!"#!$%$!&!!'
()
*
)+$#&
,-.
/
/ & 0 ,1$. - & 2
3 - 45 5 & ) / 1$
&
1 / 1$ ) &6
)
/
/
1$
& 1$ &7)/
1$
)/
-
/)
/
/
&
)'"
3
/
1$)/
0
/
8)
-0
)/
0
-&
1$
/
-)/
&
!
#!$%
9::&&:
;<9999$ ==>
7'=%?=$%@>=%# >
7'=%?=$%@>=%#>$
)$!@=$
SOD1 Aggregation
Relevance of thermodynamic stability
Lisa Lang
SOD1 Aggregation
Relevance of thermodynamic stability
Lisa Lang
List of Publications
Primary publications
I
Fibrillation precursor of superoxide dismutase 1 revealed by gradual
tuning of the protein-folding equilibrium
Lang L, Kurnik M, Danielsson J, Oliveberg M.
Proc Natl Acad Sci U S A. 2012 Oct 30;109(44):17868-73.
II
SOD1 aggregation in ALS mice shows simplistic test-tube behaviour
Lang L, Zetterström P, Brännström T, Marklund SL, Danielsson J,
Oliveberg M.
Proc Natl Acad Sci U S A. 2015 Aug 11;112(32):9878-83.
III
Structural and kinetic analysis of protein-aggregate strains in vivo using
binary epitope mapping
Bergh J, Zetterström P, Andersen PM, Brännström T, Graffmo KS, Jonsson
PA, Lang L, Danielsson J, Oliveberg M, Marklund SL.
Proc Natl Acad Sci U S A. 2015 Apr 7;112(14):4489-94.
IV
Thermodynamics of protein destabilization in live cells
Danielsson J, Mu X, Lang L, Wang H, Binolfi A, Theillet FX, Bekei B,
Logan DT, Selenko P, Wennerström H, Oliveberg M.
Proc Natl Acad Sci U S A. 2015 Oct 6;112(40):12402-7.
Additional publications
i
Cutting off functional loops from homodimeric enzyme superoxide
dismutase 1 (SOD1) leaves monomeric β-barrels
Danielsson J, Kurnik M, Lang L, Oliveberg M.
J Biol Chem. 2011 Sep 23;286(38):33070-83.
ii
Cytotoxicity of superoxide dismutase 1 in cultured cells is linked to Zn2+
chelation
Johansson AS, Vestling M, Zetterström P, Lang L, Leinartaitė L, Karlström
M, Danielsson J, Marklund SL, Oliveberg M.
PLoS One. 2012;7(4):e36104.
iii
Global structural motions from the strain of a single hydrogen bond
Danielsson J, Awad W, Saraboji K, Kurnik M, Lang L, Leinartaite L,
Marklund SL, Logan DT, Oliveberg M.
Proc Natl Acad Sci U S A. 2013 Mar 5;110(10):3829-34.
iv
Pruning the ALS-associated protein SOD1 for in-cell NMR
Danielsson J, Inomata K, Murayama S, Tochio H, Lang L, Shirakawa M,
Oliveberg M.
J Am Chem Soc. 2013 Jul 17;135(28):10266-9.
v
Solid-state NMR studies of metal-free SOD1 fibrillar structures
Banci L, Blaževitš O, Cantini F, Danielsson J, Lang L, Luchinat C, Mao J,
Oliveberg M, Ravera E.
J Biol Inorg Chem. 2014 Jun;19(4-5):659-66.
vi
Tricking a protein to swap strands
Wang H, Lang L, Logan DT, Danielsson J, Oliveberg M.
J Am Chem Soc. 2016 Dec 7;138(48):15571-15579.
vii
Physical-chemical code for quinary protein interactions in E. coli.
Xin Mu, Seongil C, Lang L, Mowray D, Dokholyan N, Danielsson J,
Oliveberg M.
Manuscript.
Contents
List of Publications ......................................................................................... vi
Contents ......................................................................................................... ix
Abbreviations ................................................................................................. x
Introduction .................................................................................................. 11
Amyotrophic lateral sclerosis ........................................................................................ 11
Superoxide dismutase 1 and ALS ................................................................................ 12
Introducing SOD1 ......................................................................................................... 14
SOD1 folding and stability ............................................................................................ 15
SOD1 misfolding and aggregation ................................................................................ 18
Aggregation pathways .................................................................................................. 21
Results and Discussion ................................................................................ 26
The aggregating state of SOD1 .................................................................................... 26
The aggregation mechanism of SOD1 ......................................................................... 29
SOD1 aggregation in vivo ............................................................................................. 32
Sequence determinants for SOD1 aggregation ............................................................ 34
Destabilisation of SOD1 in vivo .................................................................................... 38
Summary and Concluding remark ............................................................... 40
Populärvetenskaplig sammanfattning .......................................................... 44
Acknowledgements ...................................................................................... 46
References ................................................................................................... 48
Abbreviations
ALS
SOD1
hSOD1
ThT
EM
D
N
ssNMR
CNS
Amino acids
A
C
D
E
F
G
H
I
K
L
M
N
P
Q
R
S
T
V
W
Y
Alanine
Cysteine
Aspartate
Glutamate
Phenylalanine
Glycine
Histidine
Isoleucine
Lysine
Leucine
Methionine
Asparagine
Proline
Glutamine
Arginine
Serine
Threonine
Valine
Tryptophan
Tyrosine
Amyotrophic lateral sclerosis
Superoxide dismutase 1
Human SOD1
Thioflavin T
Electron microscopy
The denatured ensemble
The native ensemble
Solid-state NMR
Central nervous system
Introduction
Proteins typically gain their physiological function by folding into their
unique three-dimensional structure. For optimal function, however both the
thermodynamic stability and the rigidity of a protein structure have to be
kept within certain limits. These properties can be affected directly by mutations and indirectly by changed conditions, due to old age or environmental
stress, causing the protein to misfold and aggregate. Consequently protein
aggregates and inclusion bodies are pathological hallmarks in many agerelated diseases. The structurally ordered amyloid conformations appear to
be accessible to many, if not all, protein sequences suggesting that the protein misfolding disorders share a common mechanism. The term ‘misfolded’
is here referring to a low energy, non-natively folded protein that is significantly populated and can be part of an aggregated structure although it is not
a criterion.
Amyotrophic lateral sclerosis
Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disorder that
causes degeneration of upper and lower motor neurons, resulting in muscle
atrophy (Fig. 1). Initial symptoms are often muscle weakness rapidly followed by paralysis and subsequent death due to respiratory failure typically
within 2-3 years from symptom onset (1). ALS is the third most common
neurodegenerative disease and the most common adult-onset motor neuron
disease (2).
Although the underlying cause for ALS is yet obscure, proposed disease
mechanisms include impaired mitochondrial function (3, 4), induction of
endoplasmic reticulum (ER) stress (5, 6), reduced proteasomal activity (7, 8)
and reduced autophagy (9) as well as formation of neurotoxic aggregates
(10-12). Nevertheless, there are infallible hallmarks for ALS; the hereditary
or familial ALS (fALS) forms, with mutations in certain genes accounting
for ~10% of all ALS-cases. The remaining ~90% are considered sporadic
(sALS) and occur randomly throughout the community (2). Still there are
pathological and clinical similarities between fALS and sALS that indicate a
common disease mechanism (13). Due to the histopatological proteinaceous
inclusions found in both fALS and sALS patients (Fig. 1), ALS has been
11
classified as a protein misfolding disorder like Alzheimer’s disease and Parkinson’s disease.
A general risk factor for ALS as well as other neurodegenerative diseases is
age (14). ALS patients are apparently healthy until middle age when they
suddenly undergo rapid neurological degeneration. It is possible that agerelated decline in proteostasis and energy metabolism leads to a decreased
tolerance to stress factors and an impaired ability to prevent accumulation of
misfolded and aggregated proteins. Mutations in fALS might contribute to
these age-related changes as they move forward disease onset about 10 years
compared with sALS (13). However, despite much effort and scientific progress the cause and mechanism behind ALS remain unknown and there is
still no effective cure for this fatal disease.
Fig. 1. Schematic illustration of affected neurons in ALS and the muscle areas they innervate.
Insets show an immunohistochemically stained nerve cell and enlargement of the putative
fibrillar aggregates inside. Photograph of mouse nerve cell from Paper III. EM micrograph of
SOD1 fibers.
Superoxide dismutase 1 and ALS
Approximately 6% of all ALS patients carry mutations in the gene coding
for the free radical scavenging enzyme superoxide dismutase 1 (SOD1) (1518). Of those approximately half belong to the hereditary cases and half to
the sporadic cases, where the latter most likely are the result of incomplete
penetrance or incomplete family history information as well as de novo mutations (19, 20). Since 1993 when the first mutations where found, more than
180 mutations scattered over the SOD1 sequence have been identified (ALS
Online Genetic database, http://alsod.iop.kcl.ac.uk) (Fig. 2). The majority of
the mutations cause single amino acid replacement but there are also dele-
12
tions and insertions in the SOD1 gene causing frame shifts and truncated
protein products.
Human SOD1 (hSOD1) is an abundant intracellular protein constituting
about 0.2 and 0.4 % of total protein in grey matter of brain and in liver respectively (21). The mature protein is an antioxidant enzyme, which converts toxic superoxide radicals to hydrogen peroxide and water (21). Loss of
SOD1 function would potentially increase these toxic species and be a possible cause for neuronal damage. SOD1 knockout mice, however, have been
shown not to develop ALS (22) and most ALS-associated mutants maintained their enzymatic activity (23). Furthermore, overexpression of both
wild type and mutant hSOD1, in rodent models causes ALS-like symptoms
(24, 25), suggesting a gain of toxic function.
Genetic and experimental data show that the ALS-mutations predominantly
reduce the negative net charge (26) and/or decrease the structural stability
(27, 28) of SOD1, which increases the disposition to misfold and aggregate.
Indeed, aggregates containing mutant as well as wild type SOD1 are a hallmark of the disease and are found both in the central nervous system of diseased ALS patients (29-31) and in ALS transgenic mice models (24) suggesting that SOD1 is involved in both fALS and sALS. Furthermore, SOD1
fibrillation can be induced in vitro through prolonged shaking or stirring of
the sample (32-34). Together these observations have put research focus on
the role of SOD1 misfolding and aggregation in ALS pathology.
Fig. 2. Loci of missense mutations in the SOD1 primary and secondary structure (pdbID 2C9U).
Source: ALS Online Genetic database, http://alsod.iop.kcl.ac.uk.
13
Introducing SOD1
SOD1 is a ubiquitously expressed homodimeric metalloenzyme belonging to
the immunoglobulin-like family (35). Each of the two 153 residue subunits
is an eight-stranded β-barrel with a Greek key fold that ligates one copper
and one zinc ion (Fig. 3A) (36). The metals are coordinated by two long
active-site loops. Loop IV, or the Zn-binding loop, coordinates the structurally important zinc ion and forms, together with loop VII, the active-site
tunnel for specific substrate entrance (37). Loop VII, also called the electrostatic loop, contains the charged residues needed to guide superoxide into the
active site to the catalytically active copper ion (37). The catalytically active
copper ion in its reduced form (Cu1+) is directly bound to the β-barrel, coordinated by H46, H48, and H120 and the additional ligand H63 in the oxidised form (Cu2+) (38). H63 is shared by the structurally important Zn2+ ion,
which is further coordinated by H71, H80 and Asp83 (37, 39). Coordination
of the metals, especially in the zinc site, renders both monomeric and dimeric SOD1 remarkably stable and rigid proteins, and the native holo-dimer is
one of the most stable proteins known (40-42). The apo- state however is
significantly less stable and displays dynamic motions in the two large loops
as well as in their flanking regions (43-45). Each subunit also contains an
intramolecular disulphide bond between Cys57 and Cys146, which is rare
for intracellular proteins due to the reducing environment of the cytosol and
indicates its importance for structure and function. Accordingly, the intact
disulphide bond is required for full dimer stability by anchoring loop IV to
the β-barrel (46, 47). Consequently, the reduced flexibility of loop IV increases SOD1’s affinity for zinc, which further enhances dimer stability (48,
49). The oxidation state of the disulphide bond is also believed to be important for in vivo copper-loading, assisted by the chaperone CCS (47, 50).
Fig. 3. Chrystal structure of SOD1 (pdbID 2C9U). A. Structure of the SOD1 dimer. The active
site loops IV and VII are coloured green and blue respectively. The disulphide bond (red), zink
ion (blue) and copper ion (orange) are shown as spheres. B. The eight-stranded apoSOD1
monomer suggested being the toxic factor in ALS.
14
ing barrier (Fig. 4A). A key requirement for a two-state reaction is that folding is reversible i.e. that the ground states of refolding and unfolding are
unchanged. The relation between the two states, in terms of concentration, is
defined by the equilibrium constant KD-N.
K D−N = [D]/[N] = k u /kf ,
(Eq. 1)
Where the brackets denote concentration. KD-N is a measure of the protein
stability. At constant solution conditions and protein concentration, the expression for the energy difference between D and N is
∆GD-N = -2.3RTlogK D-N = -2.3RTlog[D]/[N]
(Eq. 2)
where R is the universal gas constant (1.987 cal K-1 mol-1) and T is the temperature in Kelvin.
