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

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

by a-crystallin

Temperature dependence and the nature of the aggregating species Agata Rekas1,2, Lucy Jankova3, David C Thorn4, Roberto Cappai5,6and John A Carver4

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: john.carver@adelaide.edu.au

(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.

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

Ca2+or the presence of Zn2+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 con-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 asand 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

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 30C 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 50C (13.4 ± 2.2 nm in diameter), although a difference in their supramolecu-lar morphology was evident: at 37C fibrils were well separated, whereas at 50C 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 37C were shorter (101.1 ± 49.6 nm) than those formed at 50C (148.0 ± 88.3 nm), including a larger number of small fragments (up to 20 nm in length) at the lower temper-ature At 37C, 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 50C, 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 50C 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 (45–60C), 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; 50C 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 +

-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.

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 30C, 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 30–33.5C, 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 33.5C, 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(kapp)¼ – EA⁄ RT + ln(A) The activation energy of fibril formation (EA) was calculated (Table 1) as the slope of the straight line fitted to a plot of ln(kapp) 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, kapp For reduced j-casein only, EA was 35.5 ± 1.1 kcalÆmol)1, showing strong temperature dependence of the rate constants (R2¼ 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

time (hours)

TFT binding @ 25 o C

0 200 400 600 800 1000 1200 1400 1600

time (hours)

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

0.10 1.00 10.00 100.00

-1 ]

0 50 100

κ-cas

κcasein κcas+0.25xαcrys κcas+0.5xα-crys κ cas+1xα-crys

C

500

100 250

50 0

0

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 (EA) 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.

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

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

1 : 1 molar ratio of j-casein⁄ a-crystallin, EA was

14.9 ± 2.0 kcalÆmol)1 (R2¼ 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 25C (Fig 2C), although the

changes in the rate constants were not significant:

(1.85 ± 0.09)· 10)1Æs)1, (1.86 ± 0.15)· 10)1Æs)1 and

(1.79 ± 0.11)· 10)1Æs)1 after preincubation at 25, 40

and 60C, respectively The initial lag times increased

from 10 to 13 to 20 min for 25, 40 and 60C

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 30–55C

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 45C (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 (EA) 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 45C, but

was significantly lower at 60C (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

60C, 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 60C was almost completely inhibited, while only partial suppression was achieved at 37C (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 28–60C for j-casein, and 39% reduction for a-synuclein between 25 and 52.5C (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.

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 (http://www.farfield-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

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

1

0

0

0

0

0

0

0

0

0

0

0

0 1 0 0

0 0

) s t u i m ( e m i

α-syn α-syn+