Báo cáo khoa học: Monitoring the prevention of amyloid fibril formation by a-crystallin Temperature dependence and the nature of the aggregating species pdf

Monitoring the prevention of amyloid fibril formation
by a-crystallin
Temperature dependence and the nature of the aggregating species
Agata Rekas
1,2
, Lucy Jankova
3
, David C. Thorn
4
, Roberto Cappai
5,6
and John A. Carver
4
1 Department of Chemistry, University of Wollongong, Australia
2 Institute for Environmental Research, Australian Nuclear Science and Technology Organization, Menai, Australia
3 ATA Scientific Pty Ltd ANSTO Woods Centre, Lucas Heights, Australia
4 School of Chemistry and Physics, The University of Adelaide, Australia
5 Department of Pathology, Bio21 Institute, The University of Melbourne, Australia
6 Mental Health Research Institute, Melbourne, Australia
a-Crystallin is a molecular chaperone of the small
heat-shock protein (sHsp) family. It is known to recog-
nize and interact with long-lived partially folded pro-
teins on their off-folding pathway to prevent their
aggregation [1,2]. Two closely related subunits of
a-crystallin exist in high concentrations in mammalian
lenses, aA- and aB-crystallin; in humans they are pres-
ent in a ratio of 3 : 1. Whereas aA-crystallin is lens
specific, aB-crystallin is also found extralenticularly in
retina, heart, skeletal muscle, skin, kidney, brain,
spinal cord and lungs, as well as in CNS glial cells and
neurons in some pathological conditions, e.g. Alzhei-

mer’s disease and dementia with Lewy bodies [3–5].
The effectiveness of a-crystallin as a chaperone in
preventing amorphous aggregation of destabilized
proteins increases with temperature [6–8]. a-Crystallin
occurs in large supramolecular assemblies of average
mass  800 kDa [9], in dynamic equilibrium with
Keywords
amyloid; dual polarization interferometry;
NMR spectroscopy; small heat shock
protein; temperature dependence
Correspondence
J. A. Carver, School of Chemistry and
Physics, The University of Adelaide,
Adelaide, South Australia 5005, Australia
Fax: +61 8 8303 4380
Tel: +61 8 8303 3110
E-mail:
(Received 22 May 2007, revised 12 October
2007, accepted 16 October 2007)
doi:10.1111/j.1742-4658.2007.06144.x
The molecular chaperone, a-crystallin, has the ability to prevent the fibril-
lar aggregation of proteins implicated in human diseases, for example,
amyloid b peptide and a-synuclein. In this study, we examine, in detail,
two aspects of a-crystallin’s fibril-suppressing ability: (a) its temperature
dependence, and (b) the nature of the aggregating species with which it
interacts. First, the efficiency of a-crystallin to suppress fibril formation in
j-casein and a-synuclein increases with temperature, despite their rate of
fibrillation also increasing in the absence of a-crystallin. This is consistent
with an increased chaperone ability of a-crystallin at higher temperatures
to protect target proteins from amorphous aggregation [GB Reddy, KP

Das, JM Petrash & WK Surewicz (2000) J Biol Chem 275, 4565–4570]. Sec-
ond, dual polarization interferometry was used to monitor real-time a-syn-
uclein aggregation in the presence and absence of aB-crystallin. In contrast
to more common methods for monitoring the time-dependent formation of
amyloid fibrils (e.g. the binding of dyes like thioflavin T), dual polarization
interferometry data did not reveal any initial lag phase, generally attributed
to the formation of prefibrillar aggregates. It was shown that aB-crystallin
interrupted a-synuclein aggregation at its earliest stages, most likely by
binding to partially folded monomers and thereby preventing their aggrega-
tion into fibrillar structures.
Abbreviations
ANS, 8-anilinonaphthalene 1-sulfonate; DPI, dual polarization interferometry; sHsp, small heat-shock protein; TEM, transmission electron
microscopy; TFT, thioflavin T.
6290 FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS
dissociated subunits. The rate of subunit exchange in
a-crystallin increases with temperature [8]. The correla-
tion between the temperature dependency of chaperone
efficiency and subunit exchange suggests that it is pri-
marily dissociated forms of sHsps that interact with
destabilized target proteins [10]. Specifically, enhanced
chaperone activity at higher temperatures has been
attributed to an increase in the subunit exchange rate
[8], and thus the availability of the dissociated, proba-
bly dimeric, active forms of a-crystallin [2,11], along
with concomitant structural changes in a-crystallin at
higher temperatures [6,12,13].
More recently, a-crystallin was found to prevent the
formation of amyloid fibrils by various proteins (e.g.
Ab peptide, apolipoprotein CII, a-synuclein) [14–18].
Fibrillar aggregation by a number of proteins, includ-

ing the aforementioned, forms the basis of many clini-
cal disorders (e.g. Ab peptide in Alzheimer’s disease,
a-synuclein in Parkinson’s disease, amylin in type II
diabetes, b2-microglobulin in dialysis-related amyloido-
sis and prion protein in Creutzfeldt–Jakob disease). In
comparison with amorphous aggregation, fibril forma-
tion is a slower and more ordered pathway of protein
aggregation, however, both processes require proteins
to adopt a conformation that is only partially folded,
either by unfolding of a structured molecule, e.g. a-lact-
albumin [19], or, in the case of intrinsically unstruc-
tured (also known as natively disordered) proteins such
as a-synuclein or j-casein, by stabilizing a conforma-
tion that is already relatively unstructured [20].
Whether a protein aggregates amorphously or forms
highly ordered fibrillar structures most likely depends
on the structural characteristics of the aggregate pre-
cursor, which is influenced by environmental condi-
tions. With a-lactalbumin, for example, removal of
Ca
2+
or the presence of Zn
2+
induces rapid formation
of amorphous aggregates, whereas lowering the pH to
2.0 (to give the so-called A state) or reducing three of
its four disulfide bonds (to give 1SS-a-lactalbumin)
leads to the formation of amyloid fibrils [21]. The first
set of conditions gives rise to a highly unstable molten
globule state with a relatively rigid conformation,

whereas the A state and 1SS-a-lactalbumin both have
considerable conformational flexibility. Furthermore,
the efficiency of interaction between sHsps and these
partially folded species varies greatly and occurs via
different binding modes, depending on the conforma-
tional properties of the target protein [19,22,23].
During amyloid fibril formation, a protein will pro-
ceed from its initial monomeric state through a series
of aggregation states, e.g. the amyloidogenic nucleus
and other prefibrillar intermediates, culminating in
formation of the mature fibril [24]. The increasing
complexity of these structures is paralleled by confor-
mational changes, often irreversible, which the protein
undergoes along its amyloid pathway. These may
include conversion to a partially folded intermediate,
partial proteolysis, b-sheet formation, ordered intermo-
lecular association and the intertwining of two or more
protofilaments [24]. Because pathological significance
has been ascribed to the early soluble intermediates
rather than mature fibrils [25–28], one approach to the
treatment of amyloid diseases involves the develop-
ment of inhibitors that not only inhibit amyloidogene-
sis in its very early stages by interacting with partially
folded or very early oligomeric species, but also result
in a product which is nontoxic or biodegradable.
The role of sHsps in amyloid fibril diseases is contro-
versial. They are upregulated in these disease states and
are known to interact with partially folded proteins.
However, while inhibiting fibril formation, aB-crystal-
lin stabilizes prefibrillar neurotoxic forms of Ab-peptide

[14,29]. By contrast, aB-crystallin interacts with a-syn-
uclein to form large nonfibrillar aggregates, implying
that it can redirect a-synuclein from a fibril-forming
pathway towards an amorphous aggregation pathway,
thus reducing the amount of physiologically stable fibril
in favour of easily degradable amorphous aggregates
[16]. There are no data available on the effect of sHsps
on the cytotoxicity of prefibrillar a-synuclein aggre-
gates, however, the unrelated Hsp70 molecular chaper-
one reduces the toxicity of prefibrillar and misfolded
detergent-insoluble a-synuclein species [30,31].
In this study, we investigated the kinetics of interac-
tion of a-crystallin with amyloid-forming a-synuclein
and j-casein. a-Synuclein is a 14.4 kDa presynaptic
protein of unknown function, which is a main compo-
nent of Lewy bodies, the amyloid-rich proteinaceous
deposits in Parkinson’s disease. It is intrinsically
unstructured, but adopts a predominantly b-sheet con-
formation during the formation of cytoplasmic amyloid
fibrils in neurons [32]. j-Casein is one of the principal
proteins of bovine milk, which together with others
caseins (e.g. a
s
and b), form a unique micellar complex
serving as a calcium phosphate transporter. Upon
reduction of its intermolecular disulfide bonds, j-casein
readily forms fibrils at physiological pH over a wide
range of temperatures [33,34], thus providing an excel-
lent model for studying the temperature-dependent
interaction of amyloid-forming proteins with sHsps. In

particular, we examined the effects of temperature on
the fibrillation rate of j-casein and a-synuclein and the
efficiency of a-crystallin to suppress this aggregation. In
addition, we investigated the interaction of aB-crystal-
lin with a-synuclein in real time using dual polarization
interferometry (DPI) [35,36], a new analytical method
A. Rekas et al. a-Crystallin and amyloid fibril formation
FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS 6291
for studying protein interaction under physiological
conditions. With regards to amyloid fibril formation, it
enables the real time study of both fibril elongation and
the initial nucleation processes. By monitoring the
thickness, average density and mass of the protein
deposition layer, it was possible to record in greater
detail the kinetics of a-synuclein aggregation in both
the absence and presence of aB-crystallin, thereby
revealing the stage at which aB-crystallin interacts with
a-synuclein to inhibit its fibril formation.
Results
Temperature dependence of a-crystallin chaperone
activity against fibril-forming target proteins
The enhanced ability of a-crystallin at elevated temper-
ature, i.e. 30 °C and above, to prevent the aggregation
and precipitation of amorphously aggregating target
proteins has been well characterized [6–8]. The aim of
our study was to determine whether a similar tempera-
ture dependency existed in the ability of a-crystallin to
prevent amyloid fibril formation by either j-casein or
a-synuclein.
j-Casein

