100µg/ml) and was shaken overnight at 37°C at 150
rpm. Protein expression was induced after 16 hours by addition of filtered 1mM isopropyl-β-Dthiogalacoside (IPTG) after which the flasks were shaken for further 4 hours in 37°C at 150 rpm.
The cells were then harvested by centrifugation at 4000 rpm for 20 minutes (rotor SLA-1500).
The pellets formed were re-suspended in Aβ lysis buffer (20mM Tris, 6M GnHCl, pH 8.0),
sonicated and then stirred at 4°C for 90 minutes. The cell debris was then removed by
centrifugation (19000 rpm in rotor SS-34) at 4°C for 1 hour.
Using the supernatant, the His-tag containing fusion construct was purified with an
affinity column. Ni-NTA agarose column (Nickel charged resins by Qiagen) was first
equilibrated with 5 column volumes of Aβ equilibration buffer (20mM Tris, 10mM Immidazole,
6M GnHCl, pH 8.0) after which the supernatant was loaded and incubated with gentle shaking at
4°C for 1 hour to enable binding of the His-tagged fusion protein to the Nickel column.
Following binding, the Nickel resins containing the bound protein were re-suspended in 5
column volumes of wash buffer (20mM Tris, 20mM Immidazole, 6M GnHCl, pH 8.0) and
gently shaken at 4°C for half an hour to get rid of unbound bacterial protein. The washing step
was repeated at least two to three times, until all unbound proteins were washed away and the
supernatants were discarded each time. Finally the Nickel resins containing mostly fusion protein
were re-suspended in 2 column volumes of elution buffer (20mM Tris, 250mM Immidazole, 6M
GnHCl, pH 8.0) and gently shaken at 4°C for 1 hour, following which the supernatant containing

36

eluted fusion protein was collected and stored at -20°C until further use. The elution step can be
repeated until all fusion protein has been recovered.
The next step in the purification is semi-preparative RP-HPLC (Reverse Phase- High
Performance Liquid Chromatography) to get rid of guanidine and other impurities. The fusion
proteins from the previous step was filtered and loaded into an Agilent Zorbax SB300 C8 semipreparative HPLC column in 10% (vol/vol) acetonitrile containing 0.1% trifluoroacetic acid
(TFA) at a flow rate of 1ml/min. The sample injection loop was washed many times at the same
acetonitrile gradient to ensure that all fusion protein gradually bound to the column and all
unbound proteins eluted, before the acetonitrile gradient was increased. Also, the high
hydrophobicity of Aβ requires HPLC to be conducted at 60°C to enable better protein recovery
than room temperature. Once the protein bound to the column, the acetonitrile gradient was
gradually increased along with the flow rate, resulting in the elution of Aβ fusion protein at 40%
acetonitrile and 0.1% TFA, at 60°C and a flow rate of 2 ml/min. The fusion protein was collected
and lyophilized, and then stored at -20°C.
The next step was cleavage of the fusion protein to yield Aβ peptide. The ProTEV Plus
enzyme was purchased from Promega (Catalog# V6101) which is a 48kDa, improved version of
the tobacco etch virus (TEV) protease (more stable in terms of enzymatic activity). It is reported
to be a highly specific proteolytic enzyme that cleaves at the position shown by the arrow:
ENLYFQ↓D thus liberating Aβ whose first residue is D. The lyophilized Aβ fusion protein was
directly dissolved in the Aβ cleavage buffer (10mM Tris, 0.5 mM EDTA, pH 8.0) and the
cleavage reaction was carried out in the presence of 1mM DTT at 4°C with 16 hour incubation.
Approximately 20µg of fusion was cleaved by using 5 units of the enzyme (1µl of 5units/µl
enzyme) following the protocol provided by Promega. The DTT was added to prevent
37

Methionine oxidation/ Cysteine disulfide bond formation at position 35 that has been reported
previously as a potential problem arising during purification[80]. As soon as Aβ was cleaved, it
began aggregating; at the end of the 16 hour incubation period all the aggregated Aβ was
sedimented by centrifugation (rotor F13S14X50Cy, 11000 rpm for 30 minutes) and then resuspended in 70% (w/w) formic acid, filtered and immediately subjected to RP-HPLC. The
peptide was loaded into an Agilent Zorbax SB300 C8 semi-preparative HPLC column in 10%
(v/v) acetonitrile and 0.1% TFA at flow rate of 1ml/min. Following rounds of washing of the
sample loop, the acetonitrile gradient was increased the same way as before. The fusion partner
being more hydrophilic eluted first followed by the Aβ peptide at 40% acetonitrile and 0.1%
TFA, at 60°C and a flow rate of 2 ml/min. It was collected, lyophilized and stored at -20°C. The
presence of the peptide was confirmed by Electrospray Ionization Mass Spectrometry (ESI).
To ensure that Aβ peptide was completely monomeric for every experiment, the
lyophilized peptide was dissolved in 2-10mM NaOH, sonicated for 2 minutes, filtered using 0.2
µm syringe filter and lyophilized for further use[80][82]. Re-suspension of Aβ peptide at pH>10
helps to ensure that the starting condition of Aβ peptide is completely random coil in
conformation which was confirmed by circular dichroism.
3.10 Thioflavin T Assay
The Thioflavin T fluorescence experiment was performed according to the protocol
previously reported [80]. Monomerized Aβ peptide in lyophilized form was re-suspended into
10mM NaOH until a concentration of roughly 100-200uM was achieved as measured by A280
absorbance (wild type ɛ280= 1730 M-1cm-1 ; F4C, F19W and M35C, F19W mutants ɛ280= 6990 M1

cm-1 ). It was further diluted into 10mM NaOH to a concentration around 80uM which were

38

kept on ice and immediately used for Thioflavin T fluorescence experiment. The reaction
mixture for the aggregation experiment was mixed in a 1cm path length quartz cuvette
comprising of the following:1)

100µl: monomerized 80µM Aβ stock solution (diluted into total volume 1000µl, hence
final Aβ peptide concentration 8µM)

2)

900µl reaction mixture: 750µl of 20mM Sodium Phosphate buffer, pH=7.4 (final
=10mM); 100µl of 10mM HCl (final= 1mM); 20µl of 5M NaCl (final=100mM); 20µl
of 2.5 mM Thioflavin T (final=50µM); 10µl of 100mM TCEP (final=1mM).

The mixture was stirred continuously at 37°C and its fluorescence was recorded every 3min for
approximately 30s (excitation: 440nm, emission: 482nm) using a Jobin Yvon Spex Fluorolog-3
spectrofluorometer equipped with a thermo stated cell holder. Just prior to recording each
reading the cuvette was inverted about 5 times for homogenous suspension of fibrils.

3.11 Trp Cys Quenching Measurement
For this experiment, monomerized Aβ peptide in lyophilized form was re-suspended into
10mM NaOH kept in ice, filtered and diluted into 20-25 mM sodium phosphate buffer (pH=7.5),
1 mM TCEP (to prevent disulfide bond formation), and various sucrose concentrations (0, 10, 20
and 30% w/w) in an ice bucket. The final concentration of the mutant proteins was kept fixed at
30µM for each set of temperature and viscosity measurements. The buffer, sucrose and TCEP
solutions were bubbled with N2O for one hour to eliminate oxygen and scavenge solvated
electrons created in the UV laser pulse.

39

3.12 Mutation Sites
Aβ peptide is an intrinsically disordered peptide lacking both Trp and Cys. In order to
measure reconfiguration, two loops were created containing one Trp and Cys each, 15-16
residues apart in sequence. The natural decay lifetime of the excited state of Trp is slower than
40µs, hence the Trp and the Cys must be engineered close enough in sequence to ensure that the
unimolecular quenching dominates (100ns-20µs) which can be attained by placing them less
than 30 residues apart for a random coil. The Trp and Cys positions had to be chosen very
carefully to ensure that it did not alter the aggregation propensity or structure of the original
peptide. The Tango [83] and Zyggregator [84] algorithms were used to pick the mutation
positions. Tango algorithm is an algorithm that predicts protein aggregation based on beta-sheet
formation propensity of a sequence. The Zyggregator method uses the physio-chemical
properties of amino acids to predict how they may influence growth of misfolded assemblies and
also specific regions of sequences that promote aggregation. The 2 loops created were F4C,
F19W (15 residue long) and F19W, M35C (16 residue long) as shown in Figure 3.1. These sites
were finally picked because Phe and Met have similar hydrophobicity compared to Trp and Cys,
and they are not charged. Additionally, Met is similar to Cys as they both contain sulfur atoms.
The wild-type (WT) along with F4C,F19W and F19W,M35C loops were expressed and purified,
after which experiments were run to verify that the mutants were similar to the WT.
Thioflavin T (ThT) is a dye which is widely used to stain and detect amyloid fibrils in
senile plaques of postmortem AD brains[85][86]. The free dye absorbs light of 385 nm and emits
fluorescence at 445 nm. However, as it associates with the fibrillar beta-sheet aggregates, the
bound complex undergoes a 120nm shift in excitation wavelength, absorbing light around 440nm
to give enhanced emission at 482nm. The ThT florescence profile is characterized by an initial
40

lag phase marked by negligible fluorescence intensity, since it cannot bind to monomer or
oligomeric intermediates in course of the aggregation pathway. However, as soon as fibrils form,
the ThT florescence starts rising and finally saturates after all monomers have been converted
into fibrils. Lag time of fiber formation is defined as the time taken to reach one-tenth of the
maximum fluorescence signal. Figure 3.2 shows the ThT fluorescence data for the Aβ42 WT
along with the 2 mutants. The calculated lag time for Aβ42 WT was ~ 1890s, which is in
agreement with previously reported data [79]. The lag time of the F19W, M35C mutant was ~
2040s and the F4C, F19W mutant was ~ 1770s, which is well within 10% of the WT lag time.
Circular Dichroism (CD) spectroscopy is commonly used to probe the secondary
structure of proteins. The differential absorption of left and right circularly polarized light, as it
passes through a protein molecule possessing backbone asymmetry, is measured at different
wavelengths and reported in milli-degrees. α-helical, β-sheet and random coil structures have
previously reported signature CD spectra which can be compared with that of the protein, to
analyze its secondary structure. Monomeric Aβ peptide is intrinsically disordered, exhibiting
random coil like behavior. However as it aggregates, it has been reported to readily assemble
into β-sheet structured aggregates which can be probed by changes in CD spectra.
Figure 3.3 shows the far-UV CD spectra of WT Aβ42, along with the 2 mutants of Aβ40
and Aβ42. The measurements were carried out in an Applied Photophysics Chirascan
spectropolarimeter, using a 1mm path length cell. The time per point was ~2.5s and bandwidth
was fixed at 1 nm. For this experiment, monomerized Aβ peptide in lyophilized form was resuspended into 10mM NaOH, filtered and diluted 20 times into 20-25 mM sodium phosphate
buffer (pH=7.5) kept in ice, to yield a final protein concentration of ~5 µM. The dead-time
between dissolving the peptide and data collection was ~10-15 min. All the mutants and wild
41

type protein yielded random coil like CD spectra, characterized by a minima around 196-198 nm
and absence of any minima in the 210-230 nm region. This helped in ensuring that the structure
of the mutant proteins was similar to the wild-type.

1
10
20
30
40
DAEFRHDSGYEVHHQKLVFFAEDVGSNKGAIIGLMVGGVVIA
↑
↑
↑
C
W
C

ThT fluorescence (A.U.)

Figure 3.1 Sequence of Aβ42 WT with arrows showing the positions of Trp and Cys mutations
for the 2 loops created.

3e+5

2e+5
wildtype
W19C35
C4W19

1e+5

0
0

20

40

60

80

100

120

time (min)
Figure 3.2 Thioflavin T fluorescence spectra of Aβ42 WT along with the 2 Trp/Cys containing
mutants.

42

Ellipticity (mdeg)

0

-200

-400

-600

190

200

210

220

230

A

1-42

wt

A

1-42

F19W M35C

A

1-42

F4C F19W

A

1-40

F19W M35C

A

1-40

F4C F19W

240

250

wavelength (nm)

Figure 3.3 The CD spectra of Aβ42 WT, and the 2 loops containing Trp/Cys, both for Aβ42 and
Aβ40.

43

3.13 Measuring Reconfiguration
Figure 3.4 shows sample decay kinetics of Trp triplet excited state as quenched by Cys,
for a typical Aβ mutant. It is best fit to a three exponential decay (as shown by the red line); a
rapid decay on the microsecond timescale shows that the tryptophan triplet is being rapidly
quenched by cysteine and this can be attributed solely to the reconfiguration of the monomer
population, as the Trp is readily accessible to the Cys. A slower decay in the 10 microsecond
timescale (faster than Trp natural lifetime) is observed that maybe attributed to the
reconfiguration of an oligomeric population which may be gradually forming or imperfect
monomerization. A very slow decay in the 100 s to 1 ms timescale may be due to different
photo-physical processes[19]. However the amplitude of the fastest decay rate corresponding to
the monomer population dominates, and is reported as
plots of exponential decay times (

. Figure 3.5 shows sample raw data

) versus viscosity ( ). Decay times at a single

temperature can be fit to lines. Each experiment was repeated and eight points corresponding to
each temperature were fitted to a line, from which

and

were calculated.

The reaction limited rates shown in Figure 3.6a were calculated from intercept of the
straight line fits from the plots shown in Figure 3.5 (a, b, c, d) such that
measured values of
loops. Since

=1/intercept. The

for Aβ42 loops were an order of magnitude higher than that of the Aβ40

is inversely proportional to the chain volume (equation 15), our measurements

show that Aβ42 is significantly more compact than Aβ40 at all temperatures, as shown in terms of
decreased average Trp-Cys distances, which is plotted in Figure 3.6c. Our measurements show
that the average Trp-Cys distances for both the loops of Aβ42 is smaller than that of Aβ40 as
quantified by average

values, which means the Aβ42 ensemble is much more compact than

Aβ40.
44

Our results are in agreement with previously reported molecular dynamics (MD)
simulation studies. Comparison between intra-molecular contact maps of Aβ42 and Aβ40 has
revealed that the two additional hydrophobic residues at the Aβ42 C-terminus, Ile41 and Ala42
significantly increase contacts within the C-terminus and also between the C-terminus and the
central hydrophobic cluster (CHC region: Leu17-Ala21). Additionally, the Tyr10 in Aβ42 also
contacts residues 13-32 more frequently than Aβ40 [87]. Another reported MD simulation study
has revealed many more close interactions between residues 25-37 and 12-20 for A42 than for
A as shown by contact maps [88], which further supports the hypothesis that stronger Cterminus interactions caused between 39-40 and 41-42 residues of A42 renders the rest of the
residues in the hydrophobic patch (residues 29-39) to form interactions with other residues
distant in sequence. Increased intra-molecular contact frequencies in Aβ42 sequence could be
responsible for the compaction in chain which we have experimentally measured. In another
analysis of MD trajectories in explicit water of Aβ42, non-local backbone H-bonds between the
residues of Lys16Leu17 and Val39Val40Ile41 was observed being formed with time that caused
a decrease in the number of water molecules around those residues, which could also lead to
compaction. Due to such non-local backbone contacts being formed, those regions may become
inaccessible to H bonding with water, facilitating inter-molecular side-chain interactions, which
could initiate aggregation in the Aβ42 sequence [89].
The diffusion limited rates shown in Figure 3.6b were calculated from the slopes of the
straight line fits shown in Figure 3.5(a, b, c, d) such that

=1/ ( slope) and we normalize by

the viscosity of water at that temperature. To find the intramolecular diffusion coefficient D, we
need a probability distribution

, such that Equation 10 predicts the measured

. Once an

appropriate distribution has been chosen for a particular solvent condition, D can be calculated
45

using

and measured

. For Aβ two models were chosen to generate

. Based on our

calculations, at all temperatures the measured D of Aβ42 (D~10-7cm2/s) is lower by an order of
magnitude than that of Aβ40 (D~10-6cm2/s) which has been summarized in Table 3.1. Previous
MD simulations (S2 ordering parameter analysis) and NMR studies [90][91] have found that the
C-terminus of Aβ42 is more structured than that of Aβ40. It has been hypothesized that the
formation of a β-hairpin in the sequence IIGLMVGGVVIA involving short strands at residues
31-34 and 38-41 reduces the C-terminal flexibility of Aβ42 peptide, which may explain the slow
intramolecular diffusion we have measured.
The simplest polymer chain model, a freely jointed chain, yields a Gaussian distribution
as given by Equation 12. Table 3.1 below summarizes the D and
fits shown in Figure 3.5, using the Gaussian distribution as

calculated from straight line

. However a more sophisticated

model is the worm-like chain model, which has been described in detail above. This model uses
as a starting point a large number (20,000,000) of wormlike-chains [28] created by a Monte
Carlo method, setting the persistence length to 4 Å. These chains are created with an excluded
diameter for each link on the chain of 4 Å. The ensemble of worm like chains were then reweighted to favor conformations in which residues of similar hydropathy are near each other,
using a tunable parameter σ (Equation 17). This re-weighted worm like chain probability
distribution was normalized and then used to calculate D and average Trp-Cys distances as
shown in Figures 3.6c, d and 3.7c, d. This model was further improved to take into account the
charge effects [92] and was used to analyze the pH dependant effects of Aβ reconfiguration,
which has been shown in Figures 3.7a, b. As shown by Equation 22 the charge effects introduced
another parameter ɤ in addition to σ. The tuning parameters were adjusted to make the calculated
reaction limited rates match the measured. When comparing measurements at pH7.5 and 10, we
46

initially set σ=0 and adjusted ɤ to make the ratios of the calculated rates at these two pH match
the measured rate ratios, and then adjusted σ to get absolute agreement with the measured rates.
As shown in Table 3.1, when the charge effects were introduced to analyze the data that were
carried out under conditions of similar charge (such as pH 7.5) the calculated D and final average
Trp-Cys distances did not change significantly, hence for conditions where charge effects did not
play a role, we assigned ɤ=0.
Reconfiguration rate can be defined as the rate which allows the disordered chain to
completely reconfigure by diffusion of the individual residues across the diameter of the chain
and is given by:-

where

is given by average radius of the chain. Using our measured diffusion coefficient D and

average Trp-Cys distance from WLC model as

, we obtain the reconfiguration rate of Aβ42 as

4x106s-1 and Aβ40 as 2.2x107s-1 at 40oC. Interestingly, our measurements show that the
reconfiguration Aβ42 is 5 times slower than Aβ40 at all temperatures. This correlates to
approximately 5 times higher concentration required for Aβ40 to aggregate with the same lag
time, and is in excellent agreement with previous reports showing that the critical concentration
of Aβ40 is 5 times higher than Aβ42 [53][78][43][77].
Using Smoluchowski model of diffusion, the bimolecular diffusion rate

for a 50µM

concentration of peptide is 1.7x105s-1. Interestingly for Aβ42, reconfiguration and bimolecular
association formation rates are observed to be within a factor of 10, which further supports the
hypothesis that when reconfiguration slows down to an optimum speed such that the

47

hydrophobic patches get to interact with each other and stick, that’s when aggregation is most
likely to occur[18][17]. As Aβ40 reconfigures 100 times faster than its ability to form bimolecular
association, stable association formation is much more difficult.
This correlation between reconfiguration and aggregation can be rationalized using a
previously reported kinetic model described in chapter 1 [17]. It was solved in Matlab using
reconfiguration rate

, calculated bimolecular association rate,

, and a previously reported

dimerization rate measured using self-quenching of tetramethylrhodamine (TMR) attached to Aβ
peptide, which showed a rapid loss of fluorescence within the lag time of fibril formation, which
could be attributed to dimer and trimer formation. Solving the kinetic model using the previously
reported values of

=1.9s-1 for Aβ42 and

=0.5s-1 for Aβ40, the rate of oligomer formation

with time was computed and plotted, shown in Figure 3.8. The black colored plot represents
Aβ42 and red represents Aβ40, comparing which we find that difference in reconfiguration
between the two explains why the former is more aggregation prone than the latter.
versus viscosity for Aβ42 peptide at pH10.