Fig. 4. The folding free-energy barrier (A) and the equilibrium denaturation curve and the chevron plot of a two-state protein (B).
The thermodynamic equilibrium is studied by gradual change of the solution
conditions from those that favour N to those favouring D (66) (Fig. 4B top
panel). Common denaturing agents are urea or guanidinium chloride
(GdmCl), pH or simply increasing the temperature. Often the intrinsic tryptophan fluorescence of the protein is used to observe the denaturation process, as it is dependent on environmental factors such as polarity and dynamics thereby shifting the excitation and emission maxima, yielding different
fluorescent properties for N and D (67). The relation between the free energy
of unfolding in different denaturant concentrations is empirically observed to
be linear (66)
16
H O
2
GD-N GD-N
mD-N denaturant
(Eq. 3)
or
H O
2
logKD-N logKD-N
mD-N denaturant
(Eq. 4)
where the slope mD-N relates to the change in solvent accessible surface area
upon denaturation (68). Solving the equation at the transition midpoint (MP
= [denaturant]1/2) where D and N are equally populated (logKD-N[denaturant] = 0)
gives logKD-N in denaturant free conditions according to -logKD-NH2O = mDN[denaturant]1/2 (Fig. 4B top panel).
Equilibrium measurements do not provide any information of the folding
pathway of the protein e.g. how the protein traverses the free-energy landscape from D to N. According to transition state theory there is a free-energy
barrier of the reaction coordinate between D and N with the transition-state
ensemble (‡) corresponding to the highest potential energy (Fig. 4A). The ‡
describes a set of short-lived protein conformations that cannot be observed
directly because of their transient nature (69) and they have the same probability to go back to D as to proceed to N. The height of the free-energy barrier, which is the amount of energy needed to reach the ‡ from N, is called the
kinetic stability of the protein (∆Gu) and is characterised by the unfolding
rate constant (ku) (Fig. 4A). Consequently there is a ∆Gf defined by the folding rate constant (kf) (Fig. 4A). The unfolding and refolding rate constants
are obtained in two complementary experiments measuring the relaxation
from one state to another. The ku is measured by rapidly mixing protein with
denaturant solution of decreasing concentration until the midpoint is
reached. Similarly the kf is acquired, but first the protein is unfolded in rather
high concentration of denaturant and subsequently refolded by rapid mixing
in solutions of increasing denaturant concentrations starting from buffer
only. The kinetic data is generally presented in a chevron plot (64) (Fig. 4B
bottom panel), named after the characteristic v-shaped (goat horn-like) relation between the logarithm of the observed rate constants of a two-state folding protein and the concentration of denaturant. Linear extrapolation, according to Eqs. 5 and 6, yield the folding and unfolding rate constants in aqueous
solution.
logf logf
H2 O
m denaturant
(Eq. 5)
H2 O
m denaturant
(Eq. 6)
log logu
17
where the two slopes, mf and mu, are the folding and unfolding m-values that
quantify the structural gain upon folding before and after the ‡ respectively.
The sum of the m-values, mD-N, is related to the total change in solvent accessible surface area upon folding or unfolding (Eq. 3, 4). Parallel to the
equilibrium measurements (Eq. 1) this model can be used to determine the
thermodynamic stability (64)
∆GD-N = -2.3RTlogKD-N = -2.3RTlog (log ku-log kf)
(Eq. 7)
confirming the two-state folding mechanism, if the two models give the
same value. Two-state proteins normally fold in the sub second timescale.
SOD1 however, due to its β-rich nature, displays relatively slow folding
rates compared to proteins with local secondary structure elements like αhelices.
SOD1 misfolding and aggregation
In addition to the native fold, proteins can also adopt low energy structures
in the form of aggregates as well as other conformations that are different
from the native structure. This process is called protein misfolding and is
often linked to human disease, although there are examples of non-disease
proteins that form aggregated structures in vivo (70). The term ‘misfolded’ is
here referring to a low energy, non-natively folded protein that is significantly populated. It can be part of an aggregated structure although it is not a
criterion. Non-native or partially folded intermediates are often implicated as
precursors for amyloid formation in protein-misfolding diseases (71-73). For
monomeric apoSOD1 however, no such species are populated during the
folding process ((74) Paper I, Paper vi (75)).
Protein misfolding and aggregation is, like protein folding, controlled by the
energy landscape with free-energy barriers and local and global minima,
where amyloid fibrils are the most stable and well-defined structures (76)
(Fig. 5). Although the aggregates can be more thermodynamically stable
than the native state, proteins still remain soluble for long periods of time as
a consequence of the high kinetic barriers that separate the native and the
aggregated states (76-78). At supersaturated conditions mechanical and
chemical stress may reduce the height of the kinetic barrier to aggregation in
vitro and similarly can age-related and mutational stress in vivo (Fig. 5) (79).
Many ALS-mutations reduce the net repulsive charge of the protein (26)
which facilitates interactions between SOD1 molecules as well as other cellular components. The reduced repulsive force lowers the kinetic barrier for
18
aggregation and potential co-aggregation (80, 81). Cytosolic proteins that
depend on movability have evolved to remain soluble but only to the levels
required to meet their function. Accordingly, the protein concentration in
vivo is typically close to the critical level where aggregation set in (82).
Many proteins in the cell are under constant competition between folding,
misfolding and aggregation. Disturbances to these delicate equilibriums are a
likely cause of impaired cellular functions and disease.
Fig. 5. Schematic free-energy profile showing the transition between the native state (N) and
the denatured states (D) as well as transitions from D or N to the aggregated states (A). Although A is more thermodynamically stable than both D and N the high kinetic barrier reduces
the probability to adopt the aggregated conformation. Mechanical and chemical stress can
reduce the kinetic barrier to aggregation in vitro and so can age-related and mutational stress in
vivo, dotted line.
Inclusions of aggregated SOD1 are found in motor neurons of both sporadic
and familial ALS patients and clearly indicate involvement of SOD1 in ALS
pathology (29-31). Several ALS-associated mutations decrease the affinity
for copper/zinc and the oxidation status of the disulphide bond (83). Therefore, in the reducing environment of the cytosol with high metal-chelating
capacity, the fraction of the destabilised disulphide reduced apoSOD1, the
suggested precursor for aggregation, will be increased due to ALS mutation.
Nonetheless, the mechanism behind pathological SOD1 aggregate formation
remains unclear. Electron micrographs of SOD1-positive inclusions from
fALS patients has identified their fibrillar morphology (10), although they
19
are not stained by the amyloid binding dyes Thioflavin S and Congo Red
(10, 84). Possibly the granule-coated morphology (10) or the quaternary
packing of the SOD1 fibrillar structures shields them from interaction with
the dyes. Even so, SOD1 has been shown to form amyloid fibrils both in
vitro at destabilising conditions (Paper I) (32, 33) and in transgenic mice
models expressing high levels of human SOD1 (32, 85).
It is still controversial whether protein aggregation is the actual cause or a
mere result of neurodegeneration in ALS and other neurodegenerative diseases. There are suggestions that aggregates are not the prime cause of neural damage but rather the toxic species that arise much earlier in the misfolding process (10, 86, 87). Unfortunately, these early species are intrinsically
difficult to isolate and study due to their transient nature; they transform into
larger aggregates or disassemble into monomers again in a too short time
frame. Studies of SOD1 aggregation kinetics in vitro has failed to record the
initial nucleation events only detecting the larger aggregates (Paper I) (32,
33, 88). Non-native metastable oligomeric structures of disulphide oxidised
SOD1 have been produced in vitro at physiological (89, 90) and acidic conditions (91) and have been suggested as cytotoxic agents in ALS. Furthermore, it is still unclear whether in vivo SOD1 aggregates are disulphide oxidised or reduced. In favour of the reduced state all four cysteines in SOD1
can be subjected to substitution in ALS patients (92-94) and SOD1 with all
four cysteines removed still form aggregates in vitro (95). Additionally, CNS
analysed from ALS transgenic mice models have been shown to contain
substantial fractions of disulphide-reduced SOD1 (83) and SOD1 in isolated
aggregates from transgenic mice is disulphide reduced (96).
Accumulating evidence indicate a prion-like spread of misfolded or aggregated protein in ALS and other neurodegenerative diseases such as Alzheimer’s disease, Parkinson’s disease, Huntington’s disease (97, 98). The
prion, a misfolded or aggregated infectious form of the protein, interacts
with other native copies of the same protein and induce their transition into
the aberrant form (99). The infective conformation is commonly called a
seed and the process is termed seeding. Accordingly, aggregated SOD1,
isolated from diseased tissue have been shown to induce aggregation of mutant and wild-type SOD1 in vitro (88, 100) and in cultured cells (101, 102).
Such prion-like spread has also been observed in culture systems (103, 104),
where SOD1 aggregates can be released to the extracellular space (105) and
transferred from cell to cell in a prion-like spread manner. Additionally, the
propagation appears not to be limited between neurons, but could include
other cell types like microglia (106) and astrocytes (107). Recently, infectivity of ALS between transgenic mice individuals has been reported (108,
109), further suggesting a prion-like spread. Spinal cord homogenates from
symptomatic ALS-mice was injected into the spinal cord of healthy hSOD1
20
transgenic mice that subsequently, in a short time, developed the disease
(108, 109) and SOD1 aggregation was observed to propagate through the
spinal cord from the site of injection. Similarly, pathological changes in ALS
patients first occur at the site of disease onset and then propagate throughout
affected tissues in a time-dependent manner (110, 111), indicating that seeded aggregation may be the key factor of disease progression.
Another typical hallmark of prion diseases is the existence of different
strains (112), which arise from the same protein but can adopt different conformations. Accordingly, SOD1 aggregate formation is a malleable process
that can give rise to different aggregate structures depending on point mutation and the conditions at which aggregation is induced (113, 114). Additionally, in vitro studies show that several distinct pathways for SOD1 aggregation are possible (115) and they are expected to have differences in
their propensity to seed aggregation. This offers an explanation for the heterologous progression and severity of fALS cases with variable phenotypes
for different SOD1 mutations (116). However, despite the number of parallels between the neurodegenerative diseases and the prion diseases they appear to differ when it comes to transmission between individuals. So far, at
least under natural circumstances, ALS is not an infectious disease(117).
Aggregation pathways
The amyloid state is not a rare phenomenon associated with a small number
of disease-related proteins but rather a generic property of a polypeptide
chain (70, 118-120) and an alternative to the native and denatured states.
Amyloid comprises a set of highly ordered aggregates, often of great interstructural diversity. In amyloid, the polypeptide chains adopt a fibrillar
structure that can act as a template or seed to incorporate new monomeric
proteins and thereby self-replicate. The fibrillar structure can also catalyse
the formation of new fibrils by providing a surface upon which new fibrils
can nucleate and grow or by breaking into smaller fibrillar fragments.
The aggregation process of amyloidogenic proteins in vitro is often studied
by optical spectroscopic methods using the amyloid binding dye thioflavin T
(ThT) as a probe. The fibrillation kinetics in vitro, where there is a limited
pool of monomers, typically displays a sigmoidal shape (Fig. 6), which can
be described by the empirical standard expression (121)
(Eq. 5)
21
where F0 is the baseline before fibrillation, A the amplitude and τ1/2 the time
for depletion of half of the monomer pool. From the fitted parameters τlag
can be calculated (121). The lag time (τlag) is defined as the intercept of the
tangent from the midpoint of the sigmoidal curve with slope vmax and the
time axis (Fig. 6).
τlag = τ1/2 - 2/vmax
(Eq. 6)
The aggregation process is complex with multiple inter-conversions; nucleation, growth, dissociation and fragmentation taking place simultaneously.
Although the empirical model above describes the aggregation reaction in
terms of apparent growth rates and lag-phases, these properties, on their
own, do not give any information about the underlying microscopic events.
A key question in protein aggregation as well as other types of molecular
assembly is to determine the importance of the different microscopic processes and their contribution to the overall reaction.
Fig. 6. The in vitro sigmoidal fibrillation profile. The fitted parameters are the maximum growth
rate (vmax), red dotted line, and the aggregation half time (τ1/2), at the point of inflection. The lag
time (τlag) is defined as the intercept of the tangent from the midpoint of the sigmoidal curve with
slope vmax and the time axis.
According to a model developed by Knowles and co-workers (122-124) the
aggregation process can be divided into two steps: nucleation, when the aggregate is formed, and elongation, when the aggregate grows by monomer
addition to the fiber ends (Fig. 7). The number of free fiber ends where
monomers can attach is increased through: primary pathways – primary nucleation (when the aggregate is formed) and secondary pathways (122-124)
(Fig. 7). The secondary pathways are divided into: surface catalysed nucleation, where nucleation of a new aggregate starts on the surface of a preformed fibril, subsequently forming a fiber of its own, and fibril fragmentation, where breakage of fibrils creates new fibril ends (122-124) (Fig. 7).