As is apparent from transmission electron microscopy
(TEM) images (Fig. 1), disulfide-reduced j-casein
formed fibrils both at 37 and 50 °C (13.4 ± 2.2 nm in
diameter), although a difference in their supramolecu-
lar morphology was evident: at 37 °C fibrils were well
separated, whereas at 50 °C there was a tendency to
further associate to form large conglomerates of tan-
gled fibrils. The length of the fibrils varied greatly, but
on average, the fibrils formed at 37 °C were shorter
(101.1 ± 49.6 nm) than those formed at 50 °C
(148.0 ± 88.3 nm), including a larger number of small
fragments (up to 20 nm in length) at the lower temper-
ature. At 37 °C, the presence of a-crystallin (up to
1 : 1 molar ratio) had little effect on the extent of fibril
formation, with longer fibrils of 94.3 ± 28.7 nm,
although the overall polydispersity was reduced. A
large number of short prefibrillar j-casein species
( 20 nm) were also present, in addition to the spheri-
cal aggregates (14–17 nm in diameter) characteristic of
a-crystallin. At 50 °C, a-crystallin caused a noticeable
reduction in the number of fibrils, including prefibrillar
species, but the average length of mature fibrils
remained large (152.2 ± 68.7 nm).
The fluorescence of j-casein-bound thioflavin T
(TFT) at 37 and 50 °C showed a sigmoidal time course
(Fig. 2A) typical of nucleation-dependent fibril forma-
tion [37]. The initial lag phase corresponds to the
formation and accumulation of oligomeric prefibrillar
partially folded intermediates that do not bind TFT
[38]. The subsequent increase in fluorescence intensity

represents elongation of the fibril [39] with a stacked
b-sheet conformation.
Under stable environmental conditions (e.g. constant
temperature), TFT fluorescence can be reliably used to
quantify the amount of stacked b sheet, and thus moni-
tor the kinetics of fibrillation. However, during experi-
ments performed at higher temperatures, a decrease in
TFT fluorescence was observed, which suggests that
either binding of TFT by proteins or the efficiency
of fluorescence are temperature dependent. For this
reason, the time course of TFT fluorescence upon
interaction with j-casein may reflect other tempera-
ture-dependent processes in addition to the formation
of amyloid fibrils. For example, at higher temperatures
(45–60 °C), the magnitude of TFT fluorescence (maxi-
mum intensity value) in the presence of j-casein alone
was much lower than at 30–37 °C (Fig. 2A), despite a
comparable number of fibrils shown by electron
microscopy. Moreover, at higher temperatures, there
was a decrease in TFT fluorescence after reaching a
maximum value (Fig. 2A; 50 °C data), which may
arise from the aggregation of fibrils into large con-
glomerates and the possible obstruction of TFT bind-
ing sites (Fig. 1). Thus, fibrillation rates (as depicted in
Fig. 2B) were reliably estimated from the TFT binding
500 nm
-casein, 37°C
-casein, 50°C
-casein + -
crystallin 37°C

-casein + -
crystallin, 50°C
Fig. 1. TEM images of reduced j-casein species formed at 37 and
50 °C in the absence and presence of an equimolar amount of
a-crystallin. Images acquired at ·40 000 magnification show a
higher level of suppression of j-casein fibrillation by a-crystallin at
50 °C than at 37 °C.
a-Crystallin and amyloid fibril formation A. Rekas et al.
6292 FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS
data using only the initial period, when the increase in
fluorescence was exponential and concentration depen-
dent (the length of this exponential period also varied
with temperature and was chosen by careful examina-
tion of fitting parameters).
j-Casein aggregation kinetics depended on tempera-
ture and the presence of the inhibitor, a-crystallin. The
rate constant for the increase in j-casein TFT binding
with time increased with temperature (Fig. 2A,B).
However, as assessed by TFT binding, the presence of
a-crystallin significantly suppressed fibril formation by
j-casein (Fig. 2A,B). At 30 °C, the fibril formation
rate was slowest and a-crystallin significantly reduced
the number of fibrils without changing the rate of fibril
formation (Fig. 2B). Percentage-wise, a-crystallin was
least effective at 30–33.5 °C, where temperature-depen-
dent increases in the rate of fibril formation were not
compensated by a concomitant increase in the ability
of a-crystallin to suppress it. Above 33.5 °C, however,
the rate of fibril formation by j-casein increased with
temperature, and so did the relative efficiency of

a-crystallin to inhibit fibril formation (Fig. 2).
Amyloid fibril elongation is known to be a first-
order reaction [37,39] (A. Rekas, unpublished data on
a-synuclein and j-casein). From the Arrhenius law,
we have: ln(k
app
) ¼ – E
A
⁄ RT + ln(A). The activation
energy of fibril formation (E
A
) was calculated
(Table 1) as the slope of the straight line fitted to a
plot of ln(k
app
) versus 1 ⁄ T, where T is temperature in
K, R is the gas constant and A is the frequency (or
pre-exponential) factor, expressed in the same units
as the apparent first-order rate constant, k
app
. For
reduced j-casein only, E
A
was 35.5 ± 1.1 kcalÆmol
)1
,
showing strong temperature dependence of the rate
constants (R
2
¼ 0.995). a-Crystallin reduced the acti-

vation energy for j-casein fibril elongation, e.g. for a
TFT binding @ 50oC
0
100
200
300
400
500
600
700
0 5 10 15
time (hours)
a.u.
TFT binding @ 25
o
C
0
200
400
600
800
1000
1200
1400
1600
0510
time (hours)
a.u.
k-cas preinc @ 25 deg
k-cas preinc @ 40 deg

k-cas preinc @ 60 deg
TFT binding @ 37oC
0
1000
2000
3000
4000
5000
6000
7000
0 5 10 15
a.u.
0.10
1.00
10.00
100.00
30 40 50 60
temperature [
o
C]
kapp [s
-1
]
0
50
100
0
0.25 0.5
κ-cas
κ-cas+0.25xα-cr

κ-cas+0.5xα-cr
κ-cas+1xα-cr
κcasein κcas+0.25xαcrys κcas+0.5xα-crys κ cas+1xα-crys
A
B
C
0 0.4 0.8
500
100
250
50
0
0 0.4 0.8
0
1
1
Fig. 2. Temperature dependence of j-casein fibril formation under reducing conditions. (A) Real-time TFT fluorescence data at 37 and 50 °C.
(B) Growth-rate constants with temperature in the presence and absence of a-crystallin at the indicated molar ratios. (C) The effects of pre-
incubation of j-casein at 25, 40 and 60 °C on its fibrillation potential.
Table 1. Comparison of activation energy (E
A
) and frequency factor (A) values for j-casein fibril elongation under reducing and nonreducing
conditions. a-Crystallin, especially at higher concentrations (0.5 : 1 and 1 : 1 w ⁄ w ratios to j-casein), reduced the activation energy and fre-
quency factor for j-casein fibril formation.
j-casein +0.25· a-crystallin +0.5· a-crystallin +1.0· a-crystallin
Reduced E
A
(kcalÆmol
)1
) 35.5 ± 1.1 13.8 ± 4.9 18.3 ± 2.6 14.9 ± 2.0

A (h
)1
) range 1.4 · 10
24
)4.2 · 10
25
9.8 · 10
5
)4.8 · 10
12
4.0 · 10
10
)1.2 · 10
14
3.3 · 10
8
)2.0 · 10
11
Native E
A
(kcalÆmol
)1
) 25.7 ± 1.9 24.9 ± 1.9 1.7 ± 4.4 )0.4 ± 1.2
A (h
)1
) range 1.9 · 10
16
)9.2 · 10
18
6.4 · 10

15
)3.2 · 10
18
5.3 · 10
)3
)5.9 · 10
3
1.2 · 10
)2
)5.3 · 10
)1
A. Rekas et al. a-Crystallin and amyloid fibril formation
FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS 6293
1 : 1 molar ratio of j-casein ⁄ a-crystallin, E
A
was
14.9 ± 2.0 kcalÆmol
)1
(R
2
¼ 0.902). Likewise, the
parameter A which is related to the frequency of inter-
actions between the molecules, decreased in the pres-
ence of a-crystallin.
Exposure of j-casein to higher temperatures for
15 min caused a slight decrease in its subsequent
fibrillation level when incubated in the presence of
reducing agent at 25 °C (Fig. 2C), although the
changes in the rate constants were not significant:
(1.85 ± 0.09) · 10