Figure 3.5e shows the raw data plot of

The trend is very different from the data at physiological pH, as the slope decreases consistently
at all temperatures. That translates into increase in diffusion with temperature, which is the usual
trend that is observed in most proteins [28][26]. However, the reversal of this trend was first seen
in the aggregation prone sequence α-synuclein, where diffusion was seen slowing down with
temperature, and it seemed to be undergoing a reverse glass like transition [18]. Aβ42 peptide
under aggregating conditions was also seen to behave like α-synuclein. Both Aβ42 and αsynuclein are extremely hydrophobic sequences and hydrophobic interactions can dominate at
higher temperatures, thus slowing down diffusion. However, at higher pH, there is an increase in
the net charge of the sequence and also overall charge distribution, which would increase
48

disorder in the sequence and prevent compaction. Our measurements are in agreement with that
and show that size of the chain is increasing with increase in pH. Charge-charge repulsion can
result in decrease of intra-molecular contacts, causing the sequence to change shape faster with
time and thus intramolecular diffusion increases. The calculated reconfiguration at pH 10 is also
faster than at pH7.5, which causes stable bimolecular association much more difficult and hence
makes the protein less aggregation prone.
Figure 3.5f shows the raw data plot of

versus viscosity when inhibitor curcumin

was added to Aβ42 peptide in equal concentrations. Curcumin is a polyphenolic compound,
derived from common Indian spice turmeric, that is known to prevent aggregation of many
proteins, including Aβ peptide [93][94]. Under aggregating conditions, Curcumin was observed
to inhibit Aβ aggregation in vitro by preventing fibril and oligomer formation in a dosedependent manner, and also blocked neurotoxicity. It was also able to cross the blood brain
barrier in a transgenic mouse model, to prevent Aβ deposition in the brain of mice as confirmed
by ELISA assay [94]. Our measurements show that curcumin was able to increase diffusion and
open up Aβ42 peptide chain sequence, thus speeding up reconfiguration and preventing
aggregation. Our results show that curcumin can be a potential drug candidate for treatment of
AD as it is able to shift reconfiguration to a safe regime, where stable bimolecular association
formation is unlikely, and hence it is able to arrest early aggregation. Curcumin has been
reported to interact with the residues 12 and 17-21 [95] which possibly disrupts intra-molecular
interactions of those residues with the hydrophobic patch far in sequence and leads to an increase
in chain size along with diffusivity, just like we have measured.

49

Absorbance (A.U.)

0.05
0.04
0.03
0.02
0.01
0.00
10-7

10-6

10-5

10-4

10-3

10-2

time (s)

Figure 3.4 Decay of Trp triplet states W19C35 at pH 7.5, 0 C as monitored by the decay of
absorption of 445 nm light and captured through oscilloscope.

50

1e-5

1e-5

6e-6

b

8e-6
1/kobs(s-1)

8e-6
1/kobs(s-1)

a

0C
10C
20C
30C
40C

4e-6

6e-6
4e-6
2e-6

2e-6
0

2

4

6

0

8

2

4

6

viscosity

viscosity
2e-5

3e-5

d

c

3e-5
1/kobs(s-1)

2e-5
1/kobs(s-1)

8

1e-5

2e-5
2e-5
1e-5

5e-6

5e-6
0

2

4

6

8

0

2

viscosity

4

6

8

viscosity
5e-6

8e-6

f

e

1/kobs(s-1)

1/kobs(s-1)

4e-6
6e-6
4e-6

3e-6
2e-6

2e-6
1e-6
0

2

4

6

8

0

2

4

6

8

viscosity

viscosity

Figure 3.5 The fast lifetimes of Trp triplet at pH 7.5 for various viscosities and temperatures as
marked. a) Aβ42 M35C, F19W loop, pH 7.5 b) Aβ42 F4C, F19W loop, pH 7.5 c) Aβ40 M35C,
F19W loop, pH 7.5 d) Aβ40 F4C, F19W loop, pH 7.5 e) Aβ42 M35C, F19W loop, pH 10 f) Aβ42
M35C, F19W loop, pH 7.5: Curcumin

51

1e+7

a

b

kD+(s )

1e+6

-1

-1

kR(s )

AM35C
AF4C
AM35C
AF4C

1e+5

1e+6

1e+5
0

10

20

30

40

0

10

T(degC)

30

40

T(degC)

2.8

10

c

2.7

-1

Dx10 (s )

2.6
2.5

d

1

6

<r>nm

20

2.4
2.3

0.1

2.2
2.1

0.01
0

10

20
T(degC)

30

40

0

10

20

30

T(degC)

Figure 3.6 Comparison of intramolecular diffusion of the 2 loops of Aβ40 and Aβ42 (peptide
concentration =30 µM) a) Measured reaction-limited rates, as determined from the intercept of
plots such as Fig. 9 (a, b, c, d) versus temperature b) Measured diffusion-limited rates as
determined from the slope of plots such as Fig. 9 a, b, c, d normalized to the viscosity of water at
each temperature. Errors from the linear fits are shown as error bars in (a) and (b). c) Average
Trp-Cys distance calculated from the reweighted worm-like chain (WLC) model as described
above d) Intramolecular diffusion coefficients determined using the WLC model.

52

40

28

a

b

M35C_pH7.5
M35C_pH10

10

<r2>1/2Ao

Dx106(cm2s-1)

100

1

26

24

0.1
22
0.01
0

10

20

30

40

0

10

T(C)

30

40

T(C)

100

28
0:1(curcumin:peptide)
1:1(curcumin:peptide)

10

<r2>1/2 Ao

Dx106(cm2s-1)

20

1

26

24

0.1
c

22

d

0.01
0

10

20

30

40

0

T(degC)

10

20

30

40

T(degC)

Figure 3.7 Modulation of reconfiguration of Aβ42 by altering solution conditions and adding an
inhibitor a) Measured intramolecular diffusion coefficients at pH 7.5 and pH 10 for Aβ42 M35C,
F19W loop b) Average Trp-Cys distance at pH 7.5 and pH 10 for Aβ42 M35C, F19W loop
determined using the WLC model from measured reaction and diffusion-limited rates obtained
from plots shown in Fig. 9e c) Measured intramolecular diffusion coefficients for Aβ42 M35C,
F19W loop, with and without inhibitor Curcumin (peptide: Curcumin=1:1) d) Average Trp-Cys
distance for Aβ42 M35C, F19W loop, with and without curcumin, determined using the WLC
model from measured reaction and diffusion-limited rates obtained from plots shown in Fig. 9f.

53

Figure 3.8 Formation of oligomer [O] versus time for Aβ42 (black) and Aβ40 (red) which was
calculated using previously reported kinetic model that correlates reconfiguration with oligomer
formation.

54

Table 3.1 Parameters for two different polymeric models that best fit the experimental data
Peptide

T
(C)
0
A1-42 C4W19 pH 7.5
10
20
30
40
0
A1-40 C4W19 pH 7.5
10
20
30
40
0
A1-42 W19C35 pH 7.5
10
20
30
40
0
A1-42 W19C35 pH 10
10
20
30
40
A1-42 W19C35 curcumin 0
10
20
30
40
0
A1-40 W19C35 pH 7.5
10
20
30
40

Gaussian
<r2>1/2 (Å)
34.6
50.0
42
37.4
36.7
84
82
79
77.5
72.1
39.4
39.6
44.0
35.7
32.2
69.3
62
48
47
42
60
58
55
50
46
80
78
69
62
56

Gaussian D
(cm2s-1)
1.2 x 10-7
8.8 x 10-7
8.4 x 10-7
6.5 x 10-7
5.5 x 10-7
3.9 x 10-6
4.8 x 10-6
4.1 x 10-6
3.4 x 10-6
2.5 x 10-6
1.7 x 10-7
2.8 x 10-7
7.4 x 10-7
4.4 x 10-7
3.9 x 10-7
4.5 x 10-6
6 x 10-6
4.6 x 10-6
5 x 10-6
3.7 x 10-6
4.1 x 10-6
3.5 x 10-6
3.3 x 10-6
3 x 10-6
2 x 10-6
3.5 x 10-6
3.3 x 10-6
3.1 x 10-6
1.8 x 10-6
1.3 x 10-6

55

WLC

1.85
0.7
1.1
1.65
1.7
-1.8
-1.7
-1.6
-1.5
-1.0
0.15
0.2
0.02
0.55
0.9
-0.1
-0.02
0.03
0.2
0.35
0.5
0.2
0.3
0.5
0.7
-0.8
-0.5
-0.4
0.1
0.6

WLC

0
0
0
0
0
0
0
0
0
0
56
55
52
52
52
57
55
52
54
55
0
0
0
0
0
0
0
0
0
0

WLC
<r>(Å)
21.0
24.0
22.5
22.0
21.9
27.1
27.0
26.5
26.0
25.5
22.4
22.4
22.9
22.1
22.0
26.3
26.1
25.5
25.1
24.5
25.5
25.4
25.0
24.5
24.1
27.0
27.0
26.0
25.6
25.0

WLC
D(cm2s-1)
2.0 x 10-7
3.9 x 10-7
2.6 x 10-7
2.0 x 10-7
1.9 x 10-7
1.6 x 10-6
2.0 x 10-6
1.9 x 10-6
1.5 x 10-6
1.4 x 10-6
9.0 x 10-8
1.5 x 10-7
2.5 x 10-7
2.5 x 10-7
2.5 x 10-7
1.2 x 10-6
1.3 x 10-6
1.3 x 10-6
1.2 x 10-6
1.7 x 10-6
1.5 x 10-6
1.5 x 10-6
1.1 x 10-6
1.3 x 10-6
1.1 x 10-6
1.3 x 10-6
1.3 x 10-6
2.2 x 10-6
1.8 x 10-6
1.7 x 10-6

Chapter 4

Protein Aggregation in Parkinson’s Disease
In this chapter, I will talk about α-synuculein protein in connection with Parkinson’s disease and
then show that its reconfiguration rates can be controlled by inhibitors or point mutations.
4.1 Introduction
Synucleins belong to a family of closely related presynaptic proteins encoded by three
distinct genes SNCA, SNCB, SNCG found abundantly in the neuronal tissues of vertebrates,
which express the proteins α, β and γ-synuclein respectively [96]. The first synuclein protein was
discovered in the electric organ synapses of the electric ray Torpedo californica, where it was
found in both nerve terminals (presynaptic membrane) and nuclear envelope (hence the name
synuclein) [97]. α-synuclein is a 140-residue protein which has been estimated to account for 1%
of the total protein in soluble cytosolic brain fractions. The protein is expressed widely across
neurons, including the nucleus, and has also been found in the neuritis of mature neurons [98]. It
has been detected in the cytosol, mitochondria, mitochondria-associated endoplasmic reticulum
(ER) membrane (MAM) and vesicles found in cells[99]. Reportedly, it can be secreted from
neuronal cells in small amounts by exocytosis under normal conditions and in the presence of
stressors, may be released in large quantities into the surrounding extracellular space to be taken
up by neighboring cells, by endocytosis [100]. Although it is extensively found in the brain, it
has also been detected in other tissues, including red blood cells in humans [101].
Parkinson’s disease (PD) was first described by Dr. James Parkinson in his essay “An
Essay on the Shaking Palsy” in the year 1817, where he documented the medical cases of six
56

people who suffered from unusual symptoms such as involuntary tremulous motion, bending of
trunk, and diminished muscle power, that worsened with time. Neuropathologically PD is
characterized by progressive loss of neurons in the substantia nigra (midbrain) and other regions
of the brain, and surviving neurons contain cytosolic filamentous inclusions known as Lewy
bodies (LBs) and Lewy neurites (LNs). LBs are large intracytoplasmic proteinacous inclusions,
5-25 µm in diameter, with a dense core of filamentous and granular materials, and LNs
correspond to abnormal neurites that contain filaments (5-10nm in diameter) similar to those
found in LB’s [102]. Abundant LBs and LNs in the cerebral cortex are the neuropathological
hallmarks of PD as well as dementia with Lewy bodies (DLBs) a common late life dementia
clinically similar to Alzheimer’s disease (AD). In general an average person undergoes regular
nigral cell loss with age; but in PD the neuron loss is considerably accelerated. The symptoms of
PD become evident after more than 70% of the dopaminergic neurons have already died.
Sporadic forms of the disease sets in at an mean age of 70 and accounts for about 95% of the
total cases, while multiple familial mutations associated with the disease causes an early-onset of
the disease before the age of 50 and the hereditary forms account for about 4% of the PD cases
[103].
Several evidence have suggested the possible role of α-synuclein in the pathogenesis of
Parkinson’s disease (PD) [96][104].
1) α-synculein is the primary component of LBs and LNs in all PD patients. Antibodies
which recognize the carboxyl-terminal region of α-synuclein, labeled whole filaments
extracted out of such inclusions, indicating that α-synculein is the major component.
Also, α-synculein can be purified and recovered in large amounts from LBs [103].

57

2) Autosomal dominant early-onset PD cases have been identified in people carrying
missense mutations in the α-synuclein gene, corresponding to A30P, E46K and A53T
substitutions in the α-synuclein protein sequence or as a result of over-expression of wild
type α-synuclein due to the gene duplication or triplication [105][106][103].
3) The transgenic mice and Drosophilia expressing human WT α-synuclein and or the
mutants, have led to PD like symptoms like motor deficits and to the formation of
neuronal inclusion bodies that resemble LBs [107][108].
4) Cells transfected with α-synuclein can develop LB-like inclusions under certain
conditions [103].

4.2 Physiological Function
The function of α-synuclein remains controversial and poorly understood [97]. In
transgenic mouse models that either over expressed wildtype human α-synuclein or A53T mutant
in the brain, showed alterations in synaptic plasticity by modulations of signaling pathways.
Such mice were reportedly prone to developing PD like symptoms, due to overexpressed αsynuclein, which caused dampening of its synaptic transmission and reduced movement [109].
However whether those alterations were arising out of neurodegenerative mechanisms linked to
α-synuclein aggregation or due to α-synuclein dysfunctioning as a result of aggregation can be
debated as the evidence found were indirect. Synaptic transmission is involved in ensuring
healthy learning memory and normal amounts of α-synuclein expression was seen to be a key
modulator of song learning in a zebra finch, which led to the conclusion that α-synuclein may be
a critical regulator of neuronal plasticity [110]. Another study showed that mice normally
expressing α-synuclein are viable, fertile, exhibit intact brain architecture (as demonstrated by

58

electro-physiological studies) and concluded that α-synuclein is an essential regulator of
neurotransmission[111]. Such data have led to the notion that α-synuclein may have a role in
synaptic transmission [112] [113]. Cysteine-string protein-α (CSPα) is another synaptic vesicle
protein whose deletion have been linked to neurodegeneration and is believed to be essential for
neuronal survival. Expression of α-synuclein in mice was able to counteract the lethality imposed
by deletion of CSPα, and together with CSPα and other proteins, was able to protect nerve
terminals against injury [114]. In primary hippocampal neuronal cell cultures, suppression of αsynuclein expression led to significant reduction in the size of the pre-synaptic vesicular pool,
which led to the conclusion that α-synuclein may interact with and regulate specific pools of
synaptic vesicles in neurons, thus modulating synaptic functions in the normal brain [115].