22
Primary pathways describe the formation of new fibrils from interaction
between monomers only and is featured by the rate constant kn and the reaction order nc (Fig 7). Where nc in the simplest case relates to the number of
monomers in the nuclei. Fibril elongation describes the association of monomers to the fibril ends with the rate constant k+. Secondary pathways, however, depend on the behaviour of already formed aggregates and can be divided into monomer dependent and monomer independent pathways (Fig 7).
Surface catalysed nucleation is monomer dependent relying on both monomeric protein and existing fibrils. Here the rate constant is k2 and the reaction
order with respect to the monomeric protein is n2 (Fig. 7).
Fig. 7. Schematic illustration of the microscopic processes of aggregation (122-124).
Fibril fragmentation on the contrary is monomer independent and increases
the number of free fibril ends at a rate depending only on the level of existing aggregates with the rate constant k- (Fig. 7). Naturally, all aggregation
reactions have to involve a primary nucleation event. However, when information about primary nucleation is unavailable the question is whether there
are secondary pathways present and which reaction that dominates the observed kinetics.
The aggregation half time, τ1/2, is generally dependent on the total monomer
concentration m(0), which can be estimated by a power law (122)
1/2 m(0)
(Eq. 7)
The half-time exponent (γ) determines the monomer dependence and can be
used to extract the reaction order, nc or n2, for a system dominated by primary or secondary pathways respectively (Fig. 7). Equation 7 becomes a linear
relationship with the slope γ when the half time is plotted versus the protein
23
concentration in a log-log plot. In the case where there are saturation effects
as shown by curvatures in the double-logarithmic plot the values of the different slopes can give insight into which mechanisms that dominate at different protein concentrations (125).
In practice secondary pathways for aggregate formation have turned out to
dominate the observed aggregation kinetics. When secondary pathways are
dominating fibril formation, the initial aggregation kinetics has been shown
to be of exponential nature (Fig. 10) with the general expression for the rate
constant κ (123)
(Eq. 8)
Where M(t) is the mass concentration of fibrils and M(t)/ctotal is the normalized fibril mass fraction and Λ is an arbitrary constant.
The lag time, τlag, is related to the initial kinetics by (123)
(Eq. 9)
In this case the kinetic parameter, λ, describes the formation of aggregates
through primary nucleation whereas κ is the kinetic parameter through secondary nucleation and/or fragmentation. These kinetic parameters are given
by
2k kn (Eq. 10)
Where k+ is the elongation rate, kn the primary nucleation rate and nc is the
primary nucleation reaction order.
2k k
(Eq. 11)
where KM is the Michaelis equilibrium constant and n2 the secondary nucleation reaction order. However, when no saturation of secondary nucleation is
detected and the dominating pathway is fibril fragmentation n2 = 0, the expression for κ reduces to
2k k
24
(Eq. 12)
where k_ is the fragmentation-rate constant. The half-time exponent γ determines the dependence of on the concentration of monomer, where γ = 0.5
for completely fragmentation-based aggregation and γ = 1.5 for secondary
nucleation-driven aggregation, without saturation effects.
25
Results and Discussion
Understanding the SOD1 aggregation process from initiation to fiber elongation and aggregate multiplication is vital for a thorough interpretation of the
role of SOD1 inclusions in ALS. The overall aim of this study is to characterise the SOD1 fibrillation pathway starting with identification of the aggregating precursor state and the aggregation mechanism, continuing with
investigation of the structural properties, i.e. which parts of the SOD1 sequence that are incorporated in the aggregates and what sequence determinants for aggregation there are. Importantly, the in vitro situation was
throughout the study compared with the complex environment of the live
cell. To understand how SOD1 molecules interact and form aggregates inside live cells is important when trying to shed light on the mechanisms behind ALS. Learning more about these processes both in vitro and in vivo can
potentially aid therapeutic intervention not only for ALS but also for other
protein misfolding diseases, as they appear to share many features.
The aggregating state of SOD1
Despite the increasing number of acknowledged amyloidogenic proteins, the
identity of the nucleating species and how the aggregates assemble remain
elusive. Data for different amyloid forming proteins and peptides point to
several potential precursors for aggregation. Many systems, like the intrinsically disordered proteins are largely unstructured prior to the aggregation
process whereas globular proteins generally need to unfold partially (71-73)
or globally (126, 127), and expose aggregation prone segments to assemble
into amyloid fibrils. Formation of fibrils can also be preceded by an oligomerisation step in which the native state can be retained and reshaped into βstructure only later in the process (128, 129). Native-like conformations can
even be maintained in the fibrils themselves under some conditions (130,
131) (Fig 5.).
The complex maturation and degradation pathways of native SOD1 include
many intermediate states that could potentially lead to aggregation. As described above, SOD1 is subjected to various post-translational modifications
before it gains its dimeric mature conformation. After synthesis, the unfolded nascent chain released into the reducing environment of the cytosol is a
26
high-risk candidate for aggregation. Like other globular proteins, SOD1
contains hydrophobic sequence elements that are buried in the native state
but could associate in a non-native fashion when the protein unfolds (60,
114). Also, the SOD1 sequence contains four cysteine residues that could
fuel aggregation by non-native disulphide bond formation (90). Finally, the
loop stretches (Fig. 2 and 3) that coordinate the structural Zn2+ and catalytic
Cu2+ in the functional state are unstructured in the in absence of metals and
could potentially interact and entangle leading to aggregation (132, 133).
Destabilisation (27), decreased net charge (26) and changed metal affinity
(23, 54) due to ALS-associated mutations and environmental factors of the
cell further affect the intracellular ratios of immature aggregation competent
species.
Fig. 8. Crystal structures and fibril morphologies of apoSOD1 variants. The apoSOD1
SH SH
with an intact disulphide link, the disulphide reduced apoSOD1
and the loop IV (green) and
barrel
VII (blue) depleted apoSOD1
. Structures were constructed from pdb entry 2XJK. EM micrographs of the respective protein fibrils formed in pure buffer at pH 6.3 and 37 °C with mechanical agitation. Adapted from Paper I.
C57-C146
Interestingly, electron microscopy pictures of fibrils formed in vitro of these
immature species reveals that both the disulphide reduced apoSOD1 as well
as the engineered model protein apoSOD1barrel, with the metal coordinating
loops IV and VII removed (Paper i (134)), form fibrils that are undistinguishable from those of the apoSOD1 with the disulphide bond intact (Fig.
8, Paper I). Hence, only the hydrophobic parts of the β-barrel are needed to
form the fibrillar core (Paper I) and neither cysteines nor loops are required
for the protein to form fibrils. The question is then if the β-barrel of SOD1
needs to unfold or partially break up in order to aggregate or if the native
conformation itself can form fibrils. As discussed earlier, proteins are dynamic entities in constant equilibrium between different states. Gradual de-
27
stabilisation of monomeric apoSOD1 in urea, i.e. increasing the concentration of denatured states (D) (Scheme 1, Eq. 3, 4), while determining the aggregation rates reveals that the more destabilised the protein is the faster it
aggregates. The increased rate of aggregation is shown by short lag times
(τlag) and half times (τ1/2) accompanied by steep growth rates (vmax) (Fig. 9,
Paper I). Interestingly, aggregation rates reach a maximum at full occupancy
of denatured protein just after the transition midpoint between N and D (Fig.
9). The same result is seen for the full length monomeric SOD1 although the
available range of globally unfolded protein is wider for SOD1barrel (Paper I).
Fig. 9. Folding and fibrillation data. A. Chevron plot of apoSOD1
showing linear fits of log ku
and log kf (Eq. 3). The vertical line at 4 M urea indicates the urea concentration where full
occupancy of denatured protein, [D], is reached. B. Concentration of denatured protein [D] in
barrel
the fibrillation assays vs. [urea], calculated from the chevron data of apoSOD1
. Total protein
barrel
concentration is 25 μM and units are in M. C. Fibrillation half times (τ1/2) of apoSOD1
vs.
[urea]. D. Corresponding data for the maximum growth rate (vmax). E. Illustration of the aggregation pathway of SOD1 showing that the aggregation starts from the globally unfolded protein.
EM micrograph of SOD1 fibers. Adapted from Paper I.
barrel
SOD1 can be destabilised either by chemical denaturation in urea, which
yields a wider concentration range of D than when the concentration is varied by mass (Fig. 9, Paper I), or by amino-acid substitutions shifting the
thermodynamic equilibrium between N and D (Fig.12, Paper II). In all cases,
apoSOD1 aggregation correlates with the concentration of globally unfolded
monomer, [D], hence D is the most probable aggregating state. Additionally,
28
equilibrium unfolding of apoSOD1 measured by NMR 1H-{15N}-HSQC
peak volumes show no indications of populated intermediates (Paper I).
A globally unfolded precursor for aggregation is compatible with many intrinsically disordered proteins and peptides that are constantly aggregation
competent as they lack the ability to bury hydrophobic sequence regions by
folding (e.g. amyloid-β peptide, α-synuclein and PolyQ). On the contrary,
many globular proteins like SOD1 need to be destabilised to aggregate. One
such example is the human plasma protein transthyretin (TTR) that recently
was shown to aggregate from unfolded precursors rather than from a partially unfolded species (135). Fibrillation from globally unfolded material has
also been shown at low pH for the WW domain (127) and β2-microglobulin
and for apomyoglobin that converts to fibrils under solution conditions correlating with the degree of denaturation (126). Additionally, recent data from
our group shows that an engineered sequence extended variant of SOD1 with
the possibility to populate two different folded structures has to completely
unfold in order to swap between these conformations (Paper vi (75)). There
is no sign of partial unfolding on the native side of the barrier, which further
confirms the need for SOD1 to globally unfold before assembling into an
alternative structure. Also, in support of the globally unfolded precursor is
the observation that ALS mutations preferentially reduce SOD1 stability,
implicating the globally unfolded state as the disease precursor (27).
The aggregation mechanism of SOD1
Aggregation is difficult to measure in the complex context of live tissue and
has therefore been studied extensively in vitro (136-141), which has led to
the development of methods for quantitative understanding of protein aggregation data (see Aggregation pathways) (122-124). Although measured at
the simplified in vitro conditions there is still great heterogeneity and complexity that obscure the molecular details of the self-assembly process. Additionally, the high variability requires the accumulation of a great number of
experiments for each examined condition or mutation to obtain good enough
statistics to optimize the precision of the fitted parameters (Paper I, II).
SOD1 is a soluble protein and fibrillation in vitro relies on mechanical agitation i.e. shaking or stirring the protein sample is necessary for aggregation to
proceed (Paper I). The aggregates grow exponentially over time, shown by
the in vitro sigmoidal fibrillation trace (Fig. 6). The sigmoidal shape of the
trace in vitro is due to the depletion of the monomer pool. The initial aggregation trace can be fitted to an exponential equation (Eq. 8, Fig. 10), which
would continue to grow exponentially if unlimited monomer was accessible.
29
Fig. 10. The in vitro sigmoidal fibrillation profile of SOD1
barrel
at 0 M urea fitted to a sigmoidal
and an exponential equation for comparison. The sigmoidal shape of the trace in vitro is due to
the depletion of the monomer pool (black). If the number of accessible monomers was unlimited
the trace would continue to grow exponentially (red).
The direct correlation between the parameters for SOD1 aggregation, log
τ1/2, log τlag and log vmax, with a slope of one (Paper I and II, Fig. 11C, 12D),
indicates that a secondary process is dominating, such as secondary nucleation on fibril surfaces or fragmentation, according to the model developed by
Knowles and co-workers (122-124). These parameters also display overall
linear dependencies on log [D] agreeing with observations for other proteins
where the concentration is varied by mass (Fig. 11B) (124, 136, 139, 142).
Interestingly the monomer concentration dependence of the parameters is
remarkably weak with slopes (γ) of approximately 0.5 (Paper I, Fig. 11B).
This does not make sense for a mechanism dominated by primary nucleation
where the formation of the nuclei is slower than the elongation (143), indicating a nucleation number of 0.5 SOD1 monomers. This weak concentration dependence rather suggests domination of secondary pathways for fibrillation, as also shown above (122-124, 143). From the weak monomer
concentration dependence, with γ = 0.5, together with the fact that agitation
is needed for SOD1 to aggregate we conclude that fragmentation is the dominating mechanism for SOD1 fiber mass growth. Hence, the observed lag
phase is just an exponential increase of fibril ends that can recruit unfolded
monomer and it contains no information about the initial nucleation event
(Fig. 10).
30
Fig. 11. Chemical denaturation with urea reveals unfolded monomers as aggregation precursors and fibril fragmentation as dominant mechanism. A. Schematic illustration of how urea
shifts the thermodynamic equilibrium between N and D. B. Correlation of aggregation rate (log
vmax and log τ1/2 ) with concentration of globally unfolded monomer (log [D]) for urea titrated
barrel
SOD1
showing γ = 0.48 and γ = -0.46, consistent with aggregate growth dominated by fragmentation. C. Log plot of vmax and τ1/2 for all individual measurements showing uniform behaviour and a slope of one characteristic of exponential growth. Adapted from Paper II.
The model developed by Knowles and co-workers (124) uses a general expression for the aggregation rate constant, κ, in the limit of secondary nucleation/fragmentation (see Aggregation pathways). Applying this model to our
SOD1 fibrillation data yields the monomer concentration dependence of κ, γ
= 0.5 (Fig. 12B) indicative of a fragmentation-assisted growth of aggregates.