)1
Æs
)1
, (1.86 ± 0.15) · 10
)1
Æs
)1
and
(1.79 ± 0.11) · 10
)1
Æs
)1
after preincubation at 25, 40
and 60 °C, respectively. The initial lag times increased
from 10 to 13 to 20 min for 25, 40 and 60 °C preincu-
bation temperature, respectively. These differences
indicate that some small irreversible structural changes
occur to j-casein with temperature, but they are not
sufficient to explain the reduction in maximum fluores-
cence intensity and the increase in fibrillation rate that
was observed for fibril formation at higher tempera-
tures in the experiments described above.
Fibril formation by j-casein under nonreducing con-
ditions (hereafter referred to as ‘native’ j-casein) was
also examined over the temperature range of 30–55 °C.
In the absence of reducing agent, the process of fibril-
lation proceeded more slowly, especially at higher tem-
peratures (Fig. 3), than under reducing conditions
(Fig. 2). At the same time, the overall efficiency of
a-crystallin to prevent fibril formation was lower, with

only equimolar amounts of a-crystallin showing signifi-
cant inhibition below 45 °C (Fig. 3B). As seen with the
reduced protein, the ability of a-crystallin to suppress
fibril formation by native j-casein increased with
temperature, as indicated by a significant reduction in
activation energy for fibril elongation (E
A
) which
at a a-crystallin ⁄ j-casein ratio of 1 : 1 (w ⁄ w)
was )0.4 ± 1.2 kcalÆmol
)1
, compared with 25.7 ±
1.9 kcalÆmol
)1
for j-casein only (Table 1). In effect, at
high concentrations of a-crystallin, the temperature
dependence of j-casein fibril formation was abrogated
by the inhibitory action of a-crystallin.
a-Synuclein
TFT fluorescence data showed differences in the fibril-
lation kinetics of a-synuclein at various temperatures
(Fig. 4A). From these data, it is evident that the rela-
tive ability of aB-crystallin to inhibit a-synuclein
aggregation increased with temperature. Also, the max-
imum fluorescence over time was relatively unchanged
upon increasing the temperature from 37 to 45 °C, but
was significantly lower at 60 °C (Fig. 4A), as observed
with j-casein at higher temperature (Figs 2 and 3).
The temperature dependence of aB-crystallin’s abil-
ity to suppress fibril formation, as shown by TFT

binding data, was supported by TEM. At 37 and
60 °C, a-synuclein, by itself, formed fibrils of compa-
rable length and morphology, however, in the pres-
ence of aB-crystallin (at a 1 : 1 molar ratio) fibril
formation at 60 °C was almost completely inhibited,
while only partial suppression was achieved at 37 °C
(Fig. 4B).
The reduction in TFT fluorescence at higher temper-
atures was demonstrated for preformed fibrils of
j-casein and a-synuclein. A constant number of fibrils
showed a 36% reduction in TFT fluorescence over the
temperature range 28–60 °C for j-casein, and 39%
reduction for a-synuclein between 25 and 52.5 °C
(Fig. 4C).
B
A
Fig. 3. Temperature dependence of j-casein fibril formation under
nonreducing conditions in the absence and presence of a-crystallin.
(A) Plots showing real-time TFT fluorescence data at 37 and 55 °C.
(B) Changes in fibril growth-rate constants with temperature at the
indicated molar ratios.
a-Crystallin and amyloid fibril formation A. Rekas et al.
6294 FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS
DPI study of the suppression of a-synuclein
aggregation by aB-crystallin
DPI was used to monitor real-time a-synuclein aggrega-
tion and the effect of aB-crystallin on this, particularly
at the very early stages of this process. In a DPI mea-
surement, the average layer density decreases during
fibrillar-type aggregation because the initial dense pro-

tein ‘monolayer’ on the surface remains attached, but
the subsequent protein deposition occurs by elongating
of (pre)fibrillar species, rather than random adherence
of nonfibrillar material. This has been observed in DPI
examination of the aggregation of other fibril-forming
proteins, i.e. the Alzheimer’s amyloid b peptide and the
familial mutants (A30P and A53T) of a-synuclein
(field-sensors.com/articles/).
The signal responses stabilized  10 min after the
injection of a-synuclein alone into channel 3 and the
resolved data showed a deposition of a protein layer
of thickness 4.114 nm, density 0.652 gÆcm
)3
and mass
2.531 ngÆmm
)2
. Over the next 4 h, a steady decrease in
layer density, and an increase in layer thickness and
mass were observed (Fig. 5). During the maturation
process, these values gradually changed, showing that
aggregation proceeded steadily. Specifically, after
60 min the protein layer thickness increased by
0
000
1
0002
0003
0004
0005
000

6
00
07
0008
TFT fluorescence
0
0
0
0
2
0004
0
0
0
6
0008
0
00
0
1
thioflavin T fluorescence
0
0
001
000
2
000
3
00
0

4
0005
02
10
90
60
3
0
2
)setu
ni
m
(em
i
t
thioflavin T fluorescence
α
-syn
α
-syn+0.5xα-crys
α
-syn+1xα-crys
73
o
C
54
o
C
06
o

C
200 nm
AB
0
002
0
0
5
2
000
3
0053
0004
005
4
0005
0
605040302
(
e
rutarepmet
o
)C
TFT fluorescence (a.u.)
κ
-casein
α
-synuclein
C
Fig. 4. (A) Temperature dependence of a-synuclein fibril formation in the absence and presence of ab-crystallin. Bar graphs show TFT fluo-

rescence data at selected time points and 37, 45 and 60 °C. Molar fractions of aB-crystallin over a-synuclein are indicated. (B) Comparison
of electron micrographs of a-synuclein species in the absence and presence of aB-crystallin (1 : 1 molar ratio) incubated for 4 h at 37 or
60 °C. (C) Temperature dependence of TFT fluorescence for 1 mgÆmL
)1
j-casein (closed symbols) and 2 mgÆmL
)1
a-synuclein (open sym-
bols); the decrease in TFT intensity accounts for the lower TFT fluorescence levels at higher temperatures shown in Figs 2A, 3A and 4A.
A. Rekas et al. a-Crystallin and amyloid fibril formation
FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS 6295
0.0746 nm on channel 3, the mass increased by
0.02 ngÆmm
)2
and the density decreased
by 0.0072 gÆcm
)3
. After 2.5 h, the thickness increased
by 0.212 nm, the mass increased by 0.05 ngÆmm
)2
and
the density decreased by 0.0167 gÆcm
)3
, from the start
of the experiment. By the end of the measurement, the
thickness had increased by 0.342 nm, the mass had
increased by 0.10 ngÆmm
)2
and the density had
decreased by 0.024 gÆcm
)3

.
By contrast, on channel 1, where aB-crystallin
was injected together with a-synuclein, the thickness,
density and mass of the protein layer were essentially
unchanged during the entire experiment (Fig. 5), i.e. the
thickness of the layer decreased by 0.014 nm, the layer
density decreased by only 0.0086 gÆcm
)3
, and the mass
decreased by 0.03 ngÆmm
)2
. The process of a-synuclein
aggregation at 25 °C without agitation is relatively slow,
so the thickness did not increase greatly over time.
In addition to demonstrating the ability of DPI to moni-
tor the aggregation of a-synuclein, this experiment
showed that the interaction of aB-crystallin with a-syn-
uclein takes place immediately after combining solutions
of both proteins and prevents formation of prefibrillar
nuclei by a-synuclein, i.e. aB-crystallin interacts with
a-synuclein early along its aggregation pathway.
To summarize, initial nucleation took place immedi-
ately after the protein was bound to the sensor surface
as the thickness and mass of the protein layer started
to increase with a simultaneous decrease in density
after only 10 min. In a parallel experiment (not
shown), no change in TFT binding was observed after
24 h incubation of a-synuclein in the absence or pres-
ence of a 0.5 molar amount of aB-crystallin at 25 °C
without agitation, i.e. under experimental conditions

mimicking those of DPI. Therefore, the DPI data refer
to prefibrillar a-synuclein aggregation.
Species specificity of aB-crystallin interaction
with a-synuclein
From the DPI results (Fig. 5), it is apparent that
aB-crystallin interacts with a-synuclein early during its
aggregation pathway (i.e. at the nucleation or proto-
fibril stage). Additional experiments were therefore
undertaken to determine whether aB-crystallin was as
effective at suppressing further fibril formation by
more advanced fibrillar forms of a-synuclein.
Time course of thioflavin T binding
As expected, in the absence of aB-crystallin, an
increase in TFT fluorescence was observed for incu-
bated a-synuclein. Fibril formation by a-synuclein, as
indicated by this increase in fluorescence, was sup-
pressed upon the addition of aB-crystallin [16]
(Fig. 6A). Interestingly, this effect was observed not
only when both proteins were present in the sample
from the beginning of incubation, but also in samples
containing a significant number of amyloid fibrils
(before addition of aB-crystallin at time points
between 25 and 65 h). Under these conditions,
aB-crystallin prevented, but did not reverse, further
fibril formation (i.e. it had no capacity to disassemble
existing fibrils), as visible from the stabilization of the
level of TFT fluorescence.
2
5.2
3