4.3 Effects of Oligomers on Neurotoxicity
One of the earliest pieces of evidence that suggested that protein inclusions do not
invariably cause neurodegeneration was the observation of incidental LBs at autopsy of aged
individuals without clinical features of PD or other neurodegenerative diseases [116]. Incidental
LBs without degeneration in asymptomatic individuals were observed in 12% autopsy cases in
another study[117]. Recent studies have revealed higher presence of soluble oligomers of αsynuclein in the CSF and human blood plasma in patients clinically diagnosed with PD, which
further led to the conclusion that they may be involved with the disease [118] [119] .
Much evidence in the field suggests that early soluble oligomers and pre-fibrillar
assemblies may be more neurotoxic and correlated with PD, compared to fibrils found in the
inclusion bodies such as LBs [113][117][120][121]. In one such study, a striking correlation was
found between the presence of prefibrillar aggregates in the cytoplasm of cell culture with Golgi

59

body fragmentation and decline of cell viability[122]. Interestingly, the increase in Golgi
fragmentation occurred much before the appearance of fibrillar inclusions. There was no further
spread of such destructive events and the cells were able to maintain Golgi structure once the
fibrillar inclusion bodies formed. Rather, co-staining of fibrillar inclusions with α-synuclein,
Golgi markers, and the presence of lysosymes and mitochondria near the juxtanuclear inclusion
bodies, led to the conclusion that inclusions might be a consequence of the cell attempting to
sequester protein aggregates and impaired orgallenes from the cytoplasm.
Cells have various quality control mechanisms that ensure that the proteins are folded
properly post translation, before being released; however when misfolded proteins accumulate in
the cell, it triggers unfolded protein response (UPR) stress signaling, which may lead to
increased expression of chaperone proteins that help to properly fold protein, or proteosomic
degradation of such misfolded proteins [13]. The accumulation of toxic protein conformations in
neurodegenerative diseases implies failure of such defense mechanisms. Another study identified
a compound that promotes inclusion body formation that was able to prevent α-synuclein
mediated toxicity [123] and led to the conclusion that fibrillar inclusion body formation may be
protective mechanisms to counteract degeneration.
In another comparison study, α-synuclein mutants that promote oligomer formation led to
an increased loss of nigral TH-positive neurons, compared to those mutants that causes increased
fibril formation, in a murine model [124]. In vivo, following lentiviral injection into the
substantia nigra region of the rat brain, the mutants that causes maximum oligomer formation in
vitro induced increased dopaminergic loss and neuronal cell death. Dissection of the rat brain
showed the increased occurrence of SDS-stable trimers corresponding to those bad mutations,
which correlated very well with cell death. Typically, the decrease in the number of TH-positive

60

cells was at least three times or more higher for those mutations that caused increased relative
trimer formation in vivo. Abnormal calcium homeostasis reportedly plays a crucial role in the
pathogenesis of neurodegenerative disorders, and in another comparison study, the smallest
oligomers (2-6 nm in height as observed by AFM) generated in vitro when injected into human
cell cultures, were able to cause elevation of intracellular calcium levels and also seeded further
aggregation of cytosolic α-synuclein in cells [125].

4.4 Structure
α-synuclein is a polypeptide chain sequence of molecular weight ~14kDa, whose amino
acid sequence can be divided into three regions: residues 1-60 which contain four 11-amino acid
imperfect repeats (that codes for amphipathic helices) with a conserved motif (KTKEGV);
residues 61-95, containing mostly of hydrophobic residues, the highly amyloidogenic NAC
region and two additional repeats; and residues 96-140 representing the highly charged Cterminal region. At physiological pH it has a net charge of -9 (24 negatively charged side chains,
and 15 positively charged side chains).
At neutral pH α-synuclein shows far-UV CD and FTIR spectra typical of an unfolded
polypeptide chain [126]. The CD spectrum is characterized by a minimum around 196 nm and
absence of bands in the 210-230 nm region, and the FTIR spectrum shows a broad peak at 1650
cm-1, suggesting that majority of molecules are disordered. The hydrodynamic properties of αsynuclein analyzed by size exclusion chromatography and small angle X ray scattering (SAXS)
were all in agreement with the CD and FTIR studies. Surprisingly, the Stokes radius measured
for α-synuclein by size exclusion chromatography was much lower than that of a completely
unfolded peptide chain of the same molecular mass, which led to the conclusion that α-synuclein

61

is slightly more compact than a typical random coil [96]. The radius of gyration (Rg) for αsynuclein calculated at neutral pH from SAXS data (around 40Å) was also significantly smaller
than that of a random coil polypeptide of the same length (52 Å) [126]. NMR studies
subsequently demonstrated the presence of transient long-range contacts within the protein, that
reportedly has a role to play in inhibiting aggregation of the protein [127] and disruption of such
contacts may lead to a more disordered conformation which could initiate aggregation.
Thus α-synuclein is considered to belong to the family of intrinsically disordered
proteins, characterized by a unique combination of low hydrophobicity and high net charge
[128]. Any changes in local environment leading to increase in hydrophobicity and decrease in
net charge can lead to partial folding of the intrinsically disordered protein structure leading to a
pre-molten globule like conformation [103]. α-synuclein has a negative charge at neutral pH (pI
4.7) and at lower pH there would be a minimization of the high net charge of the protein leading
to decreased intramolecular charge-charge repulsion and compaction. Similarly there is an
increased hydrophobic effect at higher temperatures, leading to a stronger hydrophobic driving
force for compaction. The pH and temperature induced structural transitions occurred
simultaneously in a cooperative manner, when monitored by ANS fluroscence and CD [103].
Whether α-synuclein is natively unfolded has been a subject of much debate. Until very
recently α-synuclein was known to exist as a random coil with some residual structures, which
acquires α-helical secondary structure only upon binding to lipid vesicles[129]. However one
recent study showed that α-synuclein isolated and analyzed under non-denaturing conditions
from mammalian cells and red blood cells exist predominantly as a tetramer of molecular weight
~ 58kDa that is rich in α-helical structure [130]. This study also showed that cell-derived αsynuclein behaved very differently from the in vitro recombinantly expressed monomers. It

62

showed significantly increased affinity for lipid binding and also the aggregation behavior was
vastly different, with the folded tetramers showing almost no fibril formation in vitro, as
confirmed by ThT fluorescence studies. These observations led to the important question that
whether the less aggregation-prone folded tetramer conformation was the native form of αsynuclein and if destabilization of such conformation into an unfolded disordered state precedes
its misfolding and aggregation?
This study attributed the differences in conformations to the long established recombinant
α-synuclein purification protocol which uses denaturing conditions like heating or denaturants.
A subsequent study that used modified bacterial expression protocol using a fusion construct,
avoiding heating or denaturants was able to reproduce the helical folded tetramer structure [131].
However whether or not this putative tetramer is the main physiological form of α-synuclein in
the brain is still controversial. Almost all studies conducted later could not replicate this folded
tetramer conformation of α-synuclein and confirmed that endogenous and overexpressed αsynuclein in human, mouse and rat brain predominantly exist in a monomeric state, and comigrates with the unfolded recombinantly derived unfolded monomeric α-synuclein in native
gels and gel filtration columns [132]. Similar behavior was observed in α-synuclein expressed in
mammalian cell lines, and interestingly even the modified bacterial expression protocol using a
fusion construct in absence of any denaturing conditions yielded the monomeric disordered
conformation. The concept of a protein-chameleon has been proposed in literature, according to
which the structure of α-synuclein depends on its environment and the choice between various
conformations is determined by the peculiarities of the protein's surroundings [133].

63

4.5 Mutations Associated with PD
Three autosomal dominantly inherited forms of PD have been identified, all of which is
caused by mutations in the α-synuclein sequence. α-synuclein contains seven repeat regions
(KTKEGV), within which all the familial mutations have been found. The first mutation
identified among Greek and Italian kindred was A53T which is located between the fourth and
fifth repeats [134]. The second mutation A30P located between second and third repeat was
identified among German kindred [135]. The third mutation E46K was identified in a Spanish
kindred, located within the fourth repeat region [106]. Surprisingly, within the two repeat regions
in the NAC region, no mutation has been detected to date. All the mutants are more aggregation
prone than the wild type (WT), and causes early onset of PD in people carrying them. However,
the aggregation pathways differ significantly. For example the lag time (from ThT fluorescence)
of A53T is much lower than the WT, while the lag time of A30P is higher than WT, which
means fibrillation of A53T>WT>A30P and the consumption of monomers of A53T is also the
fastest [136]. However the process of oligomerization is much faster in the case of A30P
compared with the WT and also the monomers are consumed faster than WT, forming insoluble
aggregates that could be precipitated by ultracentrifugation and analyzed using AFM and EM
[137]. The E46K mutant also increases the aggregation propensity of α-synuclein, but the effect
is less dramatic than that of A53T mutant. The relative differences in aggregation is consistent
with the reported age of onset of PD, with the mean age of onset of PD in people carrying A53T
mutant being 45 years, while that of E46K being 60 years [138]. Therefore the general
aggregation propensity of the mutants (fibrillar and non-fibrillar combined) has been observed to
be much higher than the WT.

64

Surprisingly, like WT the mutant proteins are all largely unstructured in solution [139]
but exhibit different fibrillar morphologies. The WT fibrils form twisted filaments with a
diameter 10-15 nm, while A30P mutants form straight filaments with a diameter 11-16 nm, and
A53T form larger twisted filaments with a diameter of 16-19 nm [140]. The mutants were also
observed to affect the morphology and size distribution of α-synuclein oligomeric intermediates
like protofibrils, as measured by ultracentrigugation and scanning transmission electron
microscopy [104]. The A30P promotes formation of annular, pore-like protofibrils, while the
A53T forms annular and tubular protofibrillar assemblies. The WT forms annular protofibrils,
but only after extended incubation.
Previous studies have reported that native α-synuclein adopts an ensemble of
conformations, stabilized by long range interactions, involving the N(residues 1-60) and C
termini (residues 109-140) which shields the highly amyloidogenic non-Aβ component of
Alzheimer disease amyloid (NAC) residues 61-95 from solvent and hence prevents the native
conformation from aggregating [127]. NMR studies have revealed absence of such long range
contacts in both A30P and A53T mutants, which leaves the hydrophobic NAC region free to
form intermolecular contacts and hence aggregation can be easily initiated [135]. This study
further reveals that the compaction of the N-terminal observed in the WT is reduced in the
mutants, although the C-terminal compaction is comparable to the WT.
Interestingly, all atom replica exchange molecular dynamics simulations (REMD) in
aqueous water, have predicted increased long range intramolecular protein interactions of the Cterminal with the N-terminal and NAC region of E46K mutant, compared to the WT [141].
Based on the reported data, the probability of finding structures with Rg values between 25.5 and
32.5 Å is higher for the E46K compared to WT, while the WT is more probable of populating
65

conformations of Rg values between 35.5 and 55 Å. This suggests that the E46K mutant is more
compact than the WT. NMR paramagnetic relaxation enhancement (PRE) measurements have
also detected increased interaction of the acidic C-terminal tail with NAC and N-terminal
regions. Also, extra interactions between residues 40-50 with both C-terminal and NAC regions
of the E46K mutant protein is observed, which leads to significant compaction, compared to the
WT [142]. These results are in agreement with observation that the E46K mutant elutes later than
WT α-synuclein, and is characterized by higher retention time in the size exclusion
chromatography [143], suggesting that it is more compact. This study further performed PRE
analysis on the A30P and A53T mutants, and found similar long range contacts as observed in
the WT, refuting previous claims that these mutants could be less compact than the WT. The
previous PRE reports have detected only minor decrease of long range contacts compared to
WT, which was not larger than the range of experimental error. Residual dipolar coupling (RDC)
measurements have further confirmed the presence of long range contacts in E46K, as well as in
the other two mutants A30P and A53T.
Thus recent findings that PD linked mutations could be more compact than the WT,
challenges previous conclusions that the long range contacts in α-synuclein has a protective role
to play against protein aggregation [127]. A compact structure does not necessarily shield the
NAC region from initiating intermolecular interactions. Evidence supporting this idea have been
reported recently, in which phosphorylation of α-synuclein at Ser 129 is shown to release the Cterminus from interacting with the N-terminal region as detected by NMR, yet at the same time
significantly inhibiting α-synuclein aggregation [144]. The hydrodynamic radius calculated from
the measured diffusion coefficient of phosphorylated α-synuclein is seen to be much higher than

66

that of WT, suggesting that phosphorylation strongly disrupts the ordering of monomeric αsynuclein, thus making the chain more expanded and less aggregation prone.
There have been numerous attempts to study how the mutants affect the residual structure
and other properties of the WT, but each mutant has been observed to affect it differently. NMR
studies have reported that α-synuclein is largely unfolded in solution, though the N-terminal
exhibits preference for a helical conformation while the C-terminal shows an unfolded structure
[145]. CD analysis has revealed a similar random coil secondary structure for both E46K and the
WT, when free in solution [143] and in the presence of phospholipid vesicles, both E46K and
WT undergo a structural transition to a primarily α-helical state. On the other hand, NMR
studies have revealed that the helical propensity of residues 18-31 found in the WT is absent in
A30P, exhibiting zero average population for α-helical conformation. In contrast, the A53T
mutation leaves the helical propensity of this region intact [146]. Also the A53T exhibits slightly
enhanced local preference for extended β-sheet-like conformations around the site of the
mutation. Atomic force microscopy-based single molecule experiments have revealed that the
primary random-coil structure of α-synuclein shows a tendency towards forming β-like
conformers (that resembles β-sheets) in around 7% of the cases in WT, while this tendency
increases 38% upon A30P mutation [8].
Other physiochemical property analysis has also revealed differences in behavior of the
mutants. For example, the hydration properties of A53T mutant compared to the WT using
NMR, differential scanning calorimetry and MD simulations, have revealed that the mutant
displays somewhat higher level of hydration, suggesting a bias towards more open structures
[140]. MD simulation studies also predict that there is considerable differences in the way each
of these mutants interact with membranes [147] suggesting that each would affect the biological
67

role of α-synuclein differently. Electrostatics alone is also insufficient to explain why the
mutants make the protein more aggregation prone. One of the mutants E46K increases the
positive charge of N-terminal lipid binding domain, which would be expected to enhance
interactions with the negatively charged C-terminal domain, but the same does not hold for the
other two mutants. Therefore the key factor that drives these mutants to be so much more
aggregation prone and pathogenic in nature than the WT and causes early onset of PD is still
unknown.

4.6 Expression and Purification Protocol
The α-synuclein plasmid was a kind gift from Gary Pielak (University of North Caroline,
Chapel Hill, NC). α-synuclein is a 140 residue long polypeptide chain sequence that does not
contain either Trp or Cys. First, two different Trp/Cys pair-containing α-synuclein mutants,
namely Y39W/A69C and F94W/A69C, were created using the QuikChange site-directed
mutagenesis kit (Stratagene). For mutation studies, a third mutation was then introduced in the
Y39W/A69C loop and five such triple mutants were created, namely A30P, A53T, E46K, V74E
and T72P. At each step the mutations were confirmed by DNA sequencing. The mutants were
then expressed in E. coli BL21 (DE3) cells and purified by a procedure described previously
[18]. After the cells were grown and harvested, pellets were resuspended in Lysis buffer
(containing Tris EDTA) after which protease inhibitor, DNAase, RNAase, and Lysozyme were
added to the cell suspension and stirred for 30 minutes. It was then sonicated for 4-5 minutes,
boiled for 20 minutes and centrifuged at 13,000 rpm for 30 minutes. The supernatant was
collected and the protein was precipitated by addition of 0.361 g/mL ammonium sulfate. The

68

pellet was resuspended in buffer A (20 mM Tris, 100mM NaCl, pH=8.5) and chromatographed
by anion exchange on a Q-Sepharose Fast Flow column equilibrated with buffer A, and eluted
with a linear NaCl gradient (100-1,000 mM). α-Synuclein containing fractions were collected
and further purified by gel filtration on a Hi-Prep Sephacryl S200 column equilibrated with
buffer A. The monomeric α-synuclein peak was found to be >95% pure as assessed by SDSPAGE and Coomassie blue staining. The protein concentration was determined from the
absorbance at 280 nm using extinction coefficient of 11,460 M-1 cm-1 and stored for later
experiments at -20C in the presence of 20 mM Tris, 100mM NaCl, pH=7.5 along with 1mM
tris(2-carboxyethyl)phosphine (TCEP) to prevent disulfide bond formation.

4.7 Contact Quenching Experiment
For this experiment, the respective α-synuclein mutant fractions were thawed and diluted
in 25 mM sodium phosphate buffer (pH=7.5), 10 mM TCEP (to prevent disulfide bond
formation), and various sucrose concentrations (0, 10, 20 and 30% w/w). The final concentration
of the mutant proteins was kept fixed at 30µM for each set of temperature and viscosity
measurement. The buffer, sucrose and TCEP solutions were bubbled with N2O for one hour, to
eliminate oxygen and scavenge solvated electrons created in the UV laser pulse. Triplet lifetime
decay kinetics of tryptophan as a result of cysteine quenching, was then measured with an
instrument described in Chapter 2.