The concentration dependence is the same for SOD1 destabilised by ALS
related as well as artificial mutations (Fig. 12). Fragmentation as a consequence of stirring, shaking or sonication is observed in vitro for a variety of
proteins such as β2-microglobulin (139), the yeast prions Ure2p (144) and
Sup35 (136), insulin (145), the WW domain (137), TI 127 (138) and αsynuclein (146). Fragmentation thus appears to be independent of the aminoacid sequence per se and a common feature of all protein fibers.
31
Fig. 12. SOD1 mutations reveal unfolded monomers as aggregation precursors and fibril fragmentation as dominant mechanism. A. Schematic illustration of how mutations can shift the
thermodynamic equilibrium between N and D. B. Correlation of aggregation rate (log κ) with
concentration of globally unfolded monomer (log [D]) for the mutations with γ = 0.49, consistent
with aggregate growth dominated by fragmentation. C. Distribution of the analysed mutations in
SH SH
the apoSOD1 monomer (SOD1
pdbID 2XJK) and the loop-truncated SOD1 barrel
barrel
(SOD1
pdbID 4BCZ). Colours are the same as in B. D. Log plot of vmax and τ1/2 for all individual measurements showing uniform behaviour and a slope of one characteristic of exponential growth. ALS mutations do not deviate from the general trend. Colours are the same as in B.
Adapted from Paper II.
SOD1 aggregation in vivo
Protein aggregation, like protein folding, follows the rules set by the aminoacid sequence. Aggregation however, can start from a vast number of different aggregation sites within the protein (147), which makes the process
adaptive and sensitive to experimental conditions. An important concern is
how the model for the already variable in vitro data translates to the even
more complex conditions in vivo. Comparison of SOD1 aggregation kinetics
in vitro with corresponding data from ALS-transgenic mice indicates that the
differences are smaller than expected (Paper II). Strikingly, disease progression of transgenic mice expressing human SOD1 (hSOD1) variants correlates linearly with the concentration of globally unfolded protein in the mice
tissue, [D]/Ctot (Fig. 13, Paper II). In agreement with the previously discussed test tube data (Fig. 9) this correlation highlights [D] as the aggregating state of SOD1 in the mice tissue and suggests that monomeric apoSOD1
stability determines mouse disease progression. Additionally, it indicates
that the aggregation mechanism is the same for all examined transgenic
32
mouse models regardless of the hSOD1 mutation that they express: the difference in survival time depends only on changes in concentration of globally unfolded monomer, [D].
Fig. 13. Mice survival correlates with concentrations of denatured SOD1 in ALS-transgenic mice
tissue. A. Immuohistochemistry of the ventral horn in the terminal hSOD1G93A mouse (Paper
II). B. Plot of mouse survival times v. normalised tissue [D] calculated from in vitro-determined
pwt
C6A, C111A, F50E, G51E
SOD1 (SOD1
) stability and expression levels in transgenic mice (Paper II).
The amount of pathological aggregates in live tissue is generally too low for
structural and kinetic analysis, but with a recently developed antibody mapping method the build-up of SOD1 aggregates in small samples can be monitored (148). Application of this method to tissue samples from ALStransgenic mice lines (hSOD1D90A, hSOD1G93A, hSOD1G85R and hSOD1wt)
reveals an exponential increase of SOD1 aggregates closely resembling the
behaviour from the test tube (Fig. 14 and Paper II). What differs is the sigmoidal shape of the in-vitro data reaching a plateau as the monomers are
depleted whereas the in-vivo data remains exponential due to continuous
supply of new synthesised monomer. Intriguingly, the in-vivo aggregation
growth rate (κ) calculated from the aggregation doubling times in the ALStransgenic mice tissue, depends weakly on [D] (Fig. 14) corresponding to the
fragmentation-controlled growth observed in vitro with a γ of 0.5 (Fig. 12B).
The fibrillation process appears thus to be robust and to obey the same simple rules both in vivo and in vitro; fragmentation-assisted growth seems generic to the fibrils themselves, and independent of the environmental conditions or precise structural composition of the protein.
33
Fig. 14. SOD1 aggregation in ALS-transgenic mice. A. Single exponential growth of strain A
G93A
aggregates in hSOD1
mice at two different expression levels 100% (black) and 50% (red).
D90A
hSOD1
mice have both strain A and B growth occurring in parallel thereby displaying a
second exponential phase (blue). B. Log plots of the data in A showing the common level of
aggregates in terminal mice (blue shading). C. Log plot of the aggregation rate constant (κ) vs.
G93A
normalised tissue [D] for hSOD1
with expression levels 50 % (red) and 100 % (black) yieldin vivo
D90A
ing a slope of γ
=0.52±0.05. For comparison the hSOD1
data fitted single exponentially
G93A
(blue) and the hSOD1
mice at 12 % expression loss (grey) were included (Paper II).
Regarding the toxicity mechanism, the most reductionist interpretation of
these data would be that the exponential growth of intracellular SOD1 aggregates in ALS-transgenic mice eventually overloads the cells that can no
longer function normally. Although the aggregates themselves can be harmless the aggregate load becomes too much. Indeed the aggregate load at time
of death is similar for all ALS transgenic mice individuals regardless of
SOD1 expression levels and mutation (Fig. 14B, Paper II). The aggregate
levels in terminal mice become even so high that the amount of monomers
required to sustain exponential growth exceeds the SOD1 expression capacity of the motor neurons. Thus, for ALS-transgenic mice, there is no need to
address toxicity to aggregates or elusive oligomeric SOD1 species; neuronal
damage is simply the effect of aggregate load. The human pathology, however, is more variable with stochastic disease progressions compared to the
transgenic mice. An explanation for the variability in human ALS might be
the coexistence of strains of structurally different SOD1 aggregates, as discussed in the following section.
Sequence determinants for SOD1 aggregation
Hydrophobicity is an important property for the aggregation propensity of a
polypeptide chain (80, 81, 149) and there is evidence that protein sequences
have evolved to avoid clusters of hydrophobic residues (81, 150) Another
significant feature is charge, as high net charge either globally or locally may
hinder self-association through electrostatic repulsion (81, 151, 152). Introducing negatively charged residues successively in the eight β-strands of
34
SOD1 in search of regions important for aggregation surprisingly shows an
overall uniform effect on aggregation rates. Interestingly, the response to
charge substitutions is not restricted to a few aggregation hot spots, instead
the aggregation rates of the different charge constructs correlate with their
global charge (Fig. 15A, Paper I).
Fig. 15. Correlation between SOD1 fibrillation lag time (τlag) and net charge. A. Fibrillation lag
time (τlag) correlates linearly with global net charge. B. Scheme describing the multistep process
of protein association through the formation of a transient encounter complex that subsequently
rearranges into the final complex. The transient complex is dominated by long-range electrostatic interactions thereby affected by substitution of charged side chains. Adapted from Paper I.
Complex formation of proteins can be described as multistep processes that
starts by random collisions, leading to the formation of loosely bound encounter complex that subsequently rearranges into the closely packed final
complex (153) (Fig.15). The encounter complex is typically dominated by
long-range electrostatic interactions whereas close-range contacts are less
important. Consequently, if the aggregation kinetics is rate limited by the
formation of an encounter complex the effect of mutation on association
kinetics is largest for substitutions of charged side chains and negligible for
truncation of hydrophobic side chains. The correlation of SOD1 aggregation
rates with global charge suggests that there is a loosely bound complex involved, which is determined by the net negative charge mainly affecting k+
in the expression for κ (Eq. 12). Notably, none of the tested charged β-strand
substitutions increase the aggregation rate, as would be expected if fibril
fragmentation was enhanced and k- was increased (Eq. 12). There is also a
basic correlation of aggregation rate with decreased hydrophobicity (Paper
I). Even so, point truncations of hydrophobic side chains appear to affect
aggregation mainly in the sense that they destabilise the protein and shift the
thermodynamic equilibrium towards D (Paper II) as detailed above in the
section “the aggregating state of SOD1” and “the aggregation mechanism of
SOD1” (Fig. 12B).
35
The aforementioned antibody mapping method developed to recognise unstructured SOD1 include 8 antibodies that specifically recognise peptide
regions covering the whole SOD1 sequence (148). These where used to investigate what parts of the SOD1 sequence that are incorporated into fibrils
from both test tube and ALS transgenic mice. Sequence regions that are unstructured and accessible to the antibodies are thus protruding from the aggregates whereas hidden parts are incorporated in the aggregates and give no
signal. Despite the amorphous look of the SOD1 deposits in live cells the
antibody-binding pattern of the purified and washed material suggests an
ordered fibrillar structure resembling that of the in-vitro produced fibrils
(Fig. 16, Paper II, III). Furthermore, antibody analysis of hSOD1 aggregate
structures in four different well-characterised ALS transgenic mouse lines
(hSOD1D90A, hSOD1G93A, hSOD1G85R and hSOD1wt) reveals two different
aggregate strains, A and B (Fig. 16, Paper III).
Fig. 16. SOD1 aggregation forms two structural strains (A and B) in spinal cord of transgenic
mice, which both differ from SOD1 fibrils formed in vitro. A. Epitope mapping pattern of hSOD1
aggregates from terminally ill mice from four different transgenic ALS models. B. Epitope mapping pattern of hSOD1 aggregates produced in vitro at different conditions. C. Model-free PCA
analysis of antibody binding data demonstrating two different strains of in vivo aggregates (A
and B) that differ from the hSOD1 fibrils produced in vitro (Paper III).
Solid-state NMR (ssNMR) (Paper v (154)) and limited proteolysis (113,
114) of SOD1 in vitro aggregates have previously shown that they contain
both rigid and mobile parts. This is confirmed by the antibody-binding pattern of aggregates from spinal cord of transgenic mice. In strain A, the first
three to four β-strands together with β-stand five to seven are hidden from
antibody binding whereas loop IV and the C-terminus remain flexible (Fig.
16). Strain B differs from strain A, burying only the first three to four βstrands in the aggregate while exposing the rest of the protein (Fig. 16).
36
Strain A aggregates were detected in all four mice lines, whereas strain B,
which is distinctly different from A, was only present in the D90A mouse
line. The distinct strain pattern found in transgenic mice is contrasting from
the antibody binding pattern for aggregate structures formed under in vitro
conditions (Fig. 16) as well as the variable in vitro aggregate structures detected with ssNMR (154) and limited proteolysis (113, 114). This suggests
that factors of the CNS modulate the aggregation process in vivo. Such shaping of the in vivo aggregation could stem from chaperone interactions (155,
156), crowding by surrounding macromolecules or interactions with intracellular membranes (157).
The ratio of co-existing strain A and B aggregates in D90A transgenic mice
influences their survival time: the more strain B the earlier disease onset and
faster disease progression (Fig. 17, Paper III). Additionally, strain B aggregates are more fragile than strain A aggregates when exposed to GdmCl and
mechanical stress (Fig. 17C). This explains the more aggressive growth pattern of strain B and agrees with the positive correlation between aggregate
fragility and neuroinvasiveness observed for several other disease models
(112, 158-162).
Fig. 17. Correlation between disease progression of transgenic mice and their strain A and B
levels. A. Binding pattern of the eight antibodies, covering the SOD1 sequence, to aggregates
D90A
from the youngest (red) and the oldest (blue) terminally ill hSOD1
mouse. B. Coexisting
D90A
D90A
stain A and B plotted vs. lifespan of hSOD1
mice. C. Fragility of hSOD1
aggregates from
strain A and B when exposed to chemical-mechanical titration by increasing concentration of
GdmCl and sonication. Adapted from Paper III.
Interestingly, no secondary nucleation on fibril surfaces (i.e. γ = 1.5 as described in “Aggregation pathways”) is detected for SOD1 fibrillation (Fig 11
and 12). Possibly, as indicated by the antibody-binding pattern, parts of the
SOD1 sequence that protrude from the fiber obstruct free monomers from
accessing the fiber surface and thereby inhibit nucleation on the surface. For
amyloid β peptide (Aβ) in contrast, secondary nucleation on fibril surfaces is
the dominating mechanism (142, 161) and most of the peptide is incorpo37
rated into the fiber, thereby creating a more ordered surface that can serve as
site for nucleation.
Destabilisation of SOD1 in vivo
Although proteins have evolved to function in the complex environment of
the live cell, protein thermodynamics has so far been studied almost exclusively at simplified conditions in vitro. Recent methodological advances,
however, make it possible to study protein properties in vivo with in-cell
NMR (163, 164), where isotope labelled protein is delivered into mammalian cells. This technique can be used to study protein structures, interactions
and dynamics under innate physiological conditions (165, 166). SOD1barrel
displays fully resolved NMR spectra in mammalian cells when delivered
either by conjugation of a small signal peptide to the protein facilitating the
passage through the membrane (Paper iv (167)) or by disrupting the membrane using electroporation (Paper IV).