5.3
4
5.4
5
4
32
1
0
Layer thickness (nm)
55.0
85.0
16
.0
46.0
76.
0
7
.0
4
3210
Layer density (g/cm
3
)
3
.
1
55.
1
8.1
50

.
2
3
.2
55.
2
8.2
432
1
0
)s
r
uoh(emi
t
Mass (ng/mm
2
)
α
+nys-
α
sy
r
c-B
α
n
i
e
lcunys
-
α

+nys-
α
s
yr
c
-B
α
nielcu
n
y
s
-
α
+nys-
α
syrc-B
α
nielcunys-
Fig. 5. The DPI data obtained from channel 3 (a-synuclein only;
black) and channel 1 (a-synuclein + aB-crystallin; grey), showing
a-synuclein physisorption and aggregation. a-Synuclein was at
3.5 mgÆmL
)1
and aB-crystallin at 2.5 mgÆmL
)1
. The resolved traces
of thickness, density and mass are depicted in individual panels.
The data shown are from the time of signal stabilization following
injection of protein solutions onto the sensors thermostated at
25 °C.

a-Crystallin and amyloid fibril formation A. Rekas et al.
6296 FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS
Although these data do not exclude the possibility
of an interaction between aB-crystallin and fibrillar
a-synuclein, they are consistent with the DPI results
showing that aB-crystallin interacts with monomeric or
nucleated a-synuclein prior to it being incorporated
into the growing a-synuclein fibril, and in this way pre-
vents further fibril growth. TEM images of a-synuclein
species at different stages of its fibril formation (in the
absence of aB-crystallin) are also consistent with this
proposal. Small globular a-synuclein species were
found throughout the entire time course of fibril for-
mation (Fig. 6B), including the micrograph at ‘0 h’
(which was actually about 15 min after dissolution of
the protein while being kept on ice); and during the
plateau phase (after 1 week of incubation). Consider-
ing their size of 13–19 nm in diameter, which matches
the diameter of a-synuclein fibrils, it is likely that these
species are prefibrillar intermediates.
Interaction between j-casein and a-crystallin
investigated by size-exclusion HPLC, 8-anilino-
naphthalene 1-sulfonate binding and NMR
spectroscopic studies
The interaction and complex formation of destabilized
j-casein with a-crystallin, after mixing both proteins in
the presence or absence of dithiothreitol, was investi-
gated by size-exclusion HPLC. The interaction of
a-crystallin with reduced j-casein was also investigated
by 8-anilinonaphthalene 1-sulfonate (ANS) binding

and NMR spectroscopy, and compared with an analo-
gous interaction with native j-casein. The absence of
shaking during incubation resulted in j-casein species
that did not bind TFT and were therefore nonfibrillar.
Size exclusion HPLC
Incubation of equal masses of j-casein and a-crystallin
in solution at 37 °C for 4 h led to partial formation of
a high molecular mass complex between these two pro-
teins of  1300 kDa, as shown by size-exclusion HPLC
(Fig. 7A). In its native (nonreduced) state, j-casein
exists as a large species which eluted at 5 h 45 min
from the column, the same elution time as a-crystallin,
corresponding to  830 kDa. However, the elution
time of the a-crystallin + j-casein mixture was shifted
to 5 h 28 min, implying interaction between the two
proteins which led to a complex of larger mass. Reduc-
tion of j-casein’s intermolecular disulfide bonds led to
the appearance a very large aggregate ( 6800 kDa) at
an elution time of 4 h 38 min. In the main, the pres-
ence of a-crystallin significantly decreased formation of
this large aggregate. As a result of the interaction of
a-crystallin with reduced j-casein, a complex of similar
mass (1500 kDa) to the one with native j-casein was
observed with an elution time of 5 h 26 min.
A
B
0
50
100
150

200
250
300
0 50 100 150
time (hours)
thioflavin T fluorescence
a-syn
a-syn+aB
a-syn+aB 25h
a-syn+aB 49h
a-syn+aB 65h
-syn
-syn+ B
-syn+ B 25h
-syn+ B 49h
-syn+ B 65h
72 hrs
200 nm
168 hrs
0 hrs
Fig. 6. Time course of amyloid fibril forma-
tion by a-synuclein (125 l
M) in the absence
and presence of aB-crystallin (62.5 l
M). (A)
TFT binding data. aB-crystallin was added to
the incubated a-synuclein samples (black
diamond) at the beginning of the experiment
(black squares) and at later time points, i.e.
25 h (grey triangle), 49 h (black circle) and

65 h (grey diamond). The increase in TFT
fluorescence was monitored as described
previously [16]. (B) TEM images of
a-synuclein species in the absence of
aB-crystallin at the indicated times from the
beginning of incubation at 37 °C. Images
acquired at ·60 000 magnification reveal
that small globular protein aggregates
(oligomeric intermediates, indicated by
arrows) are present alongside fibrils at all
stages of the fibril formation time course.
A. Rekas et al. a-Crystallin and amyloid fibril formation
FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS 6297
ANS binding
The level of ANS fluorescence emission (Fig. 7B)
indicated that reduced j-casein exposed much more
clustered hydrophobicity to solution than nonreduced
j-casein, which is consistent with greater unfolding of
the protein following disulfide bond reduction. Both
reduced and nonreduced j-casein, and a-crystallin,
Exposed hydrophobicity
0
500
1000
1500
2000
2500
3000
20 30 40 50 60 70
ANS fluorescence (a.u.)

a-crystallin
k-casein non-red
k-casein reduced
k-casein red. + a-crys
a-cryst
+
-cas red (theor)
HPLC elution profile
A
C
B
0
20000
40000
60000
80000
100000
33.544.555.566.577.58
elution time (min)
A(280)
k-cas
a-crystallin
cas reduced
-
-
-
ca
s
-
-

-
-
-
crys
cas red crys
14.6 kD67 kD669 kD2000 kD
temp (°C)
Fig. 7. Interaction of a-crystallin with j-casein. (A) Size-exclusion HPLC profiles of j-casein (native and reduced), a-crystallin and their mix-
tures. All proteins at 10 mgÆmL
)1
were incubated at 37 °C for 4 h prior to chromatography. a-Crystallin decreases the size of reduced
j-casein aggregates and also complexes with nonreduced j-casein. Elution times of blue dextran (2000 kDa), thyroglobulin (669 kDa), BSA
(67 kDa) and lysozyme (14.6 kDa) are indicated. (B) Maximum ANS fluorescence when bound to j-casein, a-crystallin (both proteins at
0.3 mgÆmL
)1
) and their mixtures, recorded in the temperature range from 25 to 65 °C. At lower temperatures, the interaction between these
two proteins (circles) results in greater exposure of hydrophobic regions than the sum of fluorescence values of both component proteins
(stars). (C) Superimposed 2D
1
H NMR TOCSY spectra of the NH to a,b,c region of j-casein (red), a-crystallin (blue) and their mixtures (black)
acquired at 37 °C. Each protein was dissolved in 10 m
M sodium phosphate pH 7.2, 10% D
2
O, at a concentration of 2 mgÆmL
)1
. The upper
panel shows spectra with reduced j-casein, the lower with native (nonreduced). After combining native j-casein with a-crystallin, additional
cross-peaks were observed, which are circled in green. a-Crystallin had little effect on the reduced target protein (a relatively stable unfolded
state), but additional cross-peaks were observed for the mixture under native (nonreducing) conditions.
a-Crystallin and amyloid fibril formation A. Rekas et al.

6298 FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS
showed a decrease in ANS fluorescence with increas-
ing temperature. This implies a decrease in the
amount of exposed hydrophobicity (due to self-associ-
ation) and ⁄ or a decrease in fluorescence emission
efficiency with temperature, as was observed for pro-
tein-bound ANS in the absence of conformational
changes [40]. Notwithstanding, the interaction
between j-casein and a-crystallin led to higher level
of surface hydrophobicity, as the mixture of reduced
j-casein and a-crystallin had a higher ANS-binding
level than the sum of both components (Fig. 7B).
This effect was largest at lower temperatures (25 °C)
and decreased upon heating to 70 °C, at which point
the fluorescence of the mixture was equal to the sum
of its components. This difference between theoretical
values and those of the nonreduced j-casein +
a-crystallin mixture was slightly larger than under
reduced conditions (not shown).
1
H-NMR spectroscopy
Cross-peaks from the NH to aliphatic proton regions
of
1
H 2D NMR TOCSY spectra of j-casein, a-crystal-
lin and their mixture are shown in Fig. 7C, for reduced
and native j-casein (upper and lower panels, respec-
tively). As expected, spectra of the a-crystallin aggre-
gate show only a few cross-peaks belonging to the
highly flexible C-terminal extension in both subunits of