4.8 Effect of Good and Bad Mutations on Reconfiguration
The Trp triplet decay trace, as it is being quenched by Cys in α-synuclein, is comprised of
two decays; a rapid decay in the microsecond time scale, which corresponds to the excited state
69

of tryptophan being quenched by cysteine, much faster than the natural decay rate of Trp, in the
absence of any quencher, and a very slow decay in the 100µs to 1ms time-scale due to other
photo effects. The fastest rate corresponds to monomer reconfiguration, in which Trp is readily
quenched by the Cys, and is reported as

. Figure 4.1a represents the sample

versus

viscosity at various temperatures for the Y39W-A69C loop of α-synuclein measured at pH7.5,
where the decay times at each temperature is fit to a straight line. The reaction and diffusion
limited rates can be extracted from the intercept and slope respectively, as described previously.
The trend is drastically different from all sequences measured previously, where the intercept is
seen decreasing with temperature and slope increases dramatically, and the straight lines cross
each other [18]. The analysis from both Gaussian and worm like chain analysis reveals that αsynuclein diffuses quite fast (~10-6cm2/s) at lower temperature and physiological pH. However,
as soon as it is exposed to aggregating conditions such as higher temperature (40°C) and lower
pH [103], the diffusion slows down (~10-8 cm2/s) and the chain compacts significantly, slowing
down reconfiguration. As reported previously, the reaction and diffusion limited rates and the
diffusion coefficients measured for 4 different loops spanning the 140 residue long peptide were
remarkable similar at pH7.5 [18].
Figure 4.1b shows the

versus viscosity plot at pH 7.5 and various temperatures for

the aggregation enhancing mutation A53T, in the Y39W/A69C loop. Compared to the wild-type
Y39W/A69C loop, the slope changes much more drastically, and the intercepts at each
temperature decreases significantly, reaching close to zero. The trend of the plots for all 3
aggregation enhancing familial mutations, namely A53T, A30P and E46K are quite similar.
Figure 4.2a shows the comparison of reaction limited rates of wild type with that of the 3
aggregation enhancing mutations such that

=1/intercept. The measured values of
70

are

slightly higher for the A53T mutant, and highest for the E46K mutant, while the A30P mutant
behaves oppositely, with

values slightly lower than the wild type at all temperatures. Figure

4.2b shows the comparison of measured diffusion limited rates at all temperatures where
=1/( slope) where the slopes were picked from the respective
graphs and

versus viscosity

is the viscosity of water at that temperature. The reaction and diffusion limited

rates were analyzed, using the SSS polymer theory, which has been described previously. The
probability distribution of average Trp Cys distances was chosen such that Equation 10 predicts
the measured

. Figure 4.2(c,d) shows the average distance

between Trp39 and Cys69 and

intramolecular diffusion coefficients D calculated by choosing a simple Gaussian distribution as
the

.
At higher temperatures, wildtype α-synuclein becomes more aggregation prone, with the
rising drastically, and the chain compacting at least 3 fold compared to that at lower

temperatures, as shown in Table 4.1 [18]. Figure 4.2c reveals that at all temperatures the E46K
mutant compacts the chain further. At higher temperatures, even though the wild type αsynuclein has already undergone significant compaction, a further 5Å collapse is seen for the
E46K mutant. The A53T mutant’s effects on the chain volume are less drastic than the E46K as
shown by Figure 4.2c, and it uniformly compacts the chain by 3Å at all temperatures.
Interestingly, the average distances between 39W and 69C is slightly higher for the A30P mutant
compared to the wild type at lower temperatures, and remains unchanged at higher temperatures.
Based on our measurements, we conclude that the aggregation enhancing mutants all affect the
chain volume differently. The most drastic effect is seen in the E46K mutant, and that could be
attributed to the addition of a positive charge in the N-terminal domain, which decreases
repulsion of the negatively charged C-terminal and strengthens attraction, thus compacting the
71

chain. The effect of A53T and A30P mutant maybe less pronounced because alanine has similar
hydrophobicities to both threonine and proline. Since all three mutants make the protein
aggregation prone, we can argue that maybe their effects on chain volume alone cannot
determine whether or not it aggregates.
Figure 4.2d shows the comparison of diffusion coefficients D of all the 3 mutants with
that of the Y39W/A69C loop. Interestingly, all the three familial mutations are seen consistently
slowing down the diffusion of the wild type 39W-69C loop by at least 2 fold at all temperatures,
thus facilitating aggregation. The reconfiguration rates
measured diffusion coefficient D and average Trp-Cys distances

has been calculated using the
and reported in Table 4.1. We

find that the reconfiguration of all the familial mutations is slower than the wildtype, at all
temperatures. Using Smoluchowski model of diffusion, the bimolecular diffusion rate

for a

30µM concentration of α-synuclein is 7x104s-1. At lower temperatures, when α-synuclein is less
aggregation prone, its reconfiguration rate is seen to atleast 50 times higher than the bimolecular
association formation rate. However, under aggregating conditions such as higher temperatures,
the reconfiguration of α-synuclein slows down and is seen approaching the bimolecular diffusion
regime. Since

7.8x105 s-1 for the Y39W/A69C loop at T=40°C the reconfiguration and

bimolecular association rates are within a factor of 10. We find that for all the three mutants the
reconfiguration slows down significantly at T=40°C such that it is within a factor of 5 of the
bimolecular association formation rates (Table 4.1). This allows the bad conformations to spend
more time in the aggregation prone states, thus the likelihood of formation of stable bimolecular
association increases, before they can escape by reconfiguration to some innocuous
conformation. Therefore slowing down of reconfiguration is one common factor that seem to be
driving aggregation in the mutants of α-synuclein.
72

Previously some aggregation reducing mutations of α-synuclein have been reported
[148]. V74E reportedly is a self fibrillization defective mutant, which was able to completely
suppress the fibrillation of the wildtype α-synuclein, as measured by ThT fluorescence. The
T72P mutant interestingly was able to prevent the fibrillation of wildtype, as well as the familial
mutants of α-synuclein, namely A53T, A30P and E46K. Figure 4.1b shows the

versus

viscosity plot at pH 7.5 and various temperatures for the aggregation reducing mutation V74E, in
the Y39W/A69C loop. Compared to the wild-type Y39W/A69C loop, the change in slope with
temperature is much more moderate, and the intercepts also increases. The trend of the plots for
the 2 aggregation enhancing familial mutations, namely V74E and T72P are quite similar. Figure
4.3a shows the comparison of reaction limited rates of wild type with that of the 2 aggregation
reducing mutations such that
temperatures), the drastic increase in

=1/intercept. Under aggregating conditions (higher
that is seen in the wild type is completely suppressed for

both the aggregation reducing mutations and

is just seen rising only slightly with temperature.

Figure 4.3b shows the comparison of measured diffusion limited rates at all temperatures. Using
Gaussian distribution as
distances

in equation 10, the diffusion coefficient D and average Trp Cys

were calculated, which have been shown in Figure 4.3c, d.

The analysis shows that at like higher temperature (T=40°C) the aggregation reducing
mutations V74E and T72P expands the chain at least 2 fold, and the distance between Trp39 and
Cys69 increases by more than 20 Å. Also, the slowing down of diffusion with temperature which
was seen in the wild type is completely prevented; instead compared to the Y39W/A69C loop,
the diffusion at T=40°C increases by almost 50 times for the V74E mutant and 30 times for the
T72P mutant. Proline is known to stiffen the chain and that may be attributed to the relatively
slower diffusion measured for the T72P compared to V74E [19]. This huge increase in diffusion
73

seen in the aggregation reducing mutants, translates into 10 times higher reconfiguration at
T=40°C. Because the reconfiguration rate (~106s-1) is much higher than bimolecular diffusion
rate (~104s-1), the formation of stable bimolecular association formation is unlikely and thus
aggregation is prevented. This result further corroborates our hypothesis that aggregation results
from kinetic competition between reconfiguration and bimolecular association.
Because E46K and V74E mutants showed the biggest differences in chain volume and
both causes change of charge in the polypeptide chain, we generated probability distributions
using a previously reported model that reweights ensembles of random worm like chains to
favor or disfavor conformations based on charge attraction or repulsion [28]. We started with a
set of 10 million worm-like chains generated with a persistence length 4Å and an excluded
volume of 4Å. Each chain was analyzed for non-local pair-wise interactions and assigned charge
energy Ee as shown in equation 22. From these new distributions we created a contact map of
average intramolecular distances between all pairs of residues. Figure 4.11 shows the difference
between the maps for mutants E46K and V74E and the wildtype sequence. E46K shows
significant compaction between residues 20-60 and residues 80-120 and a small expansion
between the two termini of the chain. V74E shows some compaction between residues 20-60
and 100-120, though much less than for E46K and significant expansion between residues 20-80
and the C-terminus as well as expansion between residues 60-80 and 110-130. Thus the charge
model predicts an overall compaction for the E46K and expansion for the V74E mutant. The
extra positive charge induced by Lys46 would encourage increased intramolecular contacts with
the negatively charged residues in C-terminus of α-synuclein, which could lead to compaction,
while in the wildtype Glu46 is likely to repel the C-terminus. Similarly, Glu74 instead of
uncharged Val in the NAC region would strongly disfavor contacts between the NAC region and
74

the negatively charged C-terminus region, which could open up the chain as we have measured.
However, electrostatics alone may be insufficient to explain the chain dynamics, because the
measured chain compaction or expansion clearly varies with temperature, which cannot be
explained by charge effects. That’s where hydrophobic effects come into play, and because
Val74 to Glu74 transition involves replacement of a hydrophobic residue by a polar one, it could
further expand chain by decreasing intramolecular contacts between hydrophobic residues of the
NAC region with those in other parts of the chain.

75

2e-5

0C
10C
15C
20C
30C
40C

1/kobs

2e-5
1e-5
5e-6

a

0
0

2

4
viscosity

6

8

1e-5

1/kobs

8e-6
6e-6
4e-6
b

2e-6
0

2

4

6

8

viscosity

1/kobs

8e-6
6e-6
4e-6
2e-6

c

0

2

4

6

8

viscosity

Figure 4.1 (
) versus viscosity at pH 7.5, at various temperatures as marked, for αsynuclein a) Y39W-A69C loop in the wild type sequence b) A53T mutant c) V74E mutant.
Protein concentration=30 µM.
76

b

kD+ (s-1)

kR(s-1)

a

1e+6

1e+5

1e+6

1e+5
0

10

20

30

40

0

10

90
c
80
70
60
50
40
30
20
10
0

30

40

T(C)
10

D X 106 (cm2s-1)

<r2>1/2

T(C)

20

39W69C
A53T
E46K
A30P

d

1

0.1

0.01
10

20

30

0

40

10

20

30

40

T(C)

T(C)

Figure 4.2 Effects of Aggregation-Enhancing, PD-Causing, Familial Mutations (A53T, E46K,
and A30P). a) Reaction-limited rates. b) Diffusion-limited rates, normalized to the viscosity at
water at that temperature. c) Average root mean square distance between W39 and C69
determined from reaction-limited rates d) Intramolecular diffusion coefficients determined for
each diffusion-limited rate.

77

1e+7

kD+ (s-1)

kR (s-1)

a

1e+6
39W69C
V74E
T72P
1e+5

b

1e+6

1e+5
0

10

20

30

40

0

10

T(C)

20

30

40

T(C)

80

10

c

d

D x106 (cm2 s-1)

70

<r2>1/2

60
50
40
30

1

0.1

20
10

0.01
0

10

20

30

0

40

10

20

30

40

T(C)

T(C)

Figure 4.3 Effects of Aggregation Reducing Mutations (V74E and T72P). a) Reaction-limited
rates. b) Diffusion-limited rates, normalized to the viscosity of water at that temperature. c)
Average root mean square distance between W39 and C69 determined for each reaction-limited
rate d) Intramolecular diffusion coefficients determined for each diffusion-limited rate.

78

4.9 Effect of Inhibitors on Reconfiguration
Several natural inhibitors of α-synuclein aggregation have been reported previously. The
most common class of compounds that have exhibited neuroprotective effects belong to the class
of polyphenolic compounds called flavonoids. They are found abundantly in a wide range of
fruits, vegetables, and beverages like tea, red wine, etc. In humans the daily intake of such
flavonoids can range from 50 to 800 mg depending on diet, and may well exceed that of other
dietary antioxidants like vitamin C, E and carotenoids [149]. Therefore their role in prevention of
neurodegeneration has been a subject of much interest.
One such compound is quercitin derived from different kinds of berries, red and yellow
onions, tea, wine, apples, buckwheat and beans [150]. It has been studied extensively for its
beneficial antioxidant properties. Oxidative stress has been reported to be a major risk factor for
dopamine cell degeneration in PD and reportedly causes increased aggregation of α-synuclein in
vitro [151]. Dopamine which is a principal neurotransmitter in the substantia nigra, generates
free radicals during normal metabolism. Excessive accumulation of reactive oxygen species
(ROS) produced in the brain is considered as an important factor that drives neuronal losses and
dysfunction, leading to neurodegeneration. Natural antioxidants have been widely recognized as
free radical scavenger, which can destroy ROS and prevent the progression of neurodegenerative
diseases. However, the recent studies have shown that the covalent binding of quercitin to
monomeric α-synuclein and not its redox properties is what makes it a strong candidate for fight
against neurodegenration [149]. One such study revealed that pre-incubated, oxidized quercitin
caused stronger inhibition of α-synuclein fibrillation, compared to freshly prepared quercitin.
Therefore the antioxidant activity which is the strongest in the freshly prepared quercitin is not
the main cause for inhibition of α-synuclein aggregation. Similarly, another flavonoid baicalein,
79

derived from Chinese herbal medicine Scutellaria baicalensis, inhibits α-synuclein fibril
formation and disaggregates existing fibrils more efficiently in its oxidized form [152]. Binding
studies have revealed that baicalein binds strongly to α-synuclein with a Kd~500nM, and
covalently modifies the protein molecules, possibly modulating the aggregation pathway.
However the molecular mechanism of how they were able to prevent α-synuclein aggregation is
not known. And the more important question is, are they able to prevent aggregation early
enough? Existing studies have reported their ability to prevent late stage aggregation, but there is
no information if they are able to prevent the smallest aggregates, such as dimers or trimers.
Figure 4.4b shows the

versus viscosity plot at various temperatures when

equimolar concentration of inhibitor baicalein (oxidized) was added to the F94W/A69C loop of
α-synuclein. Quercitin (oxidized) also shows similar trend (raw data not shown). Compared to
the protein in absence of any inhibitor, the change in slope with temperature is much more
moderate. The intercepts also increase and don’t change as much, when inhibitors were added.
Figure 5a shows the comparison of reaction limited rates of protein alone with that of inhibitors
quercitin and baicalein added. Under aggregating conditions, that is higher temperatures, the
drastic increase in

which is normally seen in α-synuclein, is completely suppressed by the

inhibitors, instead the

values becomes much smaller as a result of adding the inhibitors.

Figure 4.5b shows the comparison of measured diffusion limited rates at all temperatures. Using
Gaussian distribution as
distances

in equation 10, the diffusion coefficient D and average Trp Cys

were calculated, which have been shown in Figure 4.5c, d.

The analysis shows that at all temperatures, the inhibitors expand the chain and increase
diffusion of the polypeptide chain. The effect is especially pronounced under aggregation
condition such as high temperature, when α-synuclein enters the dangerous regime of diffusion,
80

and the likelihood of aggregation increases. At T=40°C, addition of inhibitor baicalein increases
diffusion of α-synuclein by almost 40 times and the chain expands more than 2 fold, with a 30Å
expansion. Quercitin on the other hand increases diffusion 30 times at T=40°C and expands the
chain as much as baicalein. The oxidized products of these inhibitors, reduce their antioxidant
activities, but increase their polarity and hydrogen binding capacity [149]. When bound to
monomeric α-synuclein, the inhibitor increases hydrophilicity of the chain, making it more
hydrated and thus expanding the chain. Also increased hydrogen bonding capacity of the
oxidized form may encourage the formation of local hydrogen bonds with α-synuclein, thus
disrupting long range interactions, and making the chain bigger and more flexible. The
reconfiguration of inhibitor baicalein is 3.4x106s-1 and that of quercitin is 2.8x106s-1 at T=40°C.
Since the reconfiguration rate is greater than the bimolecular diffusion rate, bimolecular
association is highly unlikely. Even if the encounter complex forms out of two aggregation
competent conformations, fast reconfiguration would allow one or both the conformations to
reconfigure back to an innocuous conformation, thus preventing stable bimolecular association
formation.