Fig. 18. In-cell quantification of protein stability. A. In-cell HMQC spectra of SOD1
at two
different temperatures showing that the protein is mainly folded at 17 °C (red) and unfolded at
37 °C (blue). B. ΔGD-N vs. temperature determined in mammalian (blue), bacterial (green) and in
I35A
vitro in PBS (black) of SOD1 . The D/N equilibrium was quantified from the Q153 cross-peak
volume. Destabilisation shifts both cold unfolding (Tc) and melting temperature (Tm) into the
physiological temperature region. Adapted from Paper IV.
I35A
In-cell NMR was used to explore how the thermodynamic stability of SOD1
is affected by the crowded interior of live cells (Paper IV). To avoid disulphide cross-linking and complications from unstructured loops, the SOD1
variant delivered into the cells by electroporation was the engineered
SOD1barrel, lacking cysteine and metal binding residues and the long flexible
loops IV, VII (Paper i (134)) (Fig 8). SOD1barrel with the substitution I35A is
a good in-vitro model for SOD1 as it has a thermodynamic stability close to
that of the full-length disulphide reduced apo monomeric SOD1 in vitro
(Paper IV). The in vitro-NMR spectrum of I35A SOD1barrel at 37 °C shows
mixed populations of D and N quantified by the C-terminal Q153 cross
peaks (Paper IV), making it useful as a probe for studies of perturbations to
38
the system. Live cell measurements, however, reveal that the thermodynamic
equilibrium of SOD1 is shifted towards the globally unfolded monomer (Fig.
18, Paper IV).
For comparison, in-cell data for other proteins show that the cellular interior
can act both stabilising (168-173) and destabilising (163, 174-179). A destabilisation is contradictory to the anticipated stabilising effect of steric crowding in vitro (180-182) and suggests that there are attractive interactions between SOD1 and the cellular interior. Since NMR chemical shifts and line
broadening indicate that N inside live cells remains unchanged and free of
specific interactions, it can be concluded that D is modulated inside the cell
as shown by the increased ΔCp compared to in-vitro conditions (Paper IV).
Changes in ΔCp depend generally on the altered hydrophobic hydration and
can be used as a measure of the increase in solvent-accessible surface area
upon the transition between D and N. Accordingly, for SOD1 the surface
area of intracellular D seems increased and more expanded (Paper IV) compared to the structures adopted in pure water (183). This effect can be assigned to transient interactions with the cellular interior that counterbalance
the crowding pressure. These interactions more likely rely on the protein
sequence and the context as revealed by crowding experiments with different
cosolutes (Fig. 19, Paper IV). Depending on the identity of the protein and
the environment the effect of surrounding proteins can thus be variable and
depend on both surface characteristics of surrounding molecules and the
features of the target protein itself. Moreover, recent data from our group
show that in-cell mobility of N respond predictably on surface characteristics
depending on the net-charge density, surface hydrophobicity and the electric
dipole moment of the protein (Paper vii (manuscript)).
Fig. 19. Comparison of in vitro and in vivo conditions. Osmolytes influences SOD1
stability
opposite to live cells whereas protein crowders yield different effects indicating the importance
of the amino-acid sequence of the protein solute interactions (Paper IV).
I35A
39
Summary and Concluding remark
Although this thesis work does not solve the mysteries of SOD1 aggregation
or the disease mechanism of ALS, it provides important clues to our understanding of the mechanisms and determinants for SOD1 aggregation both in
vivo and in vitro.
The most important findings of this thesis work are listed and later discussed:
1.
SOD1 fibrillation starts from the globally unfolded state.
2.
Fibril fragmentation is the dominating mechanism for SOD1 aggregate
multiplication.
3.
SOD1 fibrillation in vivo shares the features of in vitro-fibrillation with
globally unfolded precursors and fragmentation-assisted growth.
4.
SOD1 in vivo aggregates are structurally different from those produced
in vitro suggesting that the CNS environment shapes the aggregation
process and gives rise to different aggregate stains dependent on mutation.
5.
In the live cell interior, as measured by in-cell NMR, the unfolded
aggregation competent state of SOD1 is more abundant compared to
the in vitro conditions.
In Paper I we show that neither the long flexible loops IV and VII nor disulphide links are needed for SOD1 to aggregate and form fibrils. The hydrophobic parts of the β-barrel merely have to rupture and globally unfold for
the protein to aggregate. Aggregation from the globally unfolded protein is
shown both for the chemically denatured SOD1 (Paper I) and SOD1 destabilized by mutation (Paper II). Even so, fibrillation is critically dependent on
mechanical agitation to initiate and proceed. The fibrillation mechanism is
closely connected to the sample agitation, which enhances the shearing of
the fibers and thereby increases the number of fiber ends where free monomer can attach.
40
Interestingly the data from ALS-transgenic mice supports the features of in
vitro-fibrillation. The close agreement between in vitro fibrillation and
mouse pathology suggests that both processes are controlled by the concentration of unfolded protein [D] (Paper I and II). The accelerating effect of
mutations appears to be simply caused by an increase in [D] due to decreased protein stability. Additionally, the destabilisation of SOD1barrel inside
the cell as measured by in-cell NMR (Fig. 18, Paper IV) implicates that the
aggregation competent SOD1 is even more abundant in vivo. The protein
stability data used in Paper II is based on the reduced monomeric SOD1pwt
described in “SOD1 folding and stability”. When we instead estimate the
stability according to the more correct reduced hSOD1wt when calculating
tissue concentration of globally unfolded protein ([D]/ctot) the correlation
with the survival time of the transgenic mice (Fig. 13) still holds but is shifted towards lower [D]/ctot (Fig. 20A). However, when taking into account
the SOD1 destabilisation in live cells as measured by in-cell NMR (Fig. 18,
Paper IV) the correlation is improved and shifted to higher [D]/ctot (Fig.
20B).
Fig. 20. Mice survival correlates with concentrations of denatured SOD1 in ALS-transgenic mice
tissue. A. Plot of mouse survival times v. normalised tissue [D] calculated from in vitrowt
determined hSOD1 stability and expression levels in transgenic mice. B. The same data as in
A. when taking into account the SOD1 destabilisation measured by in-cell NMR.
As for he log plot of the aggregation rate constant (κ) vs. normalised tissue
[D] (Fig. 14C) when estimated to hSOD1wt stability the correlation is unchanged with γ = 0.5 although without the support of the hSOD1D90A, which
is shifted to lower [D]/ctot (Fig. 21A). The deviation of this point can stem
from the coexistence of strain A and B aggregates in the hSOD1D90A mouse
line, the estimations used in our calculations of the protein stabilities and/or
the estimation of expression levels in the mice. When taking into account the
in-cell destabilisation measured by in-cell NMR the points shift to higher
[D]/ctot although the hSOD1D90A is still somewhat deviating (Fig. 21B).
41
Fig. 21. SOD1 aggregation in ALS-transgenic mice tissue. A. Log plot of the aggregation rate
WT
G93A
constant (κ) vs. normalised tissue [D] estimated to hSOD1 stability. hSOD1
with expresin vivo
sion levels 50 % (red) and 100 % (black) yielding a slope of γ
=0.52±0.05. For comparison
D90A
G93A
data fitted single exponentially (blue) and the hSOD1
mice at 12 % expresthe hSOD1
sion loss (grey) were included. B. The same data as in A. when taking into account the SOD1
destabilisation in live cells measured by in-cell NMR.
Importantly, this result is based on a number of estimations and needs to be
further investigated taking into account the many intermediate states of the
complex maturation pathway of the dimeric, metallated and disulphide intact
SOD1 and the steady state between the different species. The effective equilibrium between native and denatured SOD1 in the cell also includes the apo
SOD1 dimer as well as CCS-bound SOD1. Additionally, ALS mutations
affect the posttranslational modifications differently, influencing the dimer
affinity (23, 54), metal binding affinity (52, 53) and the disulphide status to
varying degrees (55-57). Furthermore, the cells used in the in-cell NMR
experiment are not motor neurons and can therefor influence the protein
differently. Future studies by in-cell NMR, also in other cell lines, can evaluate how the thermodynamic stability is affected for the different ALS mutations in live cells depending on the extent of posttranslational modification.
Currently our group is looking into how the full length SOD1 monomer
without cysteine and metal binding residues is behaving inside live cells.
The antibody analysis of aggregates from ALS transgenic mice reveals the
existence of aggregate strains involving different parts of the protein depending on mutation (Paper III), which may offer an explanation for the various
disease phenotypes observed in ALS. Additionally, our results show that
SOD1 in vitro-produced fibrils differ from the fibrils found in ALS transgenic mice suggesting that the CNS influences the aggregation process in vivo.
The in vitro fibers are more heterogeneous as seen from the antibodybinding pattern (Paper III), containing a great variation of structurally different aggregates. There are many potential causes for this where the most ob-
42
vious is the presence of unknown important factors in the live tissue that
influences the aggregate formation. To look further into this our collaboration partners at Umeå University are currently performing experiments
where in vitro produced SOD1 aggregates examined with the antibody assay
(Paper III) are injected into spinal cord of transgenic mice. This result will
be compared with previous studies of injected homogenates from mice (109)
in terms of aggregate spread, disease onset and strain development in the
mice tissue.
Additional deviations of the in vitro-system compared to live tissue are the
purity of the in vitro protein sample, the conditions of the aggregation assay
exposing the protein to air water interfaces and the vessel wall that can provide a surface for seeded growth. This can also explain the difficulty in reproducible seeding of SOD1 in vitro giving rise to ambiguous results (unpublished). Finding a way to reproducibly seed the SOD1 fibrillation can
also provide more detail on the aggregation mechanism at different conditions providing a means to study fiber elongation exclusively. Accordingly,
in future studies more efforts can be focused on refining the in vitro fibrillation assay. Additionally, seeded aggregation can be a way to amplify the low
amounts of aggregates from mice tissue in order to study the in vivo aggregate structure and the different strain structures in more detail with ssNMR
and EM (184).
43
Populärvetenskaplig Sammanfattning
Proteiner är tillsammans med fetter och kolhydrater de byggstenar som
skapar levande organismer. Proteiner är i motsats till fetter och kolhydrater,
som oftast har en mer strukturell roll, de molekyler i en organism som överför information och utför arbete. Proteinerna är uppbyggda av kedjor av
aminosyror och deras funktion är beroende av den tredimensionella strukturen som varken får vara för flexibel eller för stel. Strukturen är en dynamisk jämvikt mellan veckat och uppveckat ostrukturerat protein. Tyvärr är
det många sjukdomar som uppkommer p.g.a. att proteiner inte fungerar som
de ska. Det kan bero på fel i uppbyggnaden av proteinet (mutationer) eller
förändrad miljö runt proteinet vilket påverkar dess struktur och därmed
funktion samt hur det interagerar med sin omgivning och med sig själv.
Den här avhandlingen handlar om proteinet superoxide dismutase 1 (SOD1)
som p.g.a. mutationer och förändringar i omgivningen kan klumpa ihop sig
och aggregera vilket kan orsaka sjukdomen amyotrofisk lateral skleros
(ALS) som leder till att kroppens muskler förtvinar och dör. Orsaken till
sjukdomen och mekanismen för bildandet av SOD1-aggregat är inte känd
men aggregerat SOD1 i form av amyloida fibrer hittas i både patienter och i
musmodeller av sjukdomen. Man tror att den förändrade proteinstrukturen
hos SOD1 kan ge upphov till de toxiska egenskaper som förstör nervcellerna
som styr kroppens muskler. ALS är i likhet med andra neurodegenerativa
sjukdomar som Alzheimers sjukdom och Parkinsons sjukdom klassificerad
som en protein-felveckningssjukdom där det kroppsegna proteinet undergår
en strukturell förändring och bildar cytotoxiska aggregat.
Våra resultat från studier av SOD1 aggregering visar att SOD1 börjar aggregera från det ostrukturerade uppveckade tillståndet eftersom aggregeringen
är snabbast då proteinet är helt oveckat. Vi ser samma resultat om proteinet
veckas upp kemiskt med hjälp av urea eller om det destabiliseras genom
mutationer i dess DNA (då en aminosyra byts mot en annan). Andra viktiga
iakttagelser är det svaga beroendet av SOD1-koncentrationen hos aggregeringshastigheten samt att provet måste skakas eller röras om för att aggregera. Båda dessa observationer visar att mekanismen för SOD1aggregeringen är dominerad av fiber fragmentering. Det innebär att ökningen av mängden aggregat främst sker genom att befintliga aggregat bryts sön-
44
der och i sin tur kan växa då SOD1 molekyler fäster till dess ändar. Men än
så länge vet vi inget om hur det första aggregatet bildas.
Intressant nog ser vi överensstämmande resultat in vivo där överlevnaden
hos transgena möss som uttrycker olika mutanter av humant SOD1 är relaterat till proteinernas stabilitet. Ju mer uppveckad den SOD1 mutant som uttrycks är desto kortare är överlevnaden hos musen. Även här ser vi ett svagt
beroende av proteinkoncentrationen hos hastigheten för aggregeringen vilket
indikerar att aggregeringsmekanismen är den samma i levande möss som i
provröret. Vi har även studerat SOD1 inuti levande celler med metoden incell NMR. Dessa studier visar att SOD1 är mer uppveckat och utsträckt inuti
cellen än i provröret. Denna observation talar för att koncentrationen av det
globalt uppveckade SOD1 som lätt aggregerar är högre i de trånga utrymmena i cellen än i provrörets utspädda lösning.