10–12 amino acids [41,42]. Reduced j-casein showed a
significant degree of flexibility compared with the
native species, as indicated by a large number of
intense cross-peaks. Addition of a-crystallin to j-casein
caused some additional cross-peaks to be observed,
which was particularly pronounced in the case of the
native j-casein and a-crystallin mixture, where signifi-
cant conformational flexibility was indicated by the
appearance of additional cross-peaks.
Discussion
The temperature dependence of the kinetics of fibril
formation by Ab peptide [33,39] and insulin [37] has
been described previously. The time course of fibril
formation, as monitored by TFT binding, follows a
sigmoidal curve. The prefibrillar nuclei (early oligo-
meric species) do not bind TFT. They form during the
lag time, which is followed by the fibril elongation
phase corresponding to an increase in the dye’s fluores-
cence. The subsequent plateau phase is associated with
a decrease in the concentration of small species, or the
aggregation and precipitation of fibrils [37]. The kinet-
ics of these three stages of the fibril formation process
are temperature dependent [37–39,43].
In this study, the rate of fibril formation of both
j-casein and a-synuclein increased with temperature,
as monitored by TFT binding. In the presence of
a-crystallin, the initial lag phase was longer, which
indicates that a-crystallin slowed the formation of pre-
fibrillar intermediates. The TEM data are consistent
with this conclusion. a-Crystallin undergoes a struc-

tural transition at  45 °C which leads to greater
unfolding and enhanced chaperone action against
amorphously aggregating target proteins [44,45]. This
behaviour may contribute to a-crystallin’s enhanced
ability to prevent fibril formation at higher tempera-
tures.
Our data on j-casein showed an exponential depen-
dence of the fibril formation rate on temperature.
Thus, the rate constants follow Arrhenius’ law, which
is consistent with the temperature dependence of fibril
elongation rates of the Ab peptide [39]. In addition to
decreasing the rates of fibril formation at all tempera-
tures for reduced and native j-casein, a-crystallin
decreased both the activation energy and the frequency
constant of this process. This suggests that the temper-
ature-dependent inhibition of j-casein fibrillation by
a-crystallin is a function of both ‘activating’ the
chaperone ability of a-crystallin, and of the effects of
a-crystallin on j-casein, which have not been, as yet,
described. If this mechanism of interaction occurs
in vivo, it may have important implications in the
design of chaperone-based therapeutics against amy-
loid diseases.
Fibril formation by j-casein in the presence of an
inhibitor protein, a-crystallin, is a complex process.
Possible components of this reaction include the disso-
ciation of large a-crystallin and j-casein oligomers into
smaller species, binding of a-crystallin to j-casein, con-
formational alteration of j-casein and ⁄ or a-crystallin
upon their interaction, dissociation of the complex and

subsequent conformational changes (e.g. refolding) of
j-casein. The resultant E
A
and k values are reflective
of the entire process (Table 1). Molecular collision
rates increase with temperature, and so does the disso-
ciation rate of a-crystallin oligomers. In addition, the
conformational flexibility of a-crystallin also increases
with temperature [2,6,8,11–13], making it potentially
more efficient to interact with j-casein and form a
transient complex. This is consistent with the observed
enhancement of the inhibitory effect of a-crystallin on
the rate of j-casein fibrillation at higher temperatures.
However, our NMR and fluorescence data also indi-
cate a greater unfolding of j-casein upon its interac-
tion with a-crystallin. Such partially unfolded j-casein
molecules, when released from the complex with
a-crystallin would be susceptible to association with
A. Rekas et al. a-Crystallin and amyloid fibril formation
FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS 6299
other j-casein molecules. In this way, a-crystallin may
inhibit fibril formation by binding to the target pro-
tein, but in the process ‘activate’ it to form fibrils. This
speculation is supported by our observation that
a-crystallin both decreases j-casein fibril formation
rate and lowers the activation energy of this process.
sHsps recognize relatively slowly aggregating pro-
teins, which exclude short-lived states that are present
only transiently on the rapid protein folding ⁄ unfolding
pathway [1,2]. Furthermore, sHsps do not interact with

relatively stable partially folded intermediates (i.e.
whose unfolding rate is very slow), or stable unfolded
proteins [1,46]. The conformational state of the
reduced compared with the nonreduced j-casein inter-
mediate is also a potentially contributing factor to the
difference in fibrillation rate between the two forms of
j-casein and interaction with a-crystallin. An increased
rate of protein fibrillation does not always result in a
greater efficiency of a-crystallin to prevent this. For
example, molecular crowding increased the rate of
fibril formation by a-synuclein, but also decreased the
chaperone efficiency of aB-crystallin [16]. Furthermore,
molecular crowding affects the rate of subunit
exchange of a-crystallin, where a reduced rate corre-
lates with decreased chaperone ability [47]. The situa-
tion is different in the case of increased temperature,
as this variable accelerates a-crystallin’s subunit
exchange and affects its conformation, two properties
that are strongly related to its chaperone action
[2,6,8,11–13]. In conclusion, the increased rate of target
protein fibrillar aggregation in conjunction with
changes in conformation and dynamics of a-crystallin
molecules, contribute to the increased efficiency of sup-
pression by a-crystallin of fibril formation at higher
temperatures.
At higher temperatures, the maximum TFT fluores-
cence intensity of j-casein and a-synuclein did not
correlate with the increased fibrillation rate, i.e. fluo-
rescence intensity was lower at temperatures above
45 °C and decreased sharply after reaching a maxi-

mum. However, TEM showed that the number of
j-casein fibrils after 12 h incubation at 37 °C was less
than that for j-casein fibrils incubated at 50 °C for the
same period (Fig. 1). From these results, it can be
inferred that TFT fluorescence cannot always be used
quantitatively to assess the extent of fibril formation.
At higher temperatures, as assessed by TEM (Fig. 1)
and temperature-dependent TFT fluorescence in the
presence of fibrils (Fig. 4C), the decrease in TFT fluo-
rescence over time after reaching a maximum was not
because of nonfibrillar aggregation of j-casein, which
might have reduced the number of fibrils in favour of
amorphous aggregates. The decrease can be attributed,
at least partially, to a decrease in TFT fluorescence
yield at high temperatures. A possible explanation for
this phenomenon is that thermal motion in j-casein
fibrils is enhanced at higher temperatures and disrupts
the rigid conformation of bound TFT molecules
required for high fluorescence yield [48]. By contrast,
at 45 ) 60 °C, there was a decrease in TFT fluores-
cence following a sigmoidal increase. The TEM images
(Fig. 1) showed many well-distinguished fibrils formed
at 37 °C, whereas those formed at 50 °C, although no
less numerous, tended to assemble in large conglomer-
ates of tangled fibrils which may have reduced TFT-
binding capacity [49], and this process is likely to be
time-dependent at a set temperature. It may represent
a phenomenon similar to the temperature-induced ‘pre-
cipitation’ of insulin fibrils [37,38,43]. In addition,
some degree of conformational change in j-casein

before its assembly into fibrils may be temperature
dependent, as indicated by our experiments in which
preincubation of j-casein at 60 °C limited its fibrilla-
tion potential at 25 °C in comparison with the protein
preincubated at room temperature (Fig. 2C). Interest-
ingly, this effect of decreased fluorescence at higher rel-
ative to lower temperatures was found to be less
pronounced in the presence of a-crystallin than in its
absence, despite its inhibitory effect on the amount of
fibril formation. This correlated with the absence of
fibril conglomerates in the presence of a-crystallin in
TEM images (Fig. 1), which would inhibit binding of
TFT.
The size-exclusion HPLC data indicate complex
formation between prefibrillar j-casein (native or
reduced) and a-crystallin at 37 °C. The size of this
complex is less than that of reduced j-casein aggre-
gates formed in the absence of the chaperone, which is
consistent with the data on RCM j-casein and
aB-crystallin [50]. NMR spectra of native j-casein
acquired at 37 °C in the presence of a-crystallin
showed additional conformational disorder, which
most likely arose from j-casein, because a-crystallin
does not increase its flexibility upon interaction with
target proteins [1,41,42]. The incubation conditions
(buffer and temperature) for the NMR experiments
were the same as for the HPLC studies, which showed
complex formation between both proteins, so the
observed increase in conformational disorder is likely
to be caused by j-casein unfolding upon interaction

with a-crystallin, rather than any dissociation of the
high-mass species. The interaction of a-crystallin with
native j-casein appears to cause more structural
changes than interaction with reduced j-casein,
indicating a propensity of the native j-casein to
undergo structural changes (eventually leading to fibril
a-Crystallin and amyloid fibril formation A. Rekas et al.
6300 FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS
formation) which are recognized by the molecular
chaperone a-crystallin, whereas any conformational
changes caused by a-crystallin’s interaction with
reduced j-casein are smaller, possibly because of
greater unfolding of the target protein caused by disul-
fide bond reduction. ANS-binding data provided fur-
ther evidence for unfolding of reduced j-casein and
exposure of clustered hydrophobicity upon its interac-
tion with a-crystallin. This difference in interaction
mode of a-crystallin with native and reduced j-casein
is also reflected in that activation energies and fre-
quency factors for the reduced protein do not follow a
decreasing trend in the presence of increasing amounts
of a-crystallin (Table 1), as is the case with native
j-casein.
Many methods of monitoring amyloid formation in
real time are sometimes nonspecific (e.g. they also rec-
ognize nonfibrillar proteins of high b-sheet assembly)
or depend on structural and environmental variables
(e.g. staining with amyloidophilic dyes, which are
insensitive to prefibrillar stages of aggregation and
whose binding to cross b-sheet structures depends on