4.10 Molecule that Slows Down Reconfiguration
Figure 4.4c shows the

versus viscosity plot at various temperatures when

polyphenol epigallochatechin gallete (EGCG) was added to the F94W/A69C loop of αsynuclein. Surprisingly, compared to the wild-type, the slope changes much more drastically, and
the intercepts at each temperature decreases significantly, reaching close to zero, like it was
observed in the case of the aggregation enhancing familial mutations. Figure 4.6(a,b) shows the

81

reaction and diffusion limited rates of the polypeptide chain respectively, at different
concentrations of EGCG added. Figure 4.6(c,d) shows the average distances between Cys69 and
Trp94 and the diffusion coefficients respectively, obtained using Gaussian analysis, with and
without the inhibitor. Interestingly, we find that when equimolar concentration of EGCG was
added, the diffusion increases slightly, and the chain volume remains unchanged, especially at
higher temperatures. But at 5-times higher concentration of EGCG, the effect is pronounced and
increases drastically, at all temperatures, compared to when no inhibitor was added. At
T=30°C and 40°C, the intercept of
figure 4.4c). The

versus viscosity plot becomes 0 within error (shown in

values at those temperatures approach infinity and cannot be reasonably

determined. Therefore we report the lower end of the error values as

at those temperatures,

and use that to calculate the chain volume and diffusion constant. At higher concentrations of
EGCG, the chain volume is collapsed and diffusion slows down, which results in decreased
reconfiguration, nearing that of bimolecular diffusion. Therefore our measurements predict that
at equimolar concentrations of EGCG, the polypeptide chain behaves little better than the
wildtype and reconfiguration is slightly faster, but at higher concentrations it slows down
reconfiguration of α-synuclein into the dangerous regime where bimolecular association is more
likely and thus making it more prone to aggregation.
Our measurements are in agreement with previously reported studies [153][154].
Reportedly EGCG prevents fibril formation of α-synuclein at equimolar concentrations, relative
to the protein, as observed by decrease of ThT fluorescence signal to ~10% of that observed
when no EGCG was added. This effect is also caught in our measurements, where we observe
the reconfiguration becoming slightly faster than the protein without the inhibitor. However, this
increase in reconfiguration is nothing drastic, unlike observed in the case of inhibitors quercitin
82

and baicalein, merely 2 times faster than when no inhibitor was added, so it’s still not
convincingly out of the dangerous regime. Previous studies have also reported that at higher
concentrations of EGCG, structured aggregates like fibrils is largely prevented but there is the
sudden appearance of unstructured spherical oligomers, which are caught by EM studies. Our
measurements show that at higher concentrations, the reconfiguration of α-synuclein is slowing
down, causing it to enter the dangerous regime where the chain can easily form stable
bimolecular associations. This may be facilitating the formation of the new oligomers and hence
EGCG is modulating the aggregation pathway from fibrils to oligomers. These studies have
concluded that these oligomers are non-toxic in nature, and hence EGCG may be successful in
combating the harmful effects of α-synuclein aggregation in association with PD. However,
these detected non-toxic oligomers were mostly higher order aggregates, which leave open the
more important question of whether EGCG can prevent the formation of the smaller hard-to detect toxic aggregates like dimers. Our results predict that EGCG, which is slowing down
reconfiguration and encouraging bimolecular associations, may not be an ideal aggregation
inhibitor. It is however possible that it combats α-synuclein induced toxicity in cells, by some
other mechanism and not by virtue of remodeling the aggregation pathway.
Of the polyphenolic inhibitors tested so far quercitin and baicalein have given positive
results, but the problem with using them as inhibitors is that even though binding studies show
that they bind to α-synuclein, it is not known exactly where they bind. However, specific binding
is an important issue when it comes to choosing drug candidates, and that is what led to a very
interesting molecule, called CLR01, also referred to as molecular tweezer (MT). Figure 4.8
shows the artificially generated molecular structure of MT, which displays an exceptionally high
affinity for amino acid lysine (Dissociation constant~200 µM in neutral phosphate buffer) [155].
83

It comprises of an electron-rich cavity with two anionic phosphonate groups, which facilitates
strong binding of lysine by locking the amino acid side chain through the cavity and subsequent
locking by formation of phosphonate-ammonium salt bridge, as predicted using NMR by and
Monte Carlo simulations. Thus the Lys butene group (C4H8) group is strongly threaded into the
cavity, stabilized by electrostatic and Vander Waals interactions. Lysine residues are attractive
targets since they are often involved in stabilizing long-range electrostatic interactions,
disruption of which may bring about changes in aggregation behavior of the amyloidogenic
proteins. Lysine specific MTs have been reported to prevent aggregation of many proteins,
including α-synuclein, but the molecular mechanisms behind such behavior was not known and
also the question of whether it can prevent early aggregation was open.

4.11 Molecular Tweezer
Figure 4.4a shows the

versus viscosity plot at various temperatures when

equimolar concentration of MT was added to the F94W/A69C loop of α-synuclein. It shows that
the change in slope with temperature is almost completely wiped out, and also the intercepts
increase considerably, which yields much smaller

values compared to when no inhibitor was

added. Figure 4.7a shows that at all temperatures the reaction limited rates decrease in a
concentration dependant manner with the increase in MT concentration, and at higher ratios there
is an order of magnitude difference in

values of α-synuclein. Figure 4.7b shows the diffusion

limited rates with and without the MT. Using Gaussian distribution as
diffusion coefficient D and average Trp Cys distances

in equation 10, the

were calculated, which have been shown

in Figure 4.7c, d. The analysis shows that MT increases the diffusion of α-synuclein and also

84

expands the chain volume in a concentration dependant manner, as seen by increase in measured
diffusion coefficient and the average distances between Trp94 and Cys69 positions. At T=30 and
40°C, the chain expands more than 3 fold and increase of diffusion is almost 50 times when
equimolar concentrations or higher MT was added to α-synuclein. That results in a
reconfiguration rate which is 10 times higher than the bimolecular diffusion rate of α-synuclein,
which would make stable bimolecular association formation highly unlikely and thus
aggregation is prevented.
α-synuclein has 15 lysine residues, and CLR01 binding site is located in the region
spanning residues 10-20, most likely at Lys10 and or Lys12 which was determined by mass
spectrometry [92]. There is fairly strong binding observed between α-synuclein and the MT with
KD values ranging from 10-7 to 10-4 M, as determined by binding studies done by monitoring
Trp94 fluorescence with increasing concentrations of MT added to a fixed concentration of
protein (~1µM). Binding of MT to α-synuclein causes blue-shift of Trp94 and also increases the
total fluorescence, as shown in Figure 4.9.

This suggests that the Trp is becoming more

sequestered from surrounding molecules, but the fact that Trp is becoming less solvent exposed
does not necessarily guarantee compaction, as we showed in our measurements of

.

However a still unanswered question is exactly how CLR01, when bound to one or two
Lys in α-synuclein, increases the reconfiguration rate. While bound, CLR01 changes the charge
on Lys from +1 to -1 because each CLR01 molecule has two charged phosphate groups, so the
addition of each CLR01 further decreases the net charge of the protein, which under
physiological conditions is -9. It has been reported previously that there is a possible phase
transition from a disordered globule to a swollen coil as the fraction of positively or negatively
charged residues increases while mean hydropathy remains constant [156] and that the pattern of
85

charges in the sequence affects the size of the chain significantly [157]. The charged Lys
residues located throughout the first 100 residues in the sequence of α-synuclein likely
preferentially associate with the multiple negatively charged residues, mostly located in the Cterminus, thus creating many close associations between distant residues in the α-synuclein
sequence. With CLR01 bound, -1 charged Lys may prefer association with positively charged
Lys or Arg residues nearby, creating close associations among residues located closer to each
other in the sequence. Presumably, these local associations lead to conformations that are less
compact and more diffusive.
Since MT binds to Lys10 and Lys12 causing a charge reversal of +1 to -1, a mimic of
MT binding to α-synuclein can be performed by simple reversal of charge in those two residues
in the sequence, which can be used to study the effects of MT binding to α-synuclein. Therefore,
we reversed the charge on these residues and calculated Ee to compare with unbound αsynuclein, assuming that the hydropathy remains unchanged on CLR01 binding to α-synuclein.
Fig. 4.10 shows the re-weighed probability distributions for α-synuclein alone or with CLR01
bound to Lys10 and Lys12. As each CLR01 molecule binds to one or both residues, interestingly
the distribution shifts to larger r, thus predicting that the chain becomes bigger, which matches
our measured decrease in

. Contact maps that plots average intramolecular distances have

predicted significantly closer contacts between N-terminal and C-terminal regions of α-synuclein
(residues 1-40 and 120-140) than random coil. These close contacts are inhibited upon binding of
the first CLR01 molecule, and decreases further upon binding of the second CLR01 [92].
To achieve quantitative agreement between the model and experiment, the tuning
parameters σ and

were adjusted such that the measured reaction limited rate is correctly

predicted by Equation 20. First,

was determined by adjusting the ratio of the calculated
86

reaction-limited rates for 1 and 2 CLR01 bound to match the measured ratio. Then σ was
adjusted to match the absolute rates for those sequences. The best fit for these parameters is
given by =20 and σ=0.25 and the rates have been plotted in Figure 4.10.
This model suggests that α-synuclein behaves as a random chain restrained by both
repulsive and attractive interactions (excluded volume repulsion, hydrophobic attraction, and
charge repulsion and attraction). Disruption of this balance of interactions, in this case by
reversing the charge on certain residues, substantially changes the ensemble properties and
therefore the dynamics. A more expanded chain is more diffusive and therefore more likely to
avoid bimolecular association and subsequent oligomerization. CLR01 binds to α-synuclein and
prevents its aggregation by perturbing key hydrophobic and electrostatic interactions, resulting in
a concentration dependant increase of reconfiguration rate of the protein.
Our results predict that MT can be a potential drug candidate for fighting
neurodegenerative diseases like PD. Previous reports suggest that MT inhibits α-synuclein
aggregation by inhibiting fibril formation and disaggregation of preformed fibrils. The
compound also completely blocked α-synuclein toxicity in cell lines, when α-synuclein was
either overexpressed or exogenously added. Also, in a zebrafish model MT successfully
combated neurotoxicity induced by α-synuclein, reducing neuronal apoptosis, improving
abnormal phenotype and extending survival [158]. Using radio-labeled version of the compound,
it was found that MT crosses the blood brain barrier and 1-month treatment resulted in the
decrease of protein aggregates in the brain of mice, protection of neurons and reversal of disease
progression, suggesting that this compound may actually be a breakthrough find in the fight
against neurodegenerative diseases [159].

87

5e-5
0C
10C
15C
20C
30C
40C

1/kobs(s-1)

4e-5
3e-5
2e-5
1e-5

a

0

2

4

6

8

viscosity
2e-5

1/kobs(s-1)

2e-5
1e-5
5e-6
b

0

2

4

6

8

viscosity
1e-5

1/kobs(s-1)

8e-6
6e-6
4e-6
2e-6
c

0

2

4

6

8

viscosity

Figure 4.4 (
) versus viscosity at pH 7.5, at various temperatures for inhibitor with αsynuclein Y39W-A69C loop a) Molecular tweezer, CLR01 b) inhibitor Baicalein c) inhibitor
EGCG. Protein concentration=30 µM.
88

107

a

b

-1

kD+(s )

94W-69C
Baicalein
Quercitin

kR(s-1)

107

106

105

106

105
0

10

20

30

40

0

10

T(C)

30

40

T(C)

100

10

c

d

Dx106(cm2s-1)

80
<r2>1/2Ao

20

60
40

1

0.1

20
0.01
0

10

20

30

40

0

T(C)

10

20

30

40

T(C)

Figure 4.5 Effects of Natural Inhibitors on α-synuclein. a) Reaction-limited rates. b) Diffusionlimited rates, normalized to the viscosity of water at that temperature. c) Average root mean
square distance between W94 and C69, determined from the reaction-limited rates d)
Intramolecular diffusion coefficients determined from diffusion-limited rates.

89

Figure 4.6 Effect of EGCG on α-synuclein. a) Reaction-limited rates. b) Diffusion-limited rates,
normalized to the viscosity of water at that temperature. c) Average root mean square distance
between W94 and C69, determined from the reaction-limited rates d) Intramolecular diffusion
coefficients determined from diffusion-limited rates.

90

Figure 4.7 Concentration dependant effect of Molecular Tweezer on α-synuclein. a) Reactionlimited rates. b) Diffusion-limited rates, normalized to the viscosity of water at that temperature.
c) Average root mean square distance between W94 and C69, determined from the reactionlimited rates d) Intramolecular diffusion coefficients determined from diffusion-limited rates.

91

a

Figure 4.8 Molecular Tweezer CLR01 molecule

Figure 4.9 Effect of binding of CLR01 on Trp94 fluorescence of α-synuclein. Protein
concentration was fixed at 1µM and the CLR01 concentration was gradually increased.

92

Z(r)

0.006

0.004

107
Calculated rates,
Measured rates
at 30 C
-1

unbound
CLR01
on K10
1CLR01
on K12
2 CLR01
K10/K12
unweighted

kR (s )

0.008

106

0.002

105

0.000
0

20

40

0.0

60

0.5

1.0

1.5

2.0

# CLR01 bound

r (Angstroms)

Figure 4.10 Calculated effect of CLR01 binding on α-synuclein conformation. a) Re-weighted
probability distributions, Z(r), for = 0, = 20 for α-synuclein alone (blue), α-synuclein with
one CLR01 bound at Lys10 (cyan), one CLR01 bound at Lys12 (magenta), and α-synuclein
bound to two CLR01 molecules in positions 10 and 12 (red). In each case, the bound CLR01 are
approximated by Lys→Glu substitutions. b) Calculated (red triangles) reaction-limited rates
using Eq. 9 and the probability distributions calculated in (a). The measured reaction-limited
rates are plotted as black circles.

Table 4.1 Parameters obtained from Gaussian analysis for aggregation prone mutations
Peptide

Temperature
°C

Y39W/A69C
A53T
E46K
A30P
Y39W/A69C
A53T
E46K
A30P

0
0
0
0
40
40
40
40

Diffusion
Coefficient D
(cm2/s)
2x10-6
9.5x10-7
9.6x10-7
1x10-6
3.2x10-8
1.5x10-8
1.4x10-8
1.8x10-8

Average Trp-Cys
distance
(Å)
73
70
63
79
21
17
16
21

93

Reconfiguration
(s-1)
3.8x106
1.9x106
2.5x106
1.75x106
7.8x105
5x105
5x105
4x105

20

20
40

10

60

20

20

40

10

60
0

80
-10

100
120
140

0

80

-20

100
120
140

20 40 60 80 100 120 140

-10
-20
20 40 60 80 100 120 140

Figure 4.11 Intramolecular contact plots a) E46K and b) V74E. Each plot is the difference
between the mutant and the wildtype sequence. The average distance is calculated from the
reweighted ensembles using Eqs. 6 and 7 for γ= 9.

94

Chapter 5
Conclusion
Protein aggregation is a complex process which has been linked to several
neurodegenerative diseases like Alzheimer’s and Parkinson’s. The big challenge associated with
treating these diseases is that by the time people go to the doctor the aggregation process has
already spread all over their brain, killing a big chunk of the brain cells. It is almost impossible to
control disease progression at this stage. The biochemical changes associated with the disease
happen within the body at least a decade before actual symptoms occur. Therefore in order to
treat these diseases it is really important to catch aggregation in its early stages. That was the
primary motivation behind developing this novel biophysical technique, which can predict the
early aggregation propensity of any sequence. Existing screening methods test inhibitors on their
ability to prevent higher order aggregates, because most assays can detect aggregates only
beyond a certain size. Since our technique probes the first stage in aggregation of proteins, it can
be applied towards identifying inhibitors that prevent bimolecular association of proteins.
Inhibitors that arrest early aggregation can be good drug candidates for neurodegenerative
diseases.
Within the crowded cellular environment, the proteins are always colliding against each
other, but not always do they end up sticking together. This model postulates that the speed at
which they collide is very crucial and controls whether they make stable associations.
Reconfiguration can be imagined as the measure of how fast a sequence is changing its shape
with time and if a sequence changes shape faster than its ability to make bimolecular
associations, it can escape making stable contact with another sequence. This escape by

95

reconfiguration is the salient feature of this model and gives it an edge over all other aggregation
models proposed to date.
Based on my measurements, this model successfully predicts and explains why Aβ42 is
more aggregation prone than Aβ40 peptide. The latter reconfigures much faster, and hence can
easily escape making stable associations while the former ends up making stable intermolecular
contacts more often. That would explain why senile plaques derived from postmortem
Alzheimer’s brain shows the presence of large amounts of Aβ42 aggregates. Reconfiguration also
emerges as a key factor that decides whether a mutation is good or bad. The familial mutant
proteins that cause PD, reconfigure much slower compared to the wildtype, making escape by
reconfiguration difficult and hence are much more likely to aggregate. On the other hand,
mutations that prevent aggregation speed up reconfiguration thus making protein-protein
interactions difficult. I further showed that reconfiguration can be moved around by varying
solution conditions and also with the help of small molecule inhibitors that bind to proteins.
Because reconfiguration controls early steps of aggregation of proteins, small molecule inhibitors
that modulate reconfiguration to a safe regime, may be good therapeutic candidates for fighting
neurodegenerative diseases like AD and PD.
However, this work leaves behind several questions unanswered, which could be the
scope of future work. While this model gives a good insight into the aggregation from the
dynamics point of view, the morphological aspect still remains unclear. Even though my
measurements predict compact or expanded monomeric structures under different conditions, the
exact structure is unknown. There is still the need to develop modern imaging techniques to
observe live aggregation within cells and also to detect lowest order aggregate such as a dimer.
Modern techniques such as super-resolution fluorescence imaging techniques have been
96

successful in capturing images of protein aggregates within the crowded cellular environment,
but it was limited by resolution and could only observe fibrillar structures and higher order
oligomers [160].
In the absence of structural information of lower order aggregates, we have to rely on the
kinetic studies to detect protein aggregation in its early stages. That is where this aggregation
model will be extremely useful. Using this model, I have successfully shown an excellent
correlation between reconfiguration with that of the aggregation propensity of sequences.
Therefore simply by measuring reconfiguration rate of sequences and using the aggregation
model described in chapter 1, it is possible to predict the early aggregation propensity of
sequences. This work further shows that it is possible to control the reconfiguration of sequences
by changing the sequence length, conditions, addition of inhibitors or point mutations. By
making the reconfiguration much higher than the dangerous bimolecular association regime, it is
possible to rescue the protein from aggregation. This can be used as an effective strategy to
screen inhibitors that can act in the beginning of the aggregation pathway, making bimolecular
association difficult. Such inhibitors will be useful in fighting neurodegenerative diseases like
Alzheimer’s and Parkinson’s disease.