Slutligen kan vi med hjälp av antikroppar visa att det finns två olika typer av
aggregat i möss beroende på vilken mutation i SOD1 som uttrycks. I de två
olika typerna av aggregat ingår olika delar av SOD1s aminosyrasekvensen.
Detta resultat skulle kunna förklara varför ALS-patienter med olika mutationer har olika sjukdomsförlopp och symptom.
Våra resultat löser inte gåtan bakom hur SOD1 orsakar ALS men ger viktiga
ledtrådar för att bättre förstå mekanismen bakom och avgörande faktorer för
SOD1 aggregering, samt för protein-aggregering generellt, både i provröret
och i levande organismer.
45
Acknowledgements
Mikael Oliveberg, tack för att du smittat mig med din nyfikenhet och entusiasm för forskning, att hela tiden lära nya saker och ta reda på hur det
hänger ihop. Det har varit otroligt lärorikt att ta del av din erfarenhet och
kunskap, och en fantastisk upplevelse att få vara en del av din grupp och
jobba så som vi gör. Tack också Jens Danielsson för ett gott samarbete och
för att du delar med dig av din allomfattande visdom. Tack, Pia Harrysson
för att du alltid ställer upp och för alla godsaker, speciellt dina underbara
lussekatter. Together with the group members you make group MO/PH the
best! ☺
Group MO/PH, Sylvia Röstin, Huabing Wang, Xin Mu, Sarah Leeb,
Therese Grönlund, Seongil Choi and Fan Yang, thank you for sharing all
my ups and downs in the lab, always giving good advice and spreading a
positive feeling. Sylvia, tack för du orkar lyssna och ge goda råd om allt
mellan himmel och jord ☺.
Thanks also to former MO-group members: Ellinor Haglund, Ann-Sofi
Johansson, Lina Leinartaite, Ylva Pamment, Martin Kurnik and Mikael
Karlström for the fellowship we shared and for everything you taught me as
a new student in the lab. Tack Ellinor, det var din glädje och entusiasm tillsammans med Mikaels som fick mig att vilja göra mitt exjobb och sedan
fortsätta som phD-student i grupp MO. Ann-Sofi, tack för din fina handledning, din hjälp och ditt tålamod och för att du lät mig jobba på i lugn och ro
när jag var ny. Thanks also to all project students that have come and gone,
especially Pema Choden for all your hard work and Daniel Shaal my first
and best student.
Ett stort tack till våra samarbetspartners vid Umeå Universitet, Stefan
Marklund, Tomas Brännström och Peter Andersen med grupper för
givande samarbete och intressanta diskussioner. Mitt arbete hade inte haft
samma tyngd utan jämförelsen med era fantastiska data!
Thank you everyone at DBB, PIs, students and administrative staff for making DBB such a nice workplace. Tack, Ann Nielsen och Stefan Nordlund
för all hjälp och ert engagemang i phD-programmet.
46
Ett extra stort tack till Maria Sallander, Torbjörn Astlind, Matthew Bennet, Håkan Thoren och Alex Tuuling för att ni är så glada och trevliga och
gör min vardag så mycket enklare genom att lösa små och stora problem.
Slutligen vill jag tacka min familj och mina vänner för allt ert stöd och
tålamod, fast ni många gånger sett ut som frågetecken då jag försökt förklara
vad det är jag egentligen gör på jobbet.. ☺
Ett speciellt tack till Albin, Andreas, Jakob, Lena och Hans.
47
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
48
Rowland LP & Shneider NA (2001) Amyotrophic lateral sclerosis.
N Engl J Med 344(22):1688-1700.
Renton AE, Chio A, & Traynor BJ (2014) State of play in
amyotrophic lateral sclerosis genetics. Nat Neurosci 17(1):17-23.
Menzies FM, et al. (2002) Mitochondrial dysfunction in a cell
culture model of familial amyotrophic lateral sclerosis. Brain 125(Pt
7):1522-1533.
Jaiswal MK (2013) Calcium, mitochondria, and the pathogenesis of
ALS: the good, the bad, and the ugly. Front Cell Neurosci 7:199.
Prell T, et al. (2014) Endoplasmic reticulum stress is accompanied
by activation of NF-kappaB in amyotrophic lateral sclerosis. J
Neuroimmunol 270(1-2):29-36.
Kanekura K, Suzuki H, Aiso S, & Matsuoka M (2009) ER stress and
unfolded protein response in amyotrophic lateral sclerosis. Mol
Neurobiol 39(2):81-89.
Urushitani M, Kurisu J, Tsukita K, & Takahashi R (2002)
Proteasomal inhibition by misfolded mutant superoxide dismutase 1
induces selective motor neuron death in familial amyotrophic lateral
sclerosis. J Neurochem 83(5):1030-1042.
Kabashi E, Agar JN, Strong MJ, & Durham HD (2012) Impaired
proteasome function in sporadic amyotrophic lateral sclerosis.
Amyotroph Lateral Scler 13(4):367-371.
Li L, Zhang X, & Le W (2008) Altered macroautophagy in the
spinal cord of SOD1 mutant mice. Autophagy 4(3):290-293.
Kato S, et al. (2000) New consensus research on neuropathological
aspects of familial amyotrophic lateral sclerosis with superoxide
dismutase 1 (SOD1) gene mutations: inclusions containing SOD1 in
neurons and astrocytes. Amyotroph Lateral Scler Other Motor
Neuron Disord 1(3):163-184.
Bruijn LI, et al. (1997) ALS-linked SOD1 mutant G85R mediates
damage to astrocytes and promotes rapidly progressive disease with
SOD1-containing inclusions. Neuron 18(2):327-338.
Prudencio M, Hart PJ, Borchelt DR, & Andersen PM (2009)
Variation in aggregation propensities among ALS-associated
variants of SOD1: correlation to human disease. Hum Mol Genet
18(17):3217-3226.
Li TM, Alberman E, & Swash M (1988) Comparison of sporadic
and familial disease amongst 580 cases of motor neuron disease. J
Neurol Neurosurg Psychiatry 51(6):778-784.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Haverkamp LJ, Appel V, & Appel SH (1995) Natural history of
amyotrophic lateral sclerosis in a database population. Validation of
a scoring system and a model for survival prediction. Brain 118 ( Pt
3):707-719.
Andersen PM & Al-Chalabi A (2011) Clinical genetics of
amyotrophic lateral sclerosis: what do we really know? Nat Rev
Neurol 7(11):603-615.
Andersen PM, et al. (2003) Sixteen novel mutations in the Cu/Zn
superoxide dismutase gene in amyotrophic lateral sclerosis: a decade
of discoveries, defects and disputes. Amyotroph Lateral Scler Other
Motor Neuron Disord 4(2):62-73.
Rosen DR (1993) Mutations in Cu/Zn superoxide dismutase gene
are associated with familial amyotrophic lateral sclerosis. Nature
364(6435):362.
Kiernan MC, et al. (2011) Amyotrophic lateral sclerosis. Lancet
377(9769):942-955.
Battistini S, et al. (2005) SOD1 mutations in amyotrophic lateral
sclerosis. Results from a multicenter Italian study. J Neurol
252(7):782-788.
Shaw CE, et al. (1998) Mutations in all five exons of SOD-1 may
cause ALS. Ann Neurol 43(3):390-394.
Marklund SL (1984) Extracellular superoxide dismutase in human
tissues and human cell lines. J Clin Invest 74(4):1398-1403.
Reaume AG, et al. (1996) Motor neurons in Cu/Zn superoxide
dismutase-deficient mice develop normally but exhibit enhanced cell
death after axonal injury. Nat Genet 13(1):43-47.
Hayward LJ, et al. (2002) Decreased metallation and activity in
subsets of mutant superoxide dismutases associated with familial
amyotrophic lateral sclerosis. J Biol Chem 277(18):15923-15931.
Graffmo KS, et al. (2013) Expression of wild-type human
superoxide dismutase-1 in mice causes amyotrophic lateral sclerosis.
Hum Mol Genet 22(1):51-60.
Jonsson PA, et al. (2006) Motor neuron disease in mice expressing
the wild type-like D90A mutant superoxide dismutase-1. J
Neuropathol Exp Neurol 65(12):1126-1136.
Sandelin E, Nordlund A, Andersen PM, Marklund SS, & Oliveberg
M (2007) Amyotrophic lateral sclerosis-associated copper/zinc
superoxide dismutase mutations preferentially reduce the repulsive
charge of the proteins. J Biol Chem 282(29):21230-21236.
Lindberg MJ, Bystrom R, Boknas N, Andersen PM, & Oliveberg M
(2005) Systematically perturbed folding patterns of amyotrophic
lateral sclerosis (ALS)-associated SOD1 mutants. Proc Natl Acad
Sci U S A 102(28):9754-9759.
Tiwari A & Hayward LJ (2005) Mutant SOD1 instability:
implications for toxicity in amyotrophic lateral sclerosis.
Neurodegener Dis 2(3-4):115-127.
49
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
50
Forsberg K, Andersen PM, Marklund SL, & Brannstrom T (2011)
Glial nuclear aggregates of superoxide dismutase-1 are regularly
present in patients with amyotrophic lateral sclerosis. Acta
Neuropathol 121(5):623-634.
Bosco DA, et al. (2010) Wild-type and mutant SOD1 share an
aberrant conformation and a common pathogenic pathway in ALS.
Nat Neurosci 13(11):1396-1403.
Forsberg K, et al. (2010) Novel antibodies reveal inclusions
containing non-native SOD1 in sporadic ALS patients. PLoS One
5(7):e11552.
Furukawa Y, Kaneko K, Yamanaka K, O'Halloran TV, & Nukina N
(2008) Complete loss of post-translational modifications triggers
fibrillar aggregation of SOD1 in the familial form of amyotrophic
lateral sclerosis. J Biol Chem 283(35):24167-24176.
Chattopadhyay M, et al. (2008) Initiation and elongation in
fibrillation of ALS-linked superoxide dismutase. Proc Natl Acad Sci
U S A 105(48):18663-18668.
Stathopulos PB, et al. (2004) Sonication of proteins causes
formation of aggregates that resemble amyloid. Protein Sci
13(11):3017-3027.
Richardson JS, Richardson DC, Thomas KA, Silverton EW, &
Davies DR (1976) Similarity of three-dimensional structure between
the immunoglobulin domain and the copper, zinc superoxide
dismutase subunit. J Mol Biol 102(2):221-235.
Parge HE, Hallewell RA, & Tainer JA (1992) Atomic structures of
wild-type and thermostable mutant recombinant human Cu,Zn
superoxide dismutase. Proc Natl Acad Sci U S A 89(13):6109-6113.
Getzoff ED, et al. (1983) Electrostatic recognition between
superoxide and copper, zinc superoxide dismutase. Nature
306(5940):287-290.
Rakhit R & Chakrabartty A (2006) Structure, folding, and
misfolding of Cu,Zn superoxide dismutase in amyotrophic lateral
sclerosis. Biochim Biophys Acta 1762(11-12):1025-1037.
Getzoff ED, et al. (1992) Faster superoxide dismutase mutants
designed
by
enhancing
electrostatic
guidance.
Nature
358(6384):347-351.
Roe JA, et al. (1988) Differential scanning calorimetry of Cu,Znsuperoxide dismutase, the apoprotein, and its zinc-substituted
derivatives. Biochemistry 27(3):950-958.
Rodriguez JA, et al. (2002) Familial amyotrophic lateral sclerosisassociated mutations decrease the thermal stability of distinctly
metallated species of human copper/zinc superoxide dismutase. J
Biol Chem 277(18):15932-15937.
Leinartaite L, Saraboji K, Nordlund A, Logan DT, & Oliveberg M
(2010) Folding catalysis by transient coordination of Zn2+ to the Cu
ligands of the ALS-associated enzyme Cu/Zn superoxide dismutase
1. J Am Chem Soc 132(38):13495-13504.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
Strange RW, et al. (2003) The structure of holo and metal-deficient
wild-type human Cu, Zn superoxide dismutase and its relevance to
familial amyotrophic lateral sclerosis. J Mol Biol 328(4):877-891.
Banci L, Bertini I, Cramaro F, Del Conte R, & Viezzoli MS (2003)
Solution structure of Apo Cu,Zn superoxide dismutase: role of metal
ions in protein folding. Biochemistry 42(32):9543-9553.
Durazo A, et al. (2009) Metal-free superoxide dismutase-1 and three
different amyotrophic lateral sclerosis variants share a similar
partially unfolded beta-barrel at physiological temperature. J Biol
Chem 284(49):34382-34389.
Lindberg MJ, Normark J, Holmgren A, & Oliveberg M (2004)
Folding of human superoxide dismutase: disulfide reduction
prevents dimerization and produces marginally stable monomers.
Proc Natl Acad Sci U S A 101(45):15893-15898.
Furukawa Y, Torres AS, & O'Halloran TV (2004) Oxygen-induced
maturation of SOD1: a key role for disulfide formation by the
copper chaperone CCS. EMBO J 23(14):2872-2881.