temperature or agglomeration of fibrils, as discussed
above). By contrast, DPI is a direct tool for studying
various types of protein interaction and aggregation in
a quantitative manner, by monitoring the thickness
and density of the protein deposition layer. Moreover,
DPI is sensitive to small changes in protein association
parameters and thus allows monitoring of processes
preceding formation of dye-binding amyloid species,
and thus complements dye-binding assays.
After binding to the sensor surface, aggregation of
a-synuclein commenced immediately and progressed
steadily, without the lag time indicated by TFT bind-
ing data. In contrast, freshly dissolved a-synuclein at
the same concentration and buffer as used for DPI,
but incubated at 37 °C, had an  10 h lag time before
any changes in TFT fluorescence were observed [16],
despite the rate of a-synuclein fibrillation being faster
at 37 °C with shaking than at 25 °C without agitation.
Therefore, it is likely that only nucleation or protofibr-
illar aggregation of a-synuclein was present in the
timeframe of the DPI measurement. The constant val-
ues of thickness, density and mass when a molar
equivalent of aB-crystallin was present indicate that
a-synuclein fibrillar aggregation was efficiently inhib-
ited at an early stage by aB-crystallin, i.e. before the
elongation process. Thus, aB-crystallin inhibited a-syn-
uclein aggregation at the very early prefibrillar stage,
although we cannot deduce whether it acted by pre-
venting the formation of monomeric partially folded
species, or by preventing association of a-synuclein

monomers into nuclei, or both.
We have shown aB-crystallin’s ability to inhibit
a-synuclein fibril formation both in the earliest stages
of its aggregation, and during the growth phase of
fibril formation (Fig. 6A). At each instance, addition
of aB-crystallin stopped further formation of stacked
b-sheet-rich species with comparable efficiency.
Recently, two types of a-synuclein prefibrillar interme-
diates of different structure and aggregation propensity
were characterized [38]. The first type, of a larger size,
accumulates early in fibril formation and rapidly disap-
pears during fibril elongation. A smaller, highly stable
oligomer formed by Met-oxidized a-synuclein is
observed mostly in the later stages and does not form
fibrils. It is probably this species that we observed by
TEM after the fibril formation curve reached its pla-
teau. It is therefore most likely that aB-crystallin inter-
acts with the early metastable a-synuclein oligomers,
because of their partially folded conformation and ten-
dency to association. This interaction would be con-
sistent with previously described accumulation of
nonfibrillar, potentially toxic oligomers in systems in
which fibrillization was inhibited [14,51] and with the
absence of intermediate size oligomers of amyloido-
genic apolipoprotein C-II in the presence of a-crystal-
lin [15]. aB-Crystallin may also bind to monomeric
forms of a-synuclein with partially destabilized confor-
mation if any of these were still present in the later
stages of fibrillation. Such a possibility is indicated by
NMR spectra collected at 10 °C on freshly prepared

samples (and thus not containing any fibrillar species)
and by a decreased chaperone effectiveness of aB-crys-
tallin in the presence of a-synuclein [16]. Finally, the
interaction of aB-crystallin with a-synuclein fibrils (e.g.
by capping the ends of fibrils) also can be considered,
however, addition of aB-crystallin had no effect on the
level of a-synuclein TFT binding, meaning it did not
dissociate existing fibrils.
In summary, the interactions between molecular
chaperones and amyloid-forming proteins (particularly
at the lower range of temperatures studied here) are of
significance for in vivo pathological states. In addition
to the protective role of molecular chaperones in cellu-
lar and animal models of neurodegenerative diseases
[52], their presence in amyloid deposits (e.g. aB-crystal-
lin in Lewy bodies and Ab plaques) may be involved
in facilitating protein aggregation by binding to non-
fibrillar molecules. This can occur in two ways: upon
sequestration of molecular chaperones by amyloido-
genic proteins, the availability of the former to prevent
aberrant aggregation is diminished, but also the
possibility of partial unfolding of the protein and
stabilization of an amyloidogenic intermediate by the
chaperone might make it more susceptible to
A. Rekas et al. a-Crystallin and amyloid fibril formation
FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS 6301
fibrillation or lead to pathological effects, i.e. enhanced
cell toxicity [14–18]. Our study also evaluates the appli-
cability of the TFT-binding assay to kinetic studies of
fibril formation and demonstrates the complementarity

of this method with direct measurement of protein
aggregation via fibril-forming pathway using DPI.
Experimental procedures
Full-length a-synuclein and aB-crystallin were expressed
and purified as described previously [16]. The expression
vector pET24d(+) containing the gene for human wild-type
aB-crystallin was a gift from W. de Jong (Nijmegen, the
Netherlands). Bovine a-crystallin was isolated from calf
lenses, as described previously [16]. Bovine j-casein, TFT
and ANS were purchased from Sigma-Aldrich Pty Ltd.
(Sydney, Australia).
Temperature experiments
Fibril formation by 2.5 mgÆmL
)1
j-casein dissolved in
50 mm sodium phosphate pH 7.2, in the presence of 0.0,
0.25, 0.5 and 1.0 w ⁄ w ratios of a-crystallin ⁄ j-casein, was
monitored in real time using a Fluostar Optima plate
reader. Samples, 200 lL in volume, in 96-microwell plates
(Labsystems, Sydney, Australia) were preincubated with
4 lm TFT for 15 min at the required temperature (30 °Cto
60 °C). Subsequently, for samples requiring reduction,
10 lL of dithiothreitol was added to give a final concentra-
tion of 20 mm. The plate was sealed with a transparent
sealing tape to prevent evaporation. The increase in TFT
binding was monitored by measuring changes in fluores-
cence emission at 490 nm with excitation at 450 nm, with
readings taken every 195 or 315 s. Orbital shaking of
120 r.p.m., 5 mm width, was applied for 95% of the time,
i.e. between the readings. Experiments were performed at

30.0, 33.5, 37.0, 45.0, 50.0, 55.0, 57.5 and 60.0 °C. Rate
constants of fibril formation 1 ⁄ b were calculated after fit-
ting a sigmoidal function f ¼ y
0
+ a ⁄ (1 + exp(– (x )
x
0
) ⁄ b)), where y
0
is initial fluorescence intensity, a is a con-
stant, x is time and x
0
time to 50% of maximal fluores-
cence, using sigmaplot software, to the data, excluding
points of fluorescence decline after the plateau was reached.
In order to establish whether higher temperatures induce
conformational changes that would in turn affect j-casein
fibril-forming propensity, three samples of 10 mgÆmL
)1
j-casein in 50 mm sodium phosphate (pH 7.2) were incu-
bated at 25, 40 or 60 °C for 15 min without agitation,
before being used for amyloid formation in reducing condi-
tions at 25 °C, in the manner described above.
a-Synuclein fibrillation was induced by the addition of
seeds to the final concentration of 0.9 mgÆmL
)1
. Obtained
by sonication of mature fibrils, the seeds sped up the pro-
cess of fibrillation, which would otherwise take several days
[16]. a-Synuclein (3.5 mgÆmL

)1
) solutions in NaCl ⁄ P
i
(137 mm NaCl, 2.7 mm KCl, 10 mm phosphate, pH 7.4)
were incubated with shaking at 37, 45 and 60 °C in the
absence or presence of aB-crystallin (0.5 : 1.0 or 1.0 : 1.0
molar ratio to a-synuclein), as described above for j-casein.
Consistent with species of origin, bovine j-casein was
studied with bovine a-crystallin, and human a-synuclein
with human aB-crystallin. However, fibrillation of a-synuc-
lein and j-casein at 37 °C was inhibited in a similar degree
by either a-crystallin or aB-crystallin (data not shown).
The temperature dependence of fibril-bound TFT was
measured using j-casein (reduced) and a-synuclein fibrils
preformed as described above by incubating at 37 °C for
20 h without TFT. The samples containing 1 mgÆmL
)1
j-casein or 2 mgÆmL
)1
a-synuclein fibrils in NaCl ⁄ P
i
and
4 lm TFT were equilibrated for 10 min at 30 or 25 °C,
respectively, and fluorescence was recorded on Fluostar
Optima plate reader using the above-described wavelengths.
Samples were heated in 2.5 °C increment steps allowing
10 min equilibration before taking each measurement.
TEM
TEMs of reduced j-casein (2.5 mgÆmL
)1