97

BIBLIOGRAPHY

98

BIBLIOGRAPHY

[1]

C. B. Anfinsen, “Principles that Govern the Folding of Protein Chains,” Sci. , vol. 181 ,
no. 4096 , pp. 223–230, Jul. 1973.

[2]

T. Mittag and J. D. Forman-Kay, “Atomic-level characterization of disordered protein
ensembles,” Curr. Opin. Struct. Biol., vol. 17, no. 1, pp. 3–14, Feb. 2007.

[3]

S. A. Waldauer, O. Bakajin, T. Ball, Y. Chen, S. J. DeCamp, M. Kopka, M. Jäger, V. R.
Singh, W. J. Wedemeyer, S. Weiss, S. Yao, and L. J. Lapidus, “Ruggedness in the folding
landscape of protein L,” HFSP J., vol. 2, no. 6, pp. 388–395, Dec. 2008.

[4]

S. A. Waldauer, O. Bakajin, and L. J. Lapidus, “Extremely slow intramolecular diffusion
in unfolded protein L,” Proc. Natl. Acad. Sci. , vol. 107 , no. 31 , pp. 13713–13717, Aug.
2010.

[5]

V. A. Voelz, V. R. Singh, W. J. Wedemeyer, L. J. Lapidus, and V. S. Pande, “UnfoldedState Dynamics and Structure of Protein L Characterized by Simulation and Experiment,”
J. Am. Chem. Soc., vol. 132, no. 13, pp. 4702–4709, Mar. 2010.

[6]

F. Chiti and C. M. Dobson, “Protein Misfolding, Functional Amyloid, and Human
Disease,” Annu. Rev. Biochem., vol. 75, no. 1, pp. 333–366, Jun. 2006.

[7]

F. Bemporad and F. Chiti, “Protein Misfolded Oligomers: Experimental Approaches,
Mechanism of Formation, and Structure-Toxicity Relationships,” Chem. Biol., vol. 19, no.
3, pp. 315–327, Mar. 2012.

[8]

M. Sandal, F. Valle, I. Tessari, S. Mammi, E. Bergantino, F. Musiani, M. Brucale, L.
Bubacco, and B. Samorì, “Conformational Equilibria in Monomeric α-Synuclein at the
Single-Molecule Level,” PLoS Biol, vol. 6, no. 1, p. e6, Jan. 2008.

[9]

L. F. Pease III, M. Sorci, S. Guha, D.-H. Tsai, M. R. Zachariah, M. J. Tarlov, and G.
Belfort, “Probing the Nucleus Model for Oligomer Formation during Insulin Amyloid
Fibrillogenesis,” Biophys. J., vol. 99, no. 12, pp. 3979–3985, Jul. 2014.

[10] J. Lee, E. K. Culyba, E. T. Powers, and J. W. Kelly, “Amyloid-β forms fibrils by
nucleated conformational conversion of oligomers,” Nat Chem Biol, vol. 7, no. 9, pp. 602–
609, Sep. 2011.
[11] T. R. Jahn and S. Radford, “Folding versus aggregation: Polypeptide conformations on
competing pathways,” Arch. Biochem. Biophys., 2008.

99

[12] C. M. Oslowski and F. Urano, “Chapter Four - Measuring ER Stress and the Unfolded
Protein Response Using Mammalian Tissue Culture System,” in The Unfolded Protein
Response and Cellular Stress, Part B, vol. Volume 490, P. M. C. B. T.-M. in
Enzymology, Ed. Academic Press, 2011, pp. 71–92.
[13] P. Walter and D. Ron, “The Unfolded Protein Response: From Stress Pathway to
Homeostatic Regulation,” Sci. , vol. 334 , no. 6059 , pp. 1081–1086, Nov. 2011.
[14] V. N. Uversky, J. R. Gillespie, and A. L. Fink, “Why are ‘natively unfolded’ proteins
unstructured under physiologic conditions?,” Proteins Struct. Funct. Bioinforma., vol. 41,
no. 3, pp. 415–427, 2000.
[15] V. N. et al Uversky, “Unfoldomics of human diseases: linking protein intrinsic disorder
with diseases,” 2008.
[16] V. N. Uversky, “Intrinsically disordered proteins from A to Z,” Int. J. Biochem. Cell Biol.,
vol. 43, no. 8, pp. 1090–1103, Aug. 2011.
[17] L. J. Lapidus, “Understanding protein aggregation from the view of monomer dynamics,”
Mol. BioSyst., vol. 9, no. 1, pp. 29–35, 2013.
[18] B. Ahmad, Y. Chen, and L. J. Lapidus, “Aggregation of α-synuclein is kinetically
controlled by intramolecular diffusion,” Proc. Natl. Acad. Sci. , vol. 109 , no. 7 , pp.
2336–2341, Feb. 2012.
[19] L. J. Lapidus, W. A. Eaton, and J. Hofrichter, “Measuring the rate of intramolecular
contact formation in polypeptides,” Proc. Natl. Acad. Sci. , vol. 97 , no. 13 , pp. 7220–
7225, Jun. 2000.
[20] J. Vanderkooi, “Tryptophan Phosphorescence from Proteins at Room Temperature,” in
Topics in Fluorescence Spectroscopy SE - 3, vol. 3, J. Lakowicz, Ed. Springer US, 2002,
pp. 113–136.
[21] L. J. Lapidus, W. A. Eaton, and J. Hofrichter, “Dynamics of Intramolecular Contact
Formation in Polypeptides: Distance Dependence of Quenching Rates in a RoomTemperature Glass,” Phys. Rev. Lett., vol. 87, no. 25, p. 258101, Nov. 2001.
[22] R. R. Cheng, T. Uzawa, K. W. Plaxco, and D. E. Makarov, “The Rate of Intramolecular
Loop Formation in DNA and Polypeptides: The Absence of the Diffusion-Controlled
Limit and Fractional Power-Law Viscosity Dependence,” J. Phys. Chem. B, vol. 113, no.
42, pp. 14026–14034, 2009.
[23] E. Amouyal, A. Bernas, and D. Grand, “ON THE PHOTOIONIZATION ENERGY
THRESHOLD OF TRYPTOPHAN IN AQUEOUS SOLUTIONS,” Photochem.
Photobiol., vol. 29, no. 6, pp. 1071–1077, Jun. 1979.

100

[24] A. Szabo, K. Schulten, and Z. Schulten, “1st Passage Time Approach to Diffusion
Controlled Reactions,” J. Chem. Phys., vol. 72, no. 8, pp. 4350–4357, 1980.
[25] L. J. Lapidus, P. J. Steinbach, W. A. Eaton, A. Szabo, and J. Hofrichter, “Effects of Chain
Stiffness on the Dynamics of Loop Formation in Polypeptides. Appendix: Testing a 1Dimensional Diffusion Model for Peptide Dynamics,” J. Phys. Chem. B, vol. 106, no. 44,
pp. 11628–11640, 2002.
[26] V. R. Singh, M. Kopka, Y. Chen, W. J. Wedemeyer, and L. J. Lapidus, “Dynamic
Similarity of the Unfolded States of Proteins L and G†,‡,” Biochemistry, vol. 46, no. 35,
pp. 10046–10054, 2007.
[27] J. E. Kohn, I. S. Millett, J. Jacob, B. Zagrovic, T. M. Dillon, N. Cingel, R. S. Dothager, S.
Seifert, P. Thiyagarajan, T. R. Sosnick, M. Z. Hasan, V. S. Pande, I. Ruczinski, S.
Doniach, and K. W. Plaxco, “Random-coil behavior and the dimensions of chemically
unfolded proteins,” Proc. Natl. Acad. Sci. United States Am. , vol. 101 , no. 34 , pp.
12491–12496, Aug. 2004.
[28] Y. Chen, W. J. Wedemeyer, and L. J. Lapidus, “A General Polymer Model of Unfolded
Proteins under Folding Conditions,” J. Phys. Chem. B, vol. 114, no. 48, pp. 15969–15975,
2010.
[29] M. M. Dedmon, K. Lindorff-Larsen, J. Christodoulou, M. Vendruscolo, and C. M.
Dobson, “Mapping Long-Range Interactions in α-Synuclein using Spin-Label NMR and
Ensemble Molecular Dynamics Simulations,” J. Am. Chem. Soc., vol. 127, no. 2, pp. 476–
477, 2005.
[30] T. W. Groemer, C. S. Thiel, M. Holt, D. Riedel, Y. Hua, J. Hüve, B. G. Wilhelm, and J.
Klingauf, “Amyloid Precursor Protein Is Trafficked and Secreted via Synaptic Vesicles,”
PLoS One, vol. 6, no. 4, p. e18754, Apr. 2011.
[31] D. J. Selkoe, “The cell biology of β-amyloid precursor protein and presenilin in
Alzheimer’s disease,” Trends Cell Biol., vol. 8, no. 11, pp. 447–453, Apr. 1998.
[32] T. Okamoto, S. Takeda, Y. Murayama, E. Ogata, and I. Nishimoto, “Ligand-dependent G
Protein Coupling Function of Amyloid Transmembrane Precursor,” J. Biol. Chem. , vol.
270 , no. 9 , pp. 4205–4208, Mar. 1995.
[33] S. L. Sabo, A. F. Ikin, J. D. Buxbaum, and P. Greengard, “The Alzheimer Amyloid
Precursor Protein (APP) and Fe65, an APP-Binding Protein, Regulate Cell Movement,” J.
Cell Biol. , vol. 153 , no. 7 , pp. 1403–1414, Jun. 2001.
[34] C. J. Maynard, A. I. Bush, C. L. Masters, R. Cappai, and Q.-X. Li, “Metals and amyloid-β
in Alzheimer’s disease,” Int. J. Exp. Pathol., vol. 86, no. 3, pp. 147–159, 2005.

101

[35] J. Hardy and D. J. Selkoe, “The Amyloid Hypothesis of Alzheimer’s Disease: Progress
and Problems on the Road to Therapeutics,” Sci. , vol. 297 , no. 5580 , pp. 353–356, Jul.
2002.
[36] F. M. LaFerla, K. N. Green, and S. Oddo, “Intracellular amyloid-beta in Alzheimer’s
disease,” Nat Rev Neurosci, vol. 8, pp. 499–509, 2007.
[37] C. L. Masters, G. Simms, N. A. Weinman, G. Multhaup, B. L. McDonald, and K.
Beyreuther, “Amyloid plaque core protein in Alzheimer disease and Down syndrome,”
Proc. Natl Acad. Sci. USA, vol. 82, pp. 4245–4249, 1985.
[38] G. K. Gouras, C. G. Almeida, and R. H. Takahashi, “Intraneuronal Aβ accumulation and
origin of plaques in Alzheimer’s disease,” Neurobiology of aging, vol. 26, no. 9. Elsevier,
pp. 1235–1244, 01-Oct-2005.
[39] M. Li, L. Chen, D. H. S. Lee, L.-C. Yu, and Y. Zhang, “The role of intracellular amyloid β
in Alzheimer’s disease,” Prog. Neurobiol., vol. 83, no. 3, pp. 131–139, Oct. 2007.
[40] H. YAZAWA, Z.-X. YU, TAKEDA, Y. LE, W. GONG, V. J. FERRANS, J. J.
OPPENHEIM, C. C. H. I. H. LI, and J. I. M. WANG, “β Amyloid peptide (Aβ42) is
internalized via the G-protein-coupled receptor FPRL1 and forms fibrillar aggregates in
macrophages ,” FASEB J. , vol. 15 , no. 13 , pp. 2454–2462, Nov. 2001.
[41] P. M. Clifford, S. Zarrabi, G. Siu, K. J. Kinsler, M. C. Kosciuk, V. Venkataraman, M. R.
D’Andrea, S. Dinsmore, and R. G. Nagele, “Aβ peptides can enter the brain through a
defective blood–brain barrier and bind selectively to neurons,” Brain Res., vol. 1142, no.
0, pp. 223–236, Apr. 2007.
[42] R. E. Tanzi, R. D. Moir, and S. L. Wagner, “Clearance of Alzheimer’s Aβ Peptide: The
Many Roads to Perdition,” Neuron, vol. 43, no. 5, pp. 605–608, Sep. 2004.
[43] J. D. Harper and P. T. Lansbury, “MODELS OF AMYLOID SEEDING IN
ALZHEIMER’S DISEASE AND SCRAPIE:Mechanistic Truths and Physiological
Consequences of the Time-Dependent Solubility of Amyloid Proteins,” Annu. Rev.
Biochem., vol. 66, no. 1, pp. 385–407, Jun. 1997.
[44] R. Deane, R. . Bell, A. Sagare, and B. . Zlokovic, “Clearance of amyloid-beta peptide
across the blood–brain barrier: implication for therapies in Alzheimer’s disease,” CNS
Neurol. Disord. Drug Targets, vol. 8, pp. 16–30, 2009.
[45] I. Grundke-Iqbal, K. Iqbal, Y. C. Tung, M. Quinlan, H. M. Wisniewski, and L. I. Binder,
“Abnormal phosphorylation of the microtubule-associated protein tau (tau) in Alzheimer
cytoskeletal pathology,” Proc. Natl. Acad. Sci. , vol. 83 , no. 13 , pp. 4913–4917, Jul.
1986.

102

[46] D. J. Selkoe, “Amyloid β-Protein and the Genetics of Alzheimer’s Disease,” J. Biol.
Chem. , vol. 271 , no. 31 , pp. 18295–18298, Aug. 1996.
[47] C. Haass, C. A. Lemere, A. Capell, M. Citron, P. Seubert, D. Schenk, L. Lannfelt, and D.
J. Selkoe, “The Swedish mutation causes early-onset Alzheimer’s disease by [beta]secretase cleavage within the secretory pathway,” Nat Med, vol. 1, pp. 1291–1296, 1995.
[48] D. J. Selkoe, “The molecular pathology of Alzheimer’s disease,” Neuron, vol. 6, no. 4, pp.
487–498, Apr. 1991.
[49] F. M. LaFerla and S. Oddo, “Alzheimer’s disease: Aβ, tau and synaptic dysfunction,”
Trends Mol. Med., vol. 11, no. 4, pp. 170–176, Apr. 2005.
[50] N. B. Chauhan and G. J. Siegel, “Reversal of amyloid β toxicity in Alzheimer’s disease
model Tg2576 by intraventricular antiamyloid β antibody,” J. Neurosci. Res., vol. 69, no.
1, pp. 10–23, 2002.
[51] W. A. Banks, S. A. Farr, J. E. Morley, K. M. Wolf, V. Geylis, and M. Steinitz, “Antiamyloid beta protein antibody passage across the blood–brain barrier in the SAMP8
mouse model of Alzheimer’s disease: An age-related selective uptake with reversal of
learning impairment,” Exp. Neurol., vol. 206, no. 2, pp. 248–256, Aug. 2007.
[52] T. Nakagawa, H. Zhu, N. Morishima, E. Li, J. Xu, B. A. Yankner, and J. Yuan, “Caspase12 mediates endoplasmic-reticulum-specific apoptosis and cytotoxicity by amyloid[beta],” Nature, vol. 403, no. 6765, pp. 98–103, 2000.
[53] V. H. Finder and R. Glockshuber, “Amyloid-β Aggregation,” Neurodegener. Dis., vol. 4,
no. 1, pp. 13–27, 2007.
[54] S. T. Ferreira and W. L. Klein, “The Aβ oligomer hypothesis for synapse failure and
memory loss in Alzheimer’s disease,” Neurobiol. Learn. Mem., vol. 96, no. 4, pp. 529–
543, Nov. 2011.
[55] D. J. Selkoe, “Alzheimer’s Disease Is a Synaptic Failure,” Sci. , vol. 298 , no. 5594 , pp.
789–791, Oct. 2002.
[56] L.-F. Lue, Y.-M. Kuo, A. E. Roher, L. Brachova, Y. Shen, L. Sue, T. Beach, J. H. Kurth,
R. E. Rydel, and J. Rogers, “Soluble Amyloid β Peptide Concentration as a Predictor of
Synaptic Change in Alzheimer’s Disease,” The American journal of pathology, vol. 155,
no. 3. American Society for Investigative Pathology, pp. 853–862, 01-Sep-1999.
[57] D. M. Walsh, I. Klyubin, J. V. Fadeeva, W. K. Cullen, R. Anwyl, M. S. Wolfe, M. J.
Rowan, and D. J. Selkoe, “Naturally secreted oligomers of amyloid [beta] protein potently
inhibit hippocampal long-term potentiation in vivo,” Nature, vol. 416, no. 6880, pp. 535–
539, 2002.