Hornberg A, Logan DT, Marklund SL, & Oliveberg M (2007) The
coupling between disulphide status, metallation and dimer interface
strength in Cu/Zn superoxide dismutase. J Mol Biol 365(2):333-342.
Kayatekin C, Zitzewitz JA, & Matthews CR (2008) Zinc binding
modulates the entire folding free energy surface of human Cu,Zn
superoxide dismutase. J Mol Biol 384(2):540-555.
Lamb AL, Torres AS, O'Halloran TV, & Rosenzweig AC (2001)
Heterodimeric structure of superoxide dismutase in complex with its
metallochaperone. Nat Struct Biol 8(9):751-755.
Shaw BF & Valentine JS (2007) How do ALS-associated mutations
in superoxide dismutase 1 promote aggregation of the protein?
Trends Biochem Sci 32(2):78-85.
Hough MA, et al. (2004) Dimer destabilization in superoxide
dismutase may result in disease-causing properties: structures of
motor neuron disease mutants. Proc Natl Acad Sci U S A
101(16):5976-5981.
Lindberg MJ, Tibell L, & Oliveberg M (2002) Common
denominator of Cu/Zn superoxide dismutase mutants associated with
amyotrophic lateral sclerosis: decreased stability of the apo state.
Proc Natl Acad Sci U S A 99(26):16607-16612.
Lelie HL, et al. (2011) Copper and zinc metallation status of copperzinc superoxide dismutase from amyotrophic lateral sclerosis
transgenic mice. J Biol Chem 286(4):2795-2806.
Tiwari A & Hayward LJ (2003) Familial amyotrophic lateral
sclerosis mutants of copper/zinc superoxide dismutase are
susceptible to disulfide reduction. J Biol Chem 278(8):5984-5992.
Furukawa Y & O'Halloran TV (2005) Amyotrophic lateral sclerosis
mutations have the greatest destabilizing effect on the apo- and
reduced form of SOD1, leading to unfolding and oxidative
aggregation. J Biol Chem 280(17):17266-17274.
51
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
52
Kayatekin C, Zitzewitz JA, & Matthews CR (2010) Disulfidereduced ALS variants of Cu, Zn superoxide dismutase exhibit
increased populations of unfolded species. J Mol Biol 398(2):320331.
DiDonato M, et al. (2003) ALS mutants of human superoxide
dismutase form fibrous aggregates via framework destabilization. J
Mol Biol 332(3):601-615.
Teilum K, et al. (2009) Transient structural distortion of metal-free
Cu/Zn superoxide dismutase triggers aberrant oligomerization. Proc
Natl Acad Sci U S A 106(43):18273-18278.
Nordlund A & Oliveberg M (2006) Folding of Cu/Zn superoxide
dismutase suggests structural hotspots for gain of neurotoxic
function in ALS: parallels to precursors in amyloid disease. Proc
Natl Acad Sci U S A 103(27):10218-10223.
Lepock JR, Frey HE, & Hallewell RA (1990) Contribution of
conformational stability and reversibility of unfolding to the
increased thermostability of human and bovine superoxide
dismutase mutated at free cysteines. J Biol Chem 265(35):2161221618.
Bertini I, Piccioli M, Viezzoli MS, Chiu CY, & Mullenbach GT
(1994) A spectroscopic characterization of a monomeric analog of
copper, zinc superoxide dismutase. Eur Biophys J 23(3):167-176.
Privalov PL (1979) Stability of proteins: small globular proteins.
Adv Protein Chem 33:167-241.
Fersht AR (1999) Structure and mechanism in protein science: a
guide to enzyme catalysis and protein folding (WH Freeman and Co.
New York).
Kubelka J, Hofrichter J, & Eaton WA (2004) The protein folding
'speed limit'. Curr Opin Struct Biol 14(1):76-88.
Tanford C (1970) Protein denaturation. C. Theoretical models for
the mechanism of denaturation. Adv Protein Chem 24:1-95.
Royer CA (2006) Probing protein folding and conformational
transitions with fluorescence. Chem Rev 106(5):1769-1784.
Myers JK, Pace CN, & Scholtz JM (1995) Denaturant m values and
heat capacity changes: relation to changes in accessible surface areas
of protein unfolding. Protein Sci 4(10):2138-2148.
Fersht AR (2000) Transition-state structure as a unifying basis in
protein-folding mechanisms: contact order, chain topology, stability,
and the extended nucleus mechanism. Proc Natl Acad Sci U S A
97(4):1525-1529.
Dobson CM (1999) Protein misfolding, evolution and disease.
Trends Biochem Sci 24(9):329-332.
Eichner T & Radford SE (2011) Understanding the complex
mechanisms of beta2-microglobulin amyloid assembly. FEBS J
278(20):3868-3883.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
Karamanos TK, et al. (2016) A Population Shift between Sparsely
Populated Folding Intermediates Determines Amyloidogenicity. J
Am Chem Soc 138(19):6271-6280.
Nokwe CN, et al. (2016) A Stable Mutant Predisposes Antibody
Domains to Amyloid Formation through Specific Non-Native
Interactions. J Mol Biol 428(6):1315-1332.
Kayatekin C, Cohen NR, & Matthews CR (2012) Enthalpic barriers
dominate the folding and unfolding of the human Cu, Zn superoxide
dismutase monomer. J Mol Biol 424(3-4):192-202.
Wang H, Lang L, Logan DT, Danielsson J, & Oliveberg M (2016)
Tricking a protein to swap strands. J Am Chem Soc.
Baldwin AJ, et al. (2011) Metastability of native proteins and the
phenomenon of amyloid formation. J Am Chem Soc 133(36):1416014163.
Routledge KE, Tartaglia GG, Platt GW, Vendruscolo M, & Radford
SE (2009) Competition between intramolecular and intermolecular
interactions in an amyloid-forming protein. J Mol Biol 389(4):776786.
Buell AK, et al. (2012) Detailed analysis of the energy barriers for
amyloid fibril growth. Angew Chem Int Ed Engl 51(21):5247-5251.
Ciryam P, Kundra R, Morimoto RI, Dobson CM, & Vendruscolo M
(2015) Supersaturation is a major driving force for protein
aggregation in neurodegenerative diseases. Trends Pharmacol Sci
36(2):72-77.
Chiti F, et al. (2002) Kinetic partitioning of protein folding and
aggregation. Nat Struct Biol 9(2):137-143.
Otzen DE, Kristensen O, & Oliveberg M (2000) Designed protein
tetramer zipped together with a hydrophobic Alzheimer homology: a
structural clue to amyloid assembly. Proc Natl Acad Sci U S A
97(18):9907-9912.
Tartaglia GG, Pechmann S, Dobson CM, & Vendruscolo M (2007)
Life on the edge: a link between gene expression levels and
aggregation rates of human proteins. Trends Biochem Sci 32(5):204206.
Jonsson PA, et al. (2006) Disulphide-reduced superoxide dismutase1 in CNS of transgenic amyotrophic lateral sclerosis models. Brain
129(Pt 2):451-464.
Kerman A, et al. (2010) Amyotrophic lateral sclerosis is a nonamyloid disease in which extensive misfolding of SOD1 is unique to
the familial form. Acta Neuropathol 119(3):335-344.
Wang J, et al. (2002) Fibrillar inclusions and motor neuron
degeneration in transgenic mice expressing superoxide dismutase 1
with a disrupted copper-binding site. Neurobiol Dis 10(2):128-138.
Weisberg SJ, et al. (2012) Compartmentalization of superoxide
dismutase 1 (SOD1G93A) aggregates determines their toxicity. Proc
Natl Acad Sci U S A 109(39):15811-15816.
53
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
54
Brotherton TE, Li Y, & Glass JD (2013) Cellular toxicity of mutant
SOD1 protein is linked to an easily soluble, non-aggregated form in
vitro. Neurobiol Dis 49:49-56.
Hwang YM, et al. (2010) Nonamyloid aggregates arising from
mature copper/zinc superoxide dismutases resemble those observed
in amyotrophic lateral sclerosis. J Biol Chem 285(53):41701-41711.
Redler RL, Fee L, Fay JM, Caplow M, & Dokholyan NV (2014)
Non-native soluble oligomers of Cu/Zn superoxide dismutase
(SOD1) contain a conformational epitope linked to cytotoxicity in
amyotrophic lateral sclerosis (ALS). Biochemistry 53(14):24232432.
Banci L, et al. (2007) Metal-free superoxide dismutase forms
soluble oligomers under physiological conditions: a possible general
mechanism for familial ALS. Proc Natl Acad Sci U S A
104(27):11263-11267.
Proctor EA, et al. (2016) Nonnative SOD1 trimer is toxic to motor
neurons in a model of amyotrophic lateral sclerosis. Proc Natl Acad
Sci U S A 113(3):614-619.
Andersen PM, et al. (1997) Phenotypic heterogeneity in motor
neuron disease patients with CuZn-superoxide dismutase mutations
in Scandinavia. Brain 120 ( Pt 10):1723-1737.
Eisen A, et al. (2008) SOD1 gene mutations in ALS patients from
British Columbia, Canada: clinical features, neurophysiology and
ethical issues in management. Amyotroph Lateral Scler 9(2):108119.
Brotherton T, et al. (2011) A novel ALS SOD1 C6S mutation with
implications for aggregation related toxicity and genetic counseling.
Amyotroph Lateral Scler 12(3):215-219.
Karch CM & Borchelt DR (2008) A limited role for disulfide crosslinking in the aggregation of mutant SOD1 linked to familial
amyotrophic lateral sclerosis. J Biol Chem 283(20):13528-13537.
Bergemalm D, et al. (2010) Superoxide dismutase-1 and other
proteins in inclusions from transgenic amyotrophic lateral sclerosis
model mice. J Neurochem 114(2):408-418.
Ugalde CL, Finkelstein DI, Lawson VA, & Hill AF (2016)
Pathogenic mechanisms of prion protein, amyloid-beta and alphasynuclein misfolding: the prion concept and neurotoxicity of protein
oligomers. J Neurochem.
Guest WC, et al. (2011) Generalization of the prion hypothesis to
other neurodegenerative diseases: an imperfect fit. J Toxicol Environ
Health A 74(22-24):1433-1459.
Kabir ME & Safar JG (2014) Implications of prion adaptation and
evolution paradigm for human neurodegenerative diseases. Prion
8(1):111-116.
Chia R, et al. (2010) Superoxide dismutase 1 and tgSOD1 mouse
spinal cord seed fibrils, suggesting a propagative cell death
mechanism in amyotrophic lateral sclerosis. PLoS One 5(5):e10627.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
Furukawa Y, Kaneko K, Watanabe S, Yamanaka K, & Nukina N
(2013) Intracellular seeded aggregation of mutant Cu,Zn-superoxide
dismutase associated with amyotrophic lateral sclerosis. FEBS Lett
587(16):2500-2505.
Pokrishevsky E, Grad LI, & Cashman NR (2016) TDP-43 or FUSinduced misfolded human wild-type SOD1 can propagate
intercellularly in a prion-like fashion. Sci Rep 6:22155.
Grad LI, et al. (2014) Intercellular propagated misfolding of wildtype Cu/Zn superoxide dismutase occurs via exosome-dependent
and -independent mechanisms. Proc Natl Acad Sci U S A
111(9):3620-3625.
Munch C, O'Brien J, & Bertolotti A (2011) Prion-like propagation of
mutant superoxide dismutase-1 misfolding in neuronal cells. Proc
Natl Acad Sci U S A 108(9):3548-3553.
Gomes C, Keller S, Altevogt P, & Costa J (2007) Evidence for
secretion of Cu,Zn superoxide dismutase via exosomes from a cell
model of amyotrophic lateral sclerosis. Neurosci Lett 428(1):43-46.
Polymenidou M & Cleveland DW (2011) The seeds of
neurodegeneration: prion-like spreading in ALS. Cell 147(3):498508.
Haidet-Phillips AM, et al. (2011) Astrocytes from familial and
sporadic ALS patients are toxic to motor neurons. Nat Biotechnol
29(9):824-828.
Ayers JI, Fromholt SE, O'Neal VM, Diamond JH, & Borchelt DR
(2016) Prion-like propagation of mutant SOD1 misfolding and
motor neuron disease spread along neuroanatomical pathways. Acta
Neuropathol 131(1):103-114.
Bidhendi EE, et al. (2016) Two superoxide dismutase prion strains
transmit amyotrophic lateral sclerosis-like disease. J Clin Invest.
Ravits JM & La Spada AR (2009) ALS motor phenotype
heterogeneity, focality, and spread: deconstructing motor neuron
degeneration. Neurology 73(10):805-811.
Verma A (2013) Protein aggregates and regional disease spread in
ALS is reminiscent of prion-like pathogenesis. Neurol India
61(2):107-110.
Bett C, et al. (2012) Biochemical properties of highly neuroinvasive
prion strains. PLoS Pathog 8(2):e1002522.
Chan PK, et al. (2013) Structural similarity of wild-type and ALSmutant superoxide dismutase-1 fibrils using limited proteolysis and
atomic force microscopy. Proc Natl Acad Sci U S A 110(27):1093410939.