) with and without
a-crystallin (1 : 1 molar ratio) were obtained at ·40 000
magnification from 12 h incubated samples at 37 or 50 °C,
diluted with a double amount of 50 mm sodium phosphate
pH 7.2 before being transferred to carbon- and Formvar-
coated grids (SPI, custom-coated) as described previously
[34]. Fibril dimensions were measured using image j soft-
ware ( />The temperature dependence of aB-crystallin’s ability to
suppress a-synuclein fibrillation (1 : 1 molar ratio) was also
investigated by TEM. Proteins were incubated in NaCl ⁄ P
i
with nucleation seeds (0.9 mgÆmL
)1
) for 4 h at either 37 or
60 °C with agitation. After the completion of experiments,
the samples were refrigerated and grids were prepared as
described previously [16], except that negative staining with
1% phosphotungstic acid for 45 s was used, instead of uranyl
acetate staining. Images of negatively stained samples were
taken at ·50 000 magnification on a Hitachi 7000 TEM.
TEM of a-synuclein incubations (3.5 mgÆmL
)1
,37°C) at
given time points was performed on a Hitachi 7000 TEM.
Grids were prepared in the same way as those for a-synuc-
lein temperature dependence and images were acquired at
·60 000 magnification.
DPI
The process of a-synuclein aggregation was monitored in
real time using a bench-top dual polarization interferome-

ter, the AnaLight Bio200, which measures the mass, thick-
ness and density of the protein layer bound to the surface
of an optical sensor [35]. The sensing chip used in these
a-Crystallin and amyloid fibril formation A. Rekas et al.
6302 FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS
experiments was a dual slab waveguide structure (upper
sensing waveguide and lower reference waveguide) fabri-
cated from a silicon oxynitride on a silicon substrate. DPI
detects changes in the refractive index of the medium on
the sensing waveguide as an optical phase change in com-
parison with that of the reference waveguide. Opening
windows of the upper sensing surface define the two active
areas on the chip, named channel 1 and channel 3. The
region between them acts as an instrument reference chan-
nel (channel 2).
The positively charged surface of the aminated sensor
was chosen for immobilization of a-synuclein (pI ¼ 4.7) by
physisorption.
Calibration of the sensor was performed prior to the
experiment in order to determine the exact thickness and
refractive index of the waveguide sensing layer and the
refractive index of the running buffer. For this purpose, flu-
ids of known refractive indexes, i.e. 80% w ⁄ w ethanol and
water, were applied to the sensor at a flow rate of 50 lLÆ
min
)1
for 3 min each, after which the running buffer (10 mm
Na phosphate, pH 7.2) was perfused over the sensor at
15 lLÆmin
)1

until a stable baseline was achieved in  15 min.
a-Synuclein was loaded on to both channels at a con-
centration of 3.5 mgÆmL
)1
. After 20 min, the injection of
a-synuclein was repeated on channel 3 to make sure that
the sample chamber was filled with protein solution at the
start of the experiment. A 2 : 1 molar ratio of a-synuclein
(3.5 mgÆmL
)1
) ⁄ aB-crystallin (2.5 mgÆmL
)1
)in10mm
sodium phosphate pH 7.2 was injected into channel 1 only.
The flow was stopped and the instrument measured T
m
and
T
e
(transverse magnetic and transverse electric modes of
polarization) on both channels for about 4.5 h, at 25 °C.
Delayed application of aB-crystallin to a-synuclein
during fibril formation
a-Synuclein (250 lm) was incubated in NaCl ⁄ P
i
at 37 °C.
At 0, 25, 49 and 65 h of incubation, aB-crystallin was
added to the samples containing a-synuclein at a 0.5 : 1.0
molar ratio of aB-crystallin ⁄ a-synuclein. Aliquots of 5 and
20 lL were withdrawn from incubations at given time

points and stored frozen until their analysis for TFT bind-
ing and TEM, respectively, as previously described [16].
Size-exclusion HPLC
a-Crystallin, j-casein and the resulting high molecular
weight complexes of both proteins were examined using
size-exclusion HPLC. Samples, 20 lL in volume, containing
10 mgÆmL
)1
j-casein and ⁄ or 10 mgÆmL
)1
a-crystallin in
50 mm sodium phosphate pH 7.2 0.02% NaN
3,
having been
incubated together at 37 °C for 4 h, were loaded on a Bio-
Sep-SEC-S 4000 column (300 · 4.6 mm; Phenomenex, Lane
Cove, Australia). The proteins were eluted with the same
buffer at a flow rate of 0.5 mLÆ min
)1
.
ANS binding
The relative amount of exposed hydrophobicity in j-casein,
a-crystallin and their complexes was examined using ANS,
a hydrophobic fluorescent probe. j-Casein (3 mgÆmL
)1
),
a-crystallin (0.3 mg mL
)1
), and a mixture of both, were
incubated either with or without 20 mm dithiothreitol for

4 h at 25 °C, before being diluted 10 times into 50 mm
sodium phosphate pH 7.5. ANS stock solution was then
added to the final concentration of 20 lm. Samples of non-
reduced j-casein with the same protein composition were
also prepared. A sealed quartz cuvette pathlength 1 cm
(Starna, Sydney, Australia) containing the sample was
placed in a Hitachi F-4500 fluorescence spectrophotometer
connected to a thermostat initially set to 25 °C. The sample
was allowed to equilibrate at the required temperature for
15 min with the shutter closed, before taking the measure-
ment. ANS fluorescence emission spectra were recorded
from 25 to 65 °Cin5°C intervals, with an excitation wave-
length at 385 nm, emission was monitored between 400 and
600 nm. Slit widths were 5 and 10 nm for excitation and
emission beams, respectively.
NMR spectroscopy
1
H–
1
H Watergate TOCSY NMR spectra of proteins were
acquired at 500 MHz on a Varian Inova-500 NMR
spectrometer at 37 ° C as described elsewhere [53]. The
NMR samples contained either 2 mgÆmL
)1
a-crystallin,
2mgÆmL
)1
j-casein, or a mixture of both, in solutions of
10 mm sodium phosphate (pH 7.2) in 10% D
2

O ⁄ 90% H
2
O.
In the case of ‘reduced’ j-casein, the sample was incubated
with 20 mm deuterated dithiothreitol for 2 h before adding
a-crystallin immediately prior to the experiment. For the
identification of cross-peaks, spectra of the protein mixture
were compared with those of 2 mgÆmL
)1
j-casein and of
a-crystallin (20 mgÆmL
)1
; a higher protein concentration
was required because cross-peak intensity in the case of
2mgÆmL
)1
a-crystallin was much lower than of the respec-
tive cross-peaks in the spectrum of the j-casein and a-crys-
tallin mixture, both at 2 mgÆmL
)1
).
Acknowledgements
This study was supported by a grant from the
National Health and Medical Research Council of
Australia to JC.
References
1 Carver JA, Guerreiro N, Nicholls KA & Truscott RJ
(1995) On the interaction of a-crystallin with unfolded
proteins. Biochim Biophys Acta 1252, 251–260.
A. Rekas et al. a-Crystallin and amyloid fibril formation

FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS 6303
2 Carver JA, Lindner RA, Lyon C, Canet D, Hernandez
H, Dobson CM & Redfield C (2002) The interaction of
the molecular chaperone a-crystallin with unfolding
a-lactalbumin: a structural and kinetic spectroscopic
study. J Mol Biol 318, 815–827.
3 Bhat SP & Nagineni CN (1989) aB subunit of lens-spe-
cific protein a-crystallin is present in other ocular and
non-ocular tissues. Biochem Biophys Res Commun 158,
319–325.
4 Dubin RA, Wawrousek EF & Piatigorsky J (1989)
Expression of the murine aB-crystallin gene is not
restricted to the lens. Mol Cell Biol 9, 1083–1091.
5 Iwaki T, Wisniewski T, Iwaki A, Corbin E, Tomo-
kane N, Tateishi J & Goldman JE (1992) Accumula-
tion of aB-crystallin in central nervous system glia
and neurons in pathologic conditions. Am J Pathol
140, 345–356.
6 Reddy GB, Das KP, Petrash JM & Surewicz WK
(2000) Temperature-dependent chaperone activity and
structural properties of human aA- and aB-crystallins.
J Biol Chem 275, 4565–4570.
7 Sun TX & Liang JJ (1998) Intermolecular exchange and
stabilization of recombinant human aA- and aB-crystal-
lin. J Biol Chem 273, 286–290.
8 Bova MP, Ding LL, Horwitz J & Fung BK (1997) Sub-
unit exchange of aA-crystallin. J Biol Chem 272, 29511–
29517.
9 Horwitz J, Huang QL, Ding L & Bova MP (1998) Lens
a-crystallin: chaperone-like properties. Methods Enzymol

290, 365–383.
10 Carver JA, Rekas A, Thorn DC & Wilson MR (2003)
Small heat-shock proteins and clusterin: intra- and extra-
cellular molecular chaperones with a common mechanism
of action and function? IUBMB Life 65, 661–668.
11 Sobott F, Benesch JL, Vierling E & Robinson CV
(2002) Subunit exchange of multimeric protein com-
plexes. Real-time monitoring of subunit exchange
between small heat shock proteins by using electrospray
mass spectrometry. J Biol Chem 277, 38921–38929.
12 Walsh MT, Sen AC & Chakrabarti B (1991) Micellar
subunit assembly in a three-layer model of oligomeric
a-crystallin. J Biol Chem 266, 20079–20084.
13 Das BK, Liang JJ-N & Chakrabarti B (1997) Heat-
induced conformational change and increased chaperone
activity of lens a-crystallin. Curr Eye Res 16, 303–309.
14 Stege GJ, Renkawek K, Overkamp PS, Verschuure P,
van Rijk AF, Reijnen-Aalbers A, Boelens WC, Bosman
GJ & de Jong WW (1999) The molecular chaperone aB-
crystallin enhances amyloid beta neurotoxicity. Biochem
Biophys Res Commun 262
, 152–156.
15 Hatters DM, Lindner RA, Carver JA & Howlett GJ
(2001) The molecular chaperone, a-crystallin, inhibits
amyloid formation by apolipoprotein C-II. J Biol Chem
276, 33755–33761.
16 Rekas A, Adda CG, Aquilina JA, Barnham KJ, Sunde
M, Galatis D, Williamson NA, Masters CL, Anders
RF, Robinson CV et al. (2004) Interaction of the
molecular chaperone aB-crystallin with a-synuclein:

effects on amyloid fibril formation and chaperone
activity. J Mol Biol 340, 1167–1183.
17 Santhoshkumar P & Sharma KK (2004) Inhibition of
amyloid fibrillogenesis and toxicity by a peptide chaper-
one. Mol Cell Biochem 267, 147–155.
18 Raman B, Ban T, Sakai M, Pasta SY, Ramakrishna T,
Naiki H, Goto Y, Rao Ch & M (2005) aB-Crystallin, a
small heat-shock protein, prevents the amyloid fibril
growth of an amyloid beta-peptide and beta2-micro-
globulin. Biochem J 392, 573–581.
19 Lindner RA, Kapur A & Carver JA (1997) The interac-
tion of the molecular chaperone, a-crystallin, with mol-
ten globule states of bovine a-lactalbumin. J Biol Chem
272, 27722–27729.
20 Uversky VN & Fink AL (2004) Conformational con-
straints for amyloid fibrillation: the importance of being
unfolded. Biochim Biophys Acta 1698, 131–153.
21 Goers J, Permyakov SE, Permyakov EA, Uversky VN
& Fink AL (2002) Conformational prerequisites for
a-lactalbumin fibrillation. Biochemistry 41, 12546–
12551.
22 Lindner RA, Treweek TM & Carver JA (2001) The
molecular chaperone a-crystallin is in kinetic competi-
tion with aggregation to stabilize a monomeric molten-
globule form of a-lactalbumin. Biochem J 354, 79–87.
23 Sathish HA, Koteiche HA & McHaourab HS (2004)
Binding of destabilized betaB2-crystallin mutants to
a-crystallin: the role of a folding intermediate. J Biol
Chem 279, 16425–16432.
24 Rochet JC & Lansbury PT Jr (2000) Amyloid fibrillo-

genesis: themes and variations. Curr Opin Struct Biol
10, 60–68.
25 Klein WL, Krafft GA & Finch CE (2001) Targeting
small Abeta oligomers: the solution to an Alzheimer’s
disease conundrum? Trends Neurosci 24, 219–224.
26 Volles MJ & Lansbury PT Jr (2003) Zeroing in on the
pathogenic form of a -synuclein and its mechanism of
neurotoxicity in Parkinson’s disease. Biochemistry 42,
7871–7878.
27 Sirangelo I, Malmo C, Iannuzzi C, Mezzogiorno A,
Bianco MR, Papa M & Irace G (2004) Fibrillogenesis
and cytotoxic activity of the amyloid-forming apomyo-
globin mutant W7FW14F. J Biol Chem 279, 13183–
13189.
28 Bucciantini M, Calloni G, Chiti F, Formigli L, Nosi D,
Dobson CM & Stefani M (2004) Prefibrillar amyloid
protein aggregates share common features of cytotoxic-
ity. J Biol Chem 279, 31374–31382.
29 Kudva YC, Hiddinga HJ, Butler PC, Mueske CS &
Eberhardt NL (1997) Small heat shock proteins inhibit
a-Crystallin and amyloid fibril formation A. Rekas et al.
6304 FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS
in vitro Abeta(1–42) amyloidogenesis. FEBS Lett 416,
117–121.
30 Dedmon MM, Christodoulou J, Wilson MR & Dobson
CM (2005) Heat shock protein 70 inhibits a-synuclein
fibril formation via preferential binding to prefibrillar
species. J Biol Chem 280, 14733–14740.
31 Klucken J, Shin Y, Masliah E, Hyman BT & McLean
PJ (2004) Hsp70 reduces a-synuclein aggregation and

toxicity. J Biol Chem 279, 25497–25502.
32 Conway KA, Harper JD & Lansbury PT Jr (2000)
Fibrils formed in vitro from a-synuclein and two mutant
forms linked to Parkinson’s disease are typical amyloid.
Biochemistry 39, 2552–2563.
33 Farrell HM Jr, Cooke PH, Wickham ED, Piotrowski
EG & Hoagland PD (2003) Environmental influences
on bovine j-casein: reduction and conversion to fibrillar
(amyloid) structures. J Protein Chem 22, 259–273.
34 Thorn DC, Meehan S, Sunde M, Rekas A, Gras SL,
MacPhee CE, Dobson CM, Wilson MR & Carver JA
(2005) Amyloid fibril formation by bovine milk j-casein
and its inhibition by the molecular chaperones a-s- and
beta-casein. Biochemistry 44, 17027–17036.
35 Cross GH, Reeves AA, Brand S, Popplewell JF, Peel
LL, Swann MJ & Freeman NJ (2003) A new quantita-
tive optical biosensor for protein characterisation.
Biosens Bioelectron 19, 383–390.
36 Swann MJ, Peel LL, Carrington S & Freeman NJ
(2004) Dual-polarization interferometry: an analytical
technique to measure changes in protein structure in
real time, to determine the stoichiometry of binding
events, and to differentiate between specific and nonspe-
cific interactions. Anal Biochem 329, 190–198.
37 Nielsen L, Khurana R, Coats A, Frokjaer S, Brange J,
Vyas S, Uversky VN & Fink AL (2001) Effect of
environmental factors on the kinetics of insulin fibril
formation: elucidation of the molecular mechanism.
Biochemistry 40, 6036–6046.
38 Kaylor J, Bodner N, Edridge S, Yamin G, Hong DP &

Fink AL (2005) Characterization of oligomeric inter-
mediates in a-synuclein fibrillation: FRET studies of
Y125W ⁄ Y133F ⁄ Y136F a-synuclein. J Mol Biol 353,
357–372.
39 Kusumoto Y, Lomakin A, Teplow DB & Benedek GB
(1998) Temperature dependence of amyloid beta-protein
fibrillization. Proc Natl Acad Sci USA 95, 12277–12282.
40 Poon S, Rybchyn MS, Easterbrook-Smith SB, Carver
JA, Pankhurst GJ & Wilson MR (2002) Mildly acidic
pH activates the extracellular molecular chaperone clus-
terin. J Biol Chem 277, 39532–39540.
41 Carver JA & Lindner RA (1998) NMR spectroscopy of
a-crystallin. Insights into the structure, interactions and
chaperone action of small heat-shock proteins. Int J
Biol Macromol 22, 197–209.
42 Lindner RA, Kapur A, Mariani M, Titmuss SJ &
Carver JA (1998) Structural alterations of a-crystallin
during its chaperone action. Eur J Biochem 258, 170–
183.
43 Waugh DF (1946) A fibrous modification of insulin. I.
The heat precipitate of insulin. J Am Chem Soc 68,
247–250.
44 Raman B, Ramakrishna T & Rao CM (1995) Tempera-
ture dependent chaperone-like activity of a-crystallin.
FEBS Lett 365, 133–136.
45 Raman B & Rao CM (1997) Chaperone-like activity
and temperature-induced structural changes of a-crystal-
lin. J Biol Chem 272, 23559–23564.
46 Treweek TM, Lindner RA, Mariani M & Carver JA
(2000) The small heat-shock chaperone protein, a-crys-

tallin, does not recognize stable molten globule states of
cytosolic proteins. Biochim Biophys Acta 1481, 175–188.
47 Ghahghaei A, Rekas A, Price WE & Carver JA (2007)
The effect of dextran on subunit exchange of the mole-
cular chaperone aA-crystallin. Biochim Biophys Acta
1774, 102–111.
48 Voropai ES, Samtsov MP, Kaplevskii KN, Maskevich
AA, Stepuro VI, Povarova OI, Kuznetsova IM, Turove-
rov KK, Fink AL & Uverskii VN (2003) Spectral prop-
erties of thioflavin T and its complexes with amyloid
fibrils. J Appl Spectrosc 70, 868–874.
49 Nilsson MR (2004) Techniques to study amyloid fibril
formation in vitro. Methods 34, 151–160.
50 Ecroyd H, Meehan S, Horwitz J, Aquilina JA, Benesch
JL, Robinson CV, Macphee CE & Carver JA (2007)
Mimicking phosphorylation of aB-crystallin affects its
chaperone activity. Biochem J 401, 129–141.
51 Rochet JC, Conway KA & Lansbury PT Jr (2000) Inhi-
bition of fibrillization and accumulation of prefibrillar
oligomers in mixtures of human and mouse a-synuclein.
Biochemistry 39, 10619–10626.
52 Muchowski PJ (2002) Protein misfolding, amyloid for-
mation, and neurodegeneration: a critical role for
molecular chaperones? Neuron 35, 9–12.
53 Treweek TM, Rekas A, Lindner RA, Walker MJ,
Aquilina JA, Robinson CV, Horwitz J, Der Perng M,
Quinlan RA & Carver JA (2005) R120G aB-crystallin
promotes the unfolding of reduced a-lactalbumin and is
inherently unstable. FEBS J 272, 711–724.
A. Rekas et al. a-Crystallin and amyloid fibril formation

FEBS Journal 274 (2007) 6290–6305 ª 2007 The Authors Journal compilation ª 2007 FEBS 6305