103

[58] C. A. Lemere, J. K. Blusztajn, H. Yamaguchi, T. Wisniewski, T. C. Saido, and D. J.
Selkoe, “Sequence of Deposition of Heterogeneous Amyloid β-Peptides and APO E in
Down Syndrome: Implications for Initial Events in Amyloid Plaque Formation,”
Neurobiol. Dis., vol. 3, no. 1, pp. 16–32, Feb. 1996.
[59] C. A. McLean, R. A. Cherny, F. W. Fraser, S. J. Fuller, M. J. Smith, K. Vbeyreuther, A. I.
Bush, and C. L. Masters, “Soluble pool of Aβ amyloid as a determinant of severity of
neurodegeneration in Alzheimer’s disease,” Ann. Neurol., vol. 46, no. 6, pp. 860–866,
1999.
[60] G. M. Shankar, B. L. Bloodgood, M. Townsend, D. M. Walsh, D. J. Selkoe, and B. L.
Sabatini, “Natural Oligomers of the Alzheimer Amyloid-β Protein Induce Reversible
Synapse Loss by Modulating an NMDA-Type Glutamate Receptor-Dependent Signaling
Pathway,” J. Neurosci. , vol. 27 , no. 11 , pp. 2866–2875, Mar. 2007.
[61] J. P. Cleary, D. M. Walsh, J. J. Hofmeister, G. M. Shankar, M. A. Kuskowski, D. J.
Selkoe, and K. H. Ashe, “Natural oligomers of the amyloid-[beta] protein specifically
disrupt cognitive function,” Nat Neurosci, vol. 8, no. 1, pp. 79–84, 2005.
[62] S. Lesne, M. T. AU - Koh, L. AU - Kotilinek, R. AU - Kayed, C. G. AU - Glabe, A. AU Yang, M. AU - Gallagher, and K. H. AU - Ashe, “A specific amyloid-[beta] protein
assembly in the brain impairs memory,” Nature, vol. 440, no. 7082, pp. 352–357, 2006.
[63] Y. Zhang, R. McLaughlin, C. Goodyer, and A. LeBlanc, “Selective cytotoxicity of
intracellular amyloid β peptide1–42 through p53 and Bax in cultured primary human
neurons,” J. Cell Biol. , vol. 156 , no. 3 , pp. 519–529, Feb. 2002.
[64] S. Zhang, K. Iwata, M. J. Lachenmann, J. W. Peng, S. Li, E. R. Stimson, Y. -a. Lu, A. M.
Felix, J. E. Maggio, and J. P. Lee, “The Alzheimer’s Peptide Aβ Adopts a Collapsed Coil
Structure in Water,” J. Struct. Biol., vol. 130, no. 2–3, pp. 130–141, Jun. 2000.
[65] A. E. Roher, M. O. Chaney, Y.-M. Kuo, S. D. Webster, W. B. Stine, L. J. Haverkamp, A.
S. Woods, R. J. Cotter, J. M. Tuohy, G. A. Krafft, B. S. Bonnell, and M. R. Emmerling,
“Morphology and Toxicity of Aβ-(1-42) Dimer Derived from Neuritic and Vascular
Amyloid Deposits of Alzheimer’s Disease,” J. Biol. Chem. , vol. 271 , no. 34 , pp. 20631–
20635, Aug. 1996.
[66] D. M. Walsh, B. P. Tseng, R. E. Rydel, M. B. Podlisny, and D. J. Selkoe, “The
Oligomerization of Amyloid β-Protein Begins Intracellularly in Cells Derived from
Human Brain†,” Biochemistry, vol. 39, no. 35, pp. 10831–10839, Aug. 2000.
[67] G. M. Shankar, S. Li, T. H. Mehta, A. Garcia-Munoz, N. E. Shepardson, I. Smith, F. M.
Brett, M. A. Farrell, M. J. Rowan, C. A. Lemere, C. M. Regan, D. M. Walsh, B. L.
Sabatini, and D. J. Selkoe, “Amyloid-[beta] protein dimers isolated directly from
Alzheimer’s brains impair synaptic plasticity and memory,” Nat Med, vol. 14, no. 8, pp.
837–842, 2008.
104

[68] M. Townsend, G. Shankar, T. Mehta, D. Walsh, and D. Selkoe, “Effects of secreted
oligomers of amyloid β-protein on hippocampal synaptic plasticity: a potent role for
trimers,” J Physiol, vol. 477–492, 2006.
[69] G. Shankar and D. Walsh, “Alzheimer’s disease: synaptic dysfunction and Aβ,” Mol.
Neurodegener., vol. 4, no. 1, pp. 1–13, 2009.
[70] M. Kawahara, M. Negishi-Kato, and Y. Sadakane, “Calcium dyshomeostasis and
neurotoxicity of Alzheimer’s β-amyloid protein,” Expert Rev. Neurother., vol. 9, no. 5, pp.
681–693, May 2009.
[71] D. H. Small, R. Gasperini, A. J. Vincent, A. C. Hung, and L. Foa, “The Role of AβInduced Calcium Dysregulation in the Pathogenesis of Alzheimer’s Disease,” J.
Alzheimer’s Dis., vol. 16, no. 2, pp. 225–233, Jan. 2009.
[72] D. G. Georganopoulou, L. Chang, J.-M. Nam, C. S. Thaxton, E. J. Mufson, W. L. Klein,
and C. A. Mirkin, “Nanoparticle-based detection in cerebral spinal fluid of a soluble
pathogenic biomarker for Alzheimer’s disease,” Proc. Natl. Acad. Sci. United States Am. ,
vol. 102 , no. 7 , pp. 2273–2276, Feb. 2005.
[73] K. E. Marshall and L. C. Serpell, “Insights into the Structure of Amyloid Fibrils,” Open
Biol. J., vol. 2, pp. 185–192, 2009.
[74] T. Lührs, C. Ritter, M. Adrian, D. Riek-Loher, B. Bohrmann, H. Döbeli, D. Schubert, and
R. Riek, “3D structure of Alzheimer’s amyloid-β(1–42) fibrils,” Proc. Natl. Acad. Sci.
United States Am. , vol. 102 , no. 48 , pp. 17342–17347, Nov. 2005.
[75] K. Irie, K. Murakami, Y. Masuda, A. Morimoto, H. Ohigashi, R. Ohashi, K. Takegoshi,
M. Nagao, T. Shimizu, and T. Shirasawa, “Structure of β-amyloid fibrils and its relevance
to their neurotoxicity: Implications for the pathogenesis of Alzheimer’s disease,” J. Biosci.
Bioeng., vol. 99, no. 5, pp. 437–447, May 2005.
[76] T. R. Jahn, O. S. Makin, K. L. Morris, K. E. Marshall, P. Tian, P. Sikorski, and L. C.
Serpell, “The Common Architecture of Cross-β Amyloid,” J. Mol. Biol., vol. 395, no. 4,
pp. 717–727, Jan. 2010.
[77] K. Pauwels, T. L. Williams, K. L. Morris, W. Jonckheere, A. Vandersteen, G. Kelly, J.
Schymkowitz, F. Rousseau, A. Pastore, L. C. Serpell, and K. Broersen, “Structural Basis
for Increased Toxicity of Pathological Aβ42:Aβ40 Ratios in Alzheimer Disease,” J. Biol.
Chem. , vol. 287 , no. 8 , pp. 5650–5660, Feb. 2012.
[78] I. Kuperstein, K. Broersen, I. Benilova, J. Rozenski, W. Jonckheere, M. Debulpaep, A.
Vandersteen, I. Segers‐Nolten, K. Van Der Werf, V. Subramaniam, D. Braeken, G.
Callewaert, C. Bartic, R. D&#039;Hooge, I. C. Martins, F. Rousseau, J. Schymkowitz,
and B. De Strooper, “Neurotoxicity of Alzheimer&#039;s disease Aβ peptides is induced

105

by small changes in the Aβ42 to Aβ40 ratio,” EMBO J., vol. 29, no. 19, pp. 3408–3420,
Sep. 2010.
[79] J. C. Stroud, C. Liu, P. K. Teng, and D. Eisenberg, “Toxic fibrillar oligomers of amyloid-β
have cross-β structure,” Proc. Natl. Acad. Sci. , Apr. 2012.
[80] V. H. Finder, I. Vodopivec, R. M. Nitsch, and R. Glockshuber, “The Recombinant
Amyloid-β Peptide Aβ1–42 Aggregates Faster and Is More Neurotoxic than Synthetic
Aβ1–42,” J. Mol. Biol., vol. 396, no. 1, pp. 9–18, Feb. 2010.
[81] R. B. Kapust, J. Tözsér, T. D. Copeland, and D. S. Waugh, “The P1′ specificity of tobacco
etch virus protease,” Biochem. Biophys. Res. Commun., vol. 294, no. 5, pp. 949–955,
2002.
[82] D. B. Teplow, “Preparation of Amyloid β‐Protein for Structural and Functional Studies,”
in Amyloid, Prions, and Other Protein Aggregates, Part C, vol. Volume 413, I. K. and R.
W. B. T.-M. in Enzymology, Ed. Academic Press, 2006, pp. 20–33.
[83] A.-M. Fernandez-Escamilla, F. Rousseau, J. Schymkowitz, and L. Serrano, “Prediction of
sequence-dependent and mutational effects on the aggregation of peptides and proteins,”
Nat Biotech, vol. 22, no. 10, pp. 1302–1306, 2004.
[84] G. G. Tartaglia and M. Vendruscolo, “The Zyggregator method for predicting protein
aggregation propensities,” Chem. Soc. Rev., vol. 37, no. 7, pp. 1395–1401, 2008.
[85] H. Levine, “Thioflavine T interaction with synthetic Alzheimer’s disease β-amyloid
peptides: Detection of amyloid aggregation in solution,” Protein Sci., vol. 2, no. 3, pp.
404–410, 1993.
[86] H. Levine III, “Stopped-Flow Kinetics Reveal Multiple Phases of Thioflavin T Binding to
Alzheimer β(1-40) Amyloid Fibrils,” Arch. Biochem. Biophys., vol. 342, no. 2, pp. 306–
316, Jun. 1997.
[87] M. Yang and D. B. Teplow, “Amyloid β-Protein Monomer Folding: Free-Energy Surfaces
Reveal Alloform-Specific Differences,” J. Mol. Biol., vol. 384, no. 2, pp. 450–464, Dec.
2008.
[88] R. P. Bora and R. Prabhakar, “Translational, rotational and internal dynamics of amyloid
β-peptides (Aβ40 and Aβ42) from molecular dynamics simulations,” J. Chem. Phys., vol.
131, no. 15, p. -, 2009.
[89] C. Lee and S. Ham, “Characterizing amyloid-beta protein misfolding from molecular
dynamics simulations with explicit water,” J. Comput. Chem., vol. 32, no. 2, pp. 349–355,
2011.

106

[90] A. E. Sgourakis, N. G., Yan, Y. L., McCallum, S. A., Wang, C. Y. & Garcia, “The
Alzheimer’s peptides A beta 40 and 42 adopt distinct conformations in water: A combined
MD/NMR study,” J. Mol. Biol., vol. 368, pp. 1448–1457, 2007.
[91] Y. Yan and C. Wang, “Aβ42 is More Rigid than Aβ40 at the C Terminus: Implications for
Aβ Aggregation and Toxicity,” J. Mol. Biol., vol. 364, no. 5, pp. 853–862, Dec. 2006.
[92] S. Acharya, B. M. Safaie, P. Wongkongkathep, M. I. Ivanova, A. Attar, F.-G. Klärner, T.
Schrader, J. A. Loo, G. Bitan, and L. J. Lapidus, “Molecular Basis for Preventing αSynuclein Aggregation by a Molecular Tweezer,” J. Biol. Chem. , Feb. 2014.
[93] B. Ahmad and L. J. Lapidus, “Curcumin Prevents Aggregation in α-Synuclein by
Increasing Reconfiguration Rate,” J. Biol. Chem. , vol. 287 , no. 12 , pp. 9193–9199, Mar.
2012.
[94] F. Yang, G. P. Lim, A. N. Begum, O. J. Ubeda, M. R. Simmons, S. S. Ambegaokar, P. P.
Chen, R. Kayed, C. G. Glabe, S. A. Frautschy, and G. M. Cole, “Curcumin Inhibits
Formation of Amyloid β Oligomers and Fibrils, Binds Plaques, and Reduces Amyloid in
Vivo,” J. Biol. Chem. , vol. 280 , no. 7 , pp. 5892–5901, Feb. 2005.
[95] Y. Masuda, M. Fukuchi, T. Yatagawa, M. Tada, K. Takeda, K. Irie, K. Akagi, Y. Monobe,
T. Imazawa, and K. Takegoshi, “Solid-state NMR analysis of interaction sites of curcumin
and 42-residue amyloid β-protein fibrils,” Bioorg. Med. Chem., vol. 19, no. 20, pp. 5967–
5974, Oct. 2011.
[96] V. N. Uversky, J. Li, P. Souillac, I. S. Millett, S. Doniach, R. Jakes, M. Goedert, and A. L.
Fink, “Biophysical Properties of the Synucleins and Their Propensities to Fibrillate:
INHIBITION OF α-SYNUCLEIN ASSEMBLY BY β- AND γ-SYNUCLEINS ,” J. Biol.
Chem. , vol. 277 , no. 14 , pp. 11970–11978, Apr. 2002.
[97] I. Dikiy and D. Eliezer, “Folding and misfolding of alpha-synuclein on membranes,”
Biochim. Biophys. Acta - Biomembr., vol. 1818, no. 4, pp. 1013–1018, Apr. 2012.
[98] P. J. McLean, S. Ribich, and B. T. Hyman, “Subcellular localization of α-synuclein in
primary neuronal cultures: effect of missense mutations,” in Advances in Research on
Neurodegeneration SE - 5, Y. Mizuno, D. B. Calne, R. Horowski, W. Poewe, P. Riederer,
and M. B. H. Youdim, Eds. Springer Vienna, 2000, pp. 53–63.
[99] C. Guardia-Laguarta, E. Area-Gomez, C. Rüb, Y. Liu, J. Magrané, D. Becker, W. Voos,
E. A. Schon, and S. Przedborski, “α-Synuclein Is Localized to Mitochondria-Associated
ER Membranes,” J. Neurosci. , vol. 34 , no. 1 , pp. 249–259, Jan. 2014.
[100] N. Braidy, W.-P. Gai, Y. Xu, P. Sachdev, G. Guillemin, X.-M. Jiang, J. W. Ballard, M.
Horan, Z. Fang, B. Chong, and D. K. Chan, “Uptake and mitochondrial dysfunction of
alpha-synuclein in human astrocytes, cortical neurons and fibroblasts,” Transl.
Neurodegener., vol. 2, no. 1, p. 20, 2013.
107

[101] O. Marques and T. F. Outeiro, “Alpha-synuclein: from secretion to dysfunction and
death,” Cell Death Dis, vol. 3, 2012.
[102] M. G. Spillantini, R. A. Crowther, R. Jakes, M. Hasegawa, and M. Goedert, “α-Synuclein
in filamentous inclusions of Lewy bodies from Parkinson’s disease and dementia with
Lewy bodies,” Proc. Natl. Acad. Sci. , vol. 95 , no. 11 , pp. 6469–6473, May 1998.
[103] V. N. Uversky and D. Eliezer, “Biophysics of Parkinson’s disease: structure and
aggregation of alpha-synuclein,” Curr Protein Pept Sci, vol. 10, no. 5, pp. 483–499, 2009.
[104] H. A. Lashuel, B. M. Petre, J. Wall, M. Simon, R. J. Nowak, T. Walz, and P. T. Lansbury
Jr, “α-Synuclein, Especially the Parkinson’s Disease-associated Mutants, Forms Pore-like
Annular and Tubular Protofibrils,” J. Mol. Biol., vol. 322, no. 5, pp. 1089–1102, Oct.
2002.
[105] K. A. Conway, J. D. Harper, and P. T. Lansbury, “Accelerated in vitro fibril formation by
a mutant [alpha]-synuclein linked to early-onset Parkinson disease,” Nat Med, vol. 4, no.
11, pp. 1318–1320, 1998.
[106] J. J. Zarranz, J. Alegre, J. C. Gómez-Esteban, E. Lezcano, R. Ros, I. Ampuero, L. Vidal, J.
Hoenicka, O. Rodriguez, B. Atarés, V. Llorens, E. G. Tortosa, T. del Ser, D. G. Muñoz,
and J. G. de Yebenes, “The new mutation, E46K, of α-synuclein causes parkinson and
Lewy body dementia,” Ann. Neurol., vol. 55, no. 2, pp. 164–173, 2004.
[107] E. Masliah, E. Rockenstein, I. Veinbergs, M. Mallory, M. Hashimoto, A. Takeda, Y.
Sagara, A. Sisk, and L. Mucke, “Dopaminergic Loss and Inclusion Body Formation in αSynuclein Mice: Implications for Neurodegenerative Disorders,” Sci. , vol. 287 , no. 5456
, pp. 1265–1269, Feb. 2000.
[108] M. B. Feany and W. W. Bender, “A Drosophila model of Parkinson’s disease,” Nature,
vol. 404, no. 6776, pp. 394–398, 2000.
[109] J. B. Watson, A. Hatami, H. David, E. Masliah, K. Roberts, C. E. Evans, and M. S.
Levine, “Alterations in corticostriatal synaptic plasticity in mice overexpressing human αsynuclein,” Neuroscience, vol. 159, no. 2, pp. 501–513, Mar. 2009.
[110] J. M. George, H. Jin, W. S. Woods, and D. F. Clayton, “Characterization of a novel
protein regulated during the critical period for song learning in the zebra finch,” Neuron,
vol. 15, no. 2, pp. 361–372, Jul. 2014.
[111] A. Abeliovich, Y. Schmitz, I. Fariñas, D. Choi-Lundberg, W.-H. Ho, P. E. Castillo, N.
Shinsky, J. M. G. Verdugo, M. Armanini, A. Ryan, M. Hynes, H. Phillips, D. Sulzer, and
A. Rosenthal, “Mice Lacking α-Synuclein Display Functional Deficits in the Nigrostriatal
Dopamine System,” Neuron, vol. 25, no. 1, pp. 239–252, Jul. 2014.