Furukawa Y, Kaneko K, Yamanaka K, & Nukina N (2010)
Mutation-dependent polymorphism of Cu,Zn-superoxide dismutase
aggregates in the familial form of amyotrophic lateral sclerosis. J
Biol Chem 285(29):22221-22231.
Toichi K, Yamanaka K, & Furukawa Y (2013) Disulfide scrambling
describes the oligomer formation of superoxide dismutase (SOD1)
55
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
56
proteins in the familial form of amyotrophic lateral sclerosis. J Biol
Chem 288(7):4970-4980.
Wang Q, Johnson JL, Agar NY, & Agar JN (2008) Protein
aggregation and protein instability govern familial amyotrophic
lateral sclerosis patient survival. PLoS Biol 6(7):e170.
Lee S & Kim HJ (2015) Prion-like Mechanism in Amyotrophic
Lateral Sclerosis: are Protein Aggregates the Key? Exp Neurobiol
24(1):1-7.
Dobson CM (2003) Protein folding and misfolding. Nature
426(6968):884-890.
Eisenberg D & Jucker M (2012) The amyloid state of proteins in
human diseases. Cell 148(6):1188-1203.
Chiti F & Dobson CM (2006) Protein misfolding, functional
amyloid, and human disease. Annu Rev Biochem 75:333-366.
Nielsen L, et al. (2001) Effect of environmental factors on the
kinetics of insulin fibril formation: elucidation of the molecular
mechanism. Biochemistry 40(20):6036-6046.
Cohen SI, Vendruscolo M, Dobson CM, & Knowles TP (2012)
From macroscopic measurements to microscopic mechanisms of
protein aggregation. J Mol Biol 421(2-3):160-171.
Cohen SI, et al. (2011) Nucleated polymerization with secondary
pathways. I. Time evolution of the principal moments. J Chem Phys
135(6):065105.
Knowles TP, et al. (2009) An analytical solution to the kinetics of
breakable filament assembly. Science 326(5959):1533-1537.
Meisl G, et al. (2016) Molecular mechanisms of protein aggregation
from global fitting of kinetic models. Nat Protoc 11(2):252-272.
Fandrich M, et al. (2003) Myoglobin forms amyloid fibrils by
association of unfolded polypeptide segments. Proc Natl Acad Sci U
S A 100(26):15463-15468.
Ferguson N, et al. (2006) General structural motifs of amyloid
protofilaments. Proc Natl Acad Sci U S A 103(44):16248-16253.
Bouchard M, Zurdo J, Nettleton EJ, Dobson CM, & Robinson CV
(2000) Formation of insulin amyloid fibrils followed by FTIR
simultaneously with CD and electron microscopy. Protein Sci
9(10):1960-1967.
Plakoutsi G, et al. (2005) Evidence for a mechanism of amyloid
formation involving molecular reorganisation within native-like
precursor aggregates. J Mol Biol 351(4):910-922.
Bousset L, et al. (2004) Structural characterization of the fibrillar
form of the yeast Saccharomyces cerevisiae prion Ure2p.
Biochemistry 43(17):5022-5032.
Bousset L, Thomson NH, Radford SE, & Melki R (2002) The yeast
prion Ure2p retains its native alpha-helical conformation upon
assembly into protein fibrils in vitro. EMBO J 21(12):2903-2911.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
Nordlund A, et al. (2009) Functional features cause misfolding of
the ALS-provoking enzyme SOD1. Proc Natl Acad Sci U S A
106(24):9667-9672.
Karch CM & Borchelt DR (2010) Aggregation modulating elements
in mutant human superoxide dismutase 1. Arch Biochem Biophys
503(2):175-182.
Danielsson J, Kurnik M, Lang L, & Oliveberg M (2011) Cutting off
functional loops from homodimeric enzyme superoxide dismutase 1
(SOD1) leaves monomeric beta-barrels. J Biol Chem
286(38):33070-33083.
Groenning M, Campos RI, Hirschberg D, Hammarstrom P, &
Vestergaard B (2015) Considerably Unfolded Transthyretin
Monomers Preceed and Exchange with Dynamically Structured
Amyloid Protofibrils. Sci Rep 5:11443.
Collins SR, Douglass A, Vale RD, & Weissman JS (2004)
Mechanism of prion propagation: amyloid growth occurs by
monomer addition. PLoS Biol 2(10):e321.
Ferguson N, et al. (2003) Rapid amyloid fiber formation from the
fast-folding WW domain FBP28. Proc Natl Acad Sci U S A
100(17):9814-9819.
Wright CF, Teichmann SA, Clarke J, & Dobson CM (2005) The
importance of sequence diversity in the aggregation and evolution of
proteins. Nature 438(7069):878-881.
Xue WF, Homans SW, & Radford SE (2008) Systematic analysis of
nucleation-dependent polymerization reveals new insights into the
mechanism of amyloid self-assembly. Proc Natl Acad Sci U S A
105(26):8926-8931.
Hellstrand E, Boland B, Walsh DM, & Linse S (2010) Amyloid
beta-protein aggregation produces highly reproducible kinetic data
and occurs by a two-phase process. ACS Chem Neurosci 1(1):13-18.
Fodera V, Librizzi F, Groenning M, van de Weert M, & Leone M
(2008) Secondary nucleation and accessible surface in insulin
amyloid fibril formation. J Phys Chem B 112(12):3853-3858.
Meisl G, et al. (2014) Differences in nucleation behavior underlie
the contrasting aggregation kinetics of the Abeta40 and Abeta42
peptides. Proc Natl Acad Sci U S A 111(26):9384-9389.
Ferrone F (1999) Analysis of protein aggregation kinetics. Methods
Enzymol 309:256-274.
Zhu L, Zhang XJ, Wang LY, Zhou JM, & Perrett S (2003)
Relationship between stability of folding intermediates and amyloid
formation for the yeast prion Ure2p: a quantitative analysis of the
effects of pH and buffer system. J Mol Biol 328(1):235-254.
Fodera V, et al. (2008) Thioflavin T hydroxylation at basic pH and
its effect on amyloid fibril detection. J Phys Chem B 112(47):1517415181.
57
146.
147.
148.
149.
150.
151.
152.
153.
154.
155.
156.
157.
158.
159.
160.
58
Buell AK, et al. (2014) Solution conditions determine the relative
importance of nucleation and growth processes in alpha-synuclein
aggregation. Proc Natl Acad Sci U S A 111(21):7671-7676.
Maurer-Stroh S, et al. (2010) Exploring the sequence determinants
of amyloid structure using position-specific scoring matrices. Nat
Methods 7(3):237-242.
Bergh J, et al. (2015) Structural and kinetic analysis of proteinaggregate strains in vivo using binary epitope mapping. Proc Natl
Acad Sci U S A 112(14):4489-4494.
Wurth C, Guimard NK, & Hecht MH (2002) Mutations that reduce
aggregation of the Alzheimer's Abeta42 peptide: an unbiased search
for the sequence determinants of Abeta amyloidogenesis. J Mol Biol
319(5):1279-1290.
Schwartz R, Istrail S, & King J (2001) Frequencies of amino acid
strings in globular protein sequences indicate suppression of blocks
of consecutive hydrophobic residues. Protein Sci 10(5):1023-1031.
Chiti F, et al. (2002) Studies of the aggregation of mutant proteins in
vitro provide insights into the genetics of amyloid diseases. Proc
Natl Acad Sci U S A 99 Suppl 4:16419-16426.
Schmittschmitt JP & Scholtz JM (2003) The role of protein stability,
solubility, and net charge in amyloid fibril formation. Protein Sci
12(10):2374-2378.
Schreiber G, Haran G, & Zhou HX (2009) Fundamental aspects of
protein-protein association kinetics. Chem Rev 109(3):839-860.
Banci L, et al. (2014) Solid-state NMR studies of metal-free SOD1
fibrillar structures. J Biol Inorg Chem 19(4-5):659-666.
Zetterstrom P, Graffmo KS, Andersen PM, Brannstrom T, &
Marklund SL (2011) Proteins that bind to misfolded mutant
superoxide dismutase-1 in spinal cords from transgenic amyotrophic
lateral sclerosis (ALS) model mice. J Biol Chem 286(23):2013020136.
Wang J, et al. (2009) Progressive aggregation despite chaperone
associations of a mutant SOD1-YFP in transgenic mice that develop
ALS. Proc Natl Acad Sci U S A 106(5):1392-1397.
Aisenbrey C, et al. (2008) How is protein aggregation in
amyloidogenic diseases modulated by biological membranes? Eur
Biophys J 37(3):247-255.
Colby DW & Prusiner SB (2011) Prions. Cold Spring Harb Perspect
Biol 3(1):a006833.
Xue WF, Hellewell AL, Hewitt EW, & Radford SE (2010) Fibril
fragmentation in amyloid assembly and cytotoxicity: when size
matters. Prion 4(1):20-25.
Tanaka M, Collins SR, Toyama BH, & Weissman JS (2006) The
physical basis of how prion conformations determine strain
phenotypes. Nature 442(7102):585-589.
161.
162.
163.
164.
165.
166.
167.
168.
169.
170.
171.
172.
173.
174.
175.
176.
177.
178.
Cohen SI, et al. (2013) Proliferation of amyloid-beta42 aggregates
occurs through a secondary nucleation mechanism. Proc Natl Acad
Sci U S A 110(24):9758-9763.
Abdolvahabi A, et al. (2017) How Do Gyrating Beads Accelerate
Amyloid Fibrillization? Biophys J 112(2):250-264.
Inomata K, et al. (2009) High-resolution multi-dimensional NMR
spectroscopy of proteins in human cells. Nature 458(7234):106-109.
Barbieri L, Luchinat E, & Banci L (2015) Protein interaction
patterns in different cellular environments are revealed by in-cell
NMR. Sci Rep 5:14456.
Tochio H (2012) Watching protein structure at work in living cells
using NMR spectroscopy. Curr Opin Chem Biol 16(5-6):609-613.
Serber Z, et al. (2006) Investigating macromolecules inside cultured
and injected cells by in-cell NMR spectroscopy. Nat Protoc
1(6):2701-2709.
Danielsson J, et al. (2013) Pruning the ALS-associated protein
SOD1 for in-cell NMR. J Am Chem Soc 135(28):10266-10269.
Ghaemmaghami S & Oas TG (2001) Quantitative protein stability
measurement in vivo. Nat Struct Biol 8(10):879-882.
Monteith WB & Pielak GJ (2014) Residue level quantification of
protein stability in living cells. Proc Natl Acad Sci U S A
111(31):11335-11340.
Dhar A, Ebbinghaus S, Shen Z, Mishra T, & Gruebele M (2010) The
diffusion coefficient for PGK folding in eukaryotic cells. Biophys J
99(9):L69-71.
Dhar A, et al. (2011) Protein stability and folding kinetics in the
nucleus and endoplasmic reticulum of eucaryotic cells. Biophys J
101(2):421-430.
Guo M, Xu Y, & Gruebele M (2012) Temperature dependence of
protein folding kinetics in living cells. Proc Natl Acad Sci U S A
109(44):17863-17867.
Monteith WB, Cohen RD, Smith AE, Guzman-Cisneros E, & Pielak
GJ (2015) Quinary structure modulates protein stability in cells.
Proc Natl Acad Sci U S A 112(6):1739-1742.
Ignatova Z & Gierasch LM (2007) Effects of osmolytes on protein
folding and aggregation in cells. Methods Enzymol 428:355-372.
Ignatova Z, et al. (2007) From the test tube to the cell: exploring the
folding and aggregation of a beta-clam protein. Biopolymers
88(2):157-163.
Sarkar M, Smith AE, & Pielak GJ (2013) Impact of reconstituted
cytosol on protein stability. Proc Natl Acad Sci U S A
110(48):19342-19347.
Guzman I, Gelman H, Tai J, & Gruebele M (2014) The extracellular
protein VlsE is destabilized inside cells. J Mol Biol 426(1):11-20.
Wirth AJ, Platkov M, & Gruebele M (2013) Temporal variation of a
protein folding energy landscape in the cell. J Am Chem Soc
135(51):19215-19221.
59
179.
180.
181.
182.
183.
184.
60
Smith AE, Zhou LZ, Gorensek AH, Senske M, & Pielak GJ (2016)
In-cell thermodynamics and a new role for protein surfaces. Proc
Natl Acad Sci U S A 113(7):1725-1730.
Zhou HX, Rivas G, & Minton AP (2008) Macromolecular crowding
and confinement: biochemical, biophysical, and potential
physiological consequences. Annu Rev Biophys 37:375-397.
Cheung MS, Klimov D, & Thirumalai D (2005) Molecular crowding
enhances native state stability and refolding rates of globular
proteins. Proc Natl Acad Sci U S A 102(13):4753-4758.
Ellis RJ & Minton AP (2003) Cell biology: join the crowd. Nature
425(6953):27-28.
Wuttke R, et al. (2014) Temperature-dependent solvation modulates
the dimensions of disordered proteins. Proc Natl Acad Sci U S A
111(14):5213-5218.
Lu JX, et al. (2013) Molecular structure of beta-amyloid fibrils in
Alzheimer's disease brain tissue. Cell 154(6):1257-1268.