108

[112] N. G. Kholodilov, M. Neystat, F. O. Tinmarla, E. Lo Steven, K. E. Larsen, D. Sulzer, and
R. E. Burke, “Increased Expression of Rat Synuclein in the Substantia Nigra Pars
Compacta Identified by mRNA Differential Display in a Model of Developmental Target
Injury,” J. Neurochem., vol. 73, no. 6, pp. 2586–2599, Dec. 1999.
[113] L. Stefanis, “α-Synuclein in Parkinson’s Disease,” Cold Spring Harb Perspect Med, 2012.
[114] S. Chandra, G. Gallardo, R. Fernández-Chacón, O. M. Schlüter, and T. C. Südhof, “αSynuclein Cooperates with CSPα in Preventing Neurodegeneration,” Cell, vol. 123, no. 3,
pp. 383–396, Jul. 2014.
[115] D. D. Murphy, S. M. Rueter, J. Q. Trojanowski, and V. M.-Y. Lee, “Synucleins Are
Developmentally Expressed, and α-Synuclein Regulates the Size of the Presynaptic
Vesicular Pool in Primary Hippocampal Neurons,” J. Neurosci. , vol. 20 , no. 9 , pp.
3214–3220, May 2000.
[116] F. LS, “Concentric hyalin intraneuronal inclusions of Lewy type in the brains of elderly
persons (50 incidental cases): relationship to parkinsonism,” J. Am. Geriatr. Soc., 1969.
[117] L. V Kalia, S. K. Kalia, P. J. McLean, A. M. Lozano, and A. E. Lang, “α-Synuclein
oligomers and clinical implications for Parkinson disease,” Ann. Neurol., vol. 73, no. 2,
pp. 155–169, Feb. 2013.
[118] O. M. A. El-Agnaf, S. A. Salem, K. E. Paleologou, M. D. Curran, M. J. Gibson, J. A.
Court, M. G. Schlossmacher, and D. Allsop, “Detection of oligomeric forms of αsynuclein protein in human plasma as a potential biomarker for Parkinson’s disease,”
FASEB J. , vol. 20 , no. 3 , pp. 419–425, Mar. 2006.
[119] T. et al Tokuda, “Detection of elevated levels of α-synuclein oligomers in CSF from
patients with Parkinson disease.,” Neurology, 2010.
[120] D. P. B. Karpinar Madhu Babu Gajula Kügler, Sebastian Opazo, Felipe Rezaei‐Ghaleh,
Nasrollah Wender, Nora Kim, Hai‐Young Taschenberger, Grit Falkenburger, Björn H
Heise, Henrike Kumar, Ashutosh Riedel, Dietmar Fichtner, Lars Voigt, Aaron Braus,
Gerhard H Giller, , “Pre‐fibrillar α‐synuclein variants with impaired β‐structure increase
neurotoxicity in Parkinson&#039;s disease models,” EMBO J., vol. 28, no. 20, pp. 3256–
3268, Sep. 2009.
[121] O. W. Wan and K. K. K. Chung, “The Role of Alpha-Synuclein Oligomerization and
Aggregation in Cellular and Animal Models of Parkinson’s Disease,” PLoS One, vol. 7,
no. 6, p. e38545, 2012.
[122] N. Gosavi, H.-J. Lee, J. S. Lee, S. Patel, and S.-J. Lee, “Golgi Fragmentation Occurs in
the Cells with Prefibrillar α-Synuclein Aggregates and Precedes the Formation of Fibrillar
Inclusion,” J. Biol. Chem. , vol. 277 , no. 50 , pp. 48984–48992, Dec. 2002.

109

[123] R. A. Bodner, T. F. Outeiro, S. Altmann, M. M. Maxwell, S. H. Cho, B. T. Hyman, P. J.
McLean, A. B. Young, D. E. Housman, and A. G. Kazantsev, “Pharmacological
promotion of inclusion formation: A therapeutic approach for Huntington’s and
Parkinson’s diseases,” Proc. Natl. Acad. Sci. United States Am. , vol. 103 , no. 11 , pp.
4246–4251, Mar. 2006.
[124] B. Winner, R. Jappelli, S. K. Maji, P. A. Desplats, L. Boyer, S. Aigner, C. Hetzer, T.
Loher, M. Vilar, S. Campioni, C. Tzitzilonis, A. Soragni, S. Jessberger, H. Mira, A.
Consiglio, E. Pham, E. Masliah, F. H. Gage, and R. Riek, “In vivo demonstration that αsynuclein oligomers are toxic,” Proc. Natl. Acad. Sci. , vol. 108 , no. 10 , pp. 4194–4199,
Mar. 2011.
[125] K. M. Danzer, D. Haasen, A. R. Karow, S. Moussaud, M. Habeck, A. Giese, H.
Kretzschmar, B. Hengerer, and M. Kostka, “Different Species of α-Synuclein Oligomers
Induce Calcium Influx and Seeding,” J. Neurosci. , vol. 27 , no. 34 , pp. 9220–9232, Aug.
2007.
[126] V. N. Uversky, J. Li, and A. L. Fink, “Evidence for a partially folded intermediate in
alpha-synuclein fibril formation.,” J. Biol. Chem., 2001.
[127] C. W. Bertoncini, Y.-S. Jung, C. O. Fernandez, W. Hoyer, C. Griesinger, T. M. Jovin, and
M. Zweckstetter, “Release of long-range tertiary interactions potentiates aggregation of
natively unstructured α-synuclein,” Proc. Natl. Acad. Sci. United States Am. , vol. 102 ,
no. 5 , pp. 1430–1435, Feb. 2005.
[128] V. N. Uversky, “What does it mean to be natively unfolded?,” Eur. J. Biochem., vol. 269,
no. 1, pp. 2–12, Jan. 2002.
[129] W. S. Davidson, A. Jonas, D. F. Clayton, and J. M. George, “Stabilization of α-Synuclein
Secondary Structure upon Binding to Synthetic Membranes,” J. Biol. Chem. , vol. 273 ,
no. 16 , pp. 9443–9449, Apr. 1998.
[130] T. Bartels, J. G. Choi, and D. J. Selkoe, “[agr]-Synuclein occurs physiologically as a
helically folded tetramer that resists aggregation,” Nature, vol. 477, no. 7362, pp. 107–
110, Sep. 2011.
[131] W. Wang, I. Perovic, J. Chittuluru, A. Kaganovich, L. T. T. Nguyen, J. Liao, J. R. Auclair,
D. Johnson, A. Landeru, A. K. Simorellis, S. Ju, M. R. Cookson, F. J. Asturias, J. N. Agar,
B. N. Webb, C. Kang, D. Ringe, G. A. Petsko, T. C. Pochapsky, and Q. Q. Hoang, “A
soluble α-synuclein construct forms a dynamic tetramer,” Proc. Natl. Acad. Sci. , vol. 108
, no. 43 , pp. 17797–17802, Oct. 2011.
[132] B. Fauvet, M. K. Mbefo, M.-B. Fares, C. Desobry, S. Michael, M. T. Ardah, E. Tsika, P.
Coune, M. Prudent, N. Lion, D. Eliezer, D. J. Moore, B. Schneider, P. Aebischer, O. M.
El-Agnaf, E. Masliah, and H. A. Lashuel, “α-Synuclein in Central Nervous System and

110

from Erythrocytes, Mammalian Cells, and Escherichia coli Exists Predominantly as
Disordered Monomer,” J. Biol. Chem. , vol. 287 , no. 19 , pp. 15345–15364, May 2012.
[133] V. N. Uversky, “A Protein-Chameleon: Conformational Plasticity of α-Synuclein, a
Disordered Protein Involved in Neurodegenerative Disorders,” J. Biomol. Struct. Dyn.,
vol. 21, no. 2, 2003.
[134] M. H. Polymeropoulos, C. Lavedan, E. Leroy, S. E. Ide, A. Dehejia, A. Dutra, B. Pike, H.
Root, J. Rubenstein, R. Boyer, E. S. Stenroos, S. Chandrasekharappa, A. Athanassiadou,
T. Papapetropoulos, W. G. Johnson, A. M. Lazzarini, R. C. Duvoisin, G. Di Iorio, L. I.
Golbe, and R. L. Nussbaum, “Mutation in the α-Synuclein Gene Identified in Families
with Parkinson’s Disease,” Sci. , vol. 276 , no. 5321 , pp. 2045–2047, Jun. 1997.
[135] C. W. Bertoncini, C. O. Fernandez, C. Griesinger, T. M. Jovin, and M. Zweckstetter,
“Familial Mutants of α-Synuclein with Increased Neurotoxicity Have a Destabilized
Conformation,” J. Biol. Chem. , vol. 280 , no. 35 , pp. 30649–30652, Sep. 2005.
[136] K. A. Conway, S.-J. Lee, J.-C. Rochet, T. T. Ding, R. E. Williamson, and P. T. Lansbury,
“Acceleration of oligomerization, not fibrillization, is a shared property of both αsynuclein mutations linked to early-onset Parkinson’s disease: Implications for
pathogenesis and therapy,” Proc. Natl. Acad. Sci. , vol. 97 , no. 2 , pp. 571–576, Jan.
2000.
[137] L. Narhi, S. J. Wood, S. Steavenson, Y. Jiang, G. M. Wu, D. Anafi, S. A. Kaufman, F.
Martin, K. Sitney, P. Denis, J.-C. Louis, J. Wypych, A. L. Biere, and M. Citron, “Both
Familial Parkinson’s Disease Mutations Accelerate α-Synuclein Aggregation,” J. Biol.
Chem. , vol. 274 , no. 14 , pp. 9843–9846, Apr. 1999.
[138] E. A. Greenbaum, C. L. Graves, A. J. Mishizen-Eberz, M. A. Lupoli, D. R. Lynch, S. W.
Englander, P. H. Axelsen, and B. I. Giasson, “The E46K Mutation in α-Synuclein
Increases Amyloid Fibril Formation,” J. Biol. Chem. , vol. 280 , no. 9 , pp. 7800–7807,
Mar. 2005.
[139] L. R. Lemkau, G. Comellas, S. W. Lee, L. K. Rikardsen, W. S. Woods, J. M. George, and
C. M. Rienstra, “Site-Specific Perturbations of Alpha-Synuclein Fibril Structure by the
Parkinson’s Disease Associated Mutations A53T and E46K,” PLoS One, vol. 8, no. 3, p.
e49750, Mar. 2013.
[140] E. Hazy, M. Bokor, L. Kalmar, A. Gelencser, P. Kamasa, K.-H. Han, K. Tompa, and P.
Tompa, “Distinct Hydration Properties of Wild-Type and Familial Point Mutant A53T of
α-Synuclein Associated with Parkinson’s Disease,” Biophys. J., vol. 101, no. 9, pp. 2260–
2266, Jul. 2014.
[141] O. Wise-Scira, A. Dunn, A. K. Aloglu, I. T. Sakallioglu, and O. Coskuner, “Structures of
the E46K Mutant-Type α-Synuclein Protein and Impact of E46K Mutation on the

111

Structures of the Wild-Type α-Synuclein Protein,” ACS Chem. Neurosci., vol. 4, no. 3, pp.
498–508, Jan. 2013.
[142] C. C. Rospigliosi, S. McClendon, A. W. Schmid, T. F. Ramlall, P. Barré, H. A. Lashuel,
and D. Eliezer, “E46K Parkinson’s-Linked Mutation Enhances C-Terminal-to-N-Terminal
Contacts in α-Synuclein,” J. Mol. Biol., vol. 388, no. 5, pp. 1022–1032, May 2009.
[143] R. A. Fredenburg, C. Rospigliosi, R. K. Meray, J. C. Kessler, H. A. Lashuel, D. Eliezer,
and P. T. Lansbury, “The Impact of the E46K Mutation on the Properties of α-Synuclein
in Its Monomeric and Oligomeric States†,” Biochemistry, vol. 46, no. 24, pp. 7107–7118,
May 2007.
[144] K. E. Paleologou, A. W. Schmid, C. C. Rospigliosi, H.-Y. Kim, G. R. Lamberto, R. A.
Fredenburg, P. T. Lansbury, C. O. Fernandez, D. Eliezer, M. Zweckstetter, and H. A.
Lashuel, “Phosphorylation at Ser-129 but Not the Phosphomimics S129E/D Inhibits the
Fibrillation of α-Synuclein,” J. Biol. Chem. , vol. 283 , no. 24 , pp. 16895–16905, Jun.
2008.
[145] D. Eliezer, E. Kutluay, R. Bussell Jr, and G. Browne, “Conformational properties of αsynuclein in its free and lipid-associated states,” J. Mol. Biol., vol. 307, no. 4, pp. 1061–
1073, Apr. 2001.
[146] R. Bussell and D. Eliezer, “Residual Structure and Dynamics in Parkinson’s Diseaseassociated Mutants of α-Synuclein,” J. Biol. Chem. , vol. 276 , no. 49 , pp. 45996–46003,
Dec. 2001.
[147] J. D. Perlmutter, A. R. Braun, and J. N. Sachs, “Curvature Dynamics of α-Synuclein
Familial Parkinson Disease Mutants: MOLECULAR SIMULATIONS OF THE
MICELLE- AND BILAYER-BOUND FORMS ,” J. Biol. Chem. , vol. 284 , no. 11 , pp.
7177–7189, Mar. 2009.
[148] H.-J. Koo, M. Y. Choi, and H. Im, “Aggregation-defective α-synuclein mutants inhibit the
fibrillation of Parkinson’s disease-linked α-synuclein variants,” Biochem. Biophys. Res.
Commun., vol. 386, no. 1, pp. 165–169, Aug. 2009.
[149] M. Zhu, S. Han, and A. L. Fink, “Oxidized quercetin inhibits α-synuclein fibrillization,”
Biochim. Biophys. Acta - Gen. Subj., vol. 1830, no. 4, pp. 2872–2881, Apr. 2013.
[150] M. Caruana, T. Högen, J. Levin, A. Hillmer, A. Giese, and N. Vassallo, “Inhibition and
disaggregation of α-synuclein oligomers by natural polyphenolic compounds,” FEBS
letters, vol. 585, no. 8. Elsevier Science B.V., pp. 1113–1120, 20-Apr-2011.
[151] N. Ostrerova-Golts, L. Petrucelli, J. Hardy, J. M. Lee, M. Farer, and B. Wolozin, “The
A53T α-Synuclein Mutation Increases Iron-Dependent Aggregation and Toxicity,” J.
Neurosci. , vol. 20 , no. 16 , pp. 6048–6054, Aug. 2000.

112

[152] M. Zhu, S. Rajamani, J. Kaylor, S. Han, F. Zhou, and A. L. Fink, “The Flavonoid
Baicalein Inhibits Fibrillation of α-Synuclein and Disaggregates Existing Fibrils,” J. Biol.
Chem. , vol. 279 , no. 26 , pp. 26846–26857, Jun. 2004.
[153] D. E. Ehrnhoefer, J. Bieschke, A. Boeddrich, M. Herbst, L. Masino, R. Lurz, S.
Engemann, A. Pastore, and E. E. Wanker, “EGCG redirects amyloidogenic polypeptides
into unstructured, off-pathway oligomers,” Nat Struct Mol Biol, vol. 15, no. 6, pp. 558–
566, Jun. 2008.
[154] J. Bieschke, J. Russ, R. P. Friedrich, D. E. Ehrnhoefer, H. Wobst, K. Neugebauer, and E.
E. Wanker, “EGCG remodels mature α-synuclein and amyloid-β fibrils and reduces
cellular toxicity,” Proc. Natl. Acad. Sci. , vol. 107 , no. 17 , pp. 7710–7715, Apr. 2010.
[155] M. Fokkens, T. Schrader, and F.-G. Klärner, “A Molecular Tweezer for Lysine and
Arginine,” J. Am. Chem. Soc., vol. 127, no. 41, pp. 14415–14421, Sep. 2005.
[156] A. H. Mao, S. L. Crick, A. Vitalis, C. L. Chicoine, and R. V Pappu, “Net charge per
residue modulates conformational ensembles of intrinsically disordered proteins,” Proc.
Natl. Acad. Sci. , vol. 107 , no. 18 , pp. 8183–8188, May 2010.
[157] R. K. Das and R. V Pappu, “Conformations of intrinsically disordered proteins are
influenced by linear sequence distributions of oppositely charged residues,” Proc. Natl.
Acad. Sci. , Jul. 2013.
[158] S. Prabhudesai, S. Sinha, A. Attar, A. Kotagiri, A. Fitzmaurice, R. Lakshmanan, M.
Ivanova, J. Loo, F.-G. Klärner, T. Schrader, M. Stahl, G. Bitan, and J. Bronstein, “A
Novel ‘Molecular Tweezer’ Inhibitor of α-Synuclein Neurotoxicity in Vitro and in Vivo,”
Neurotherapeutics, vol. 9, no. 2, pp. 464–476, 2012.
[159] A. Attar, C. Ripoli, E. Riccardi, P. Maiti, D. D. Li Puma, T. Liu, J. Hayes, M. R. Jones, K.
Lichti-Kaiser, F. Yang, G. D. Gale, C. Tseng, M. Tan, C.-W. Xie, J. L. Straudinger, F.-G.
Klärner, T. Schrader, S. A. Frautschy, C. Grassi, and G. Bitan, “Protection of primary
neurons and mouse brain from Alzheimer’s pathology by molecular tweezers,” Brain ,
vol. 135 , no. 12 , pp. 3735–3748, Dec. 2012.
[160] S. J. Sahl, L. E. Weiss, W. C. Duim, J. Frydman, and W. E. Moerner, “Cellular Inclusion
Bodies of Mutant Huntingtin Exon 1 Obscure Small Fibrillar Aggregate Species,” Sci.
Rep., vol. 2, Nov. 2012.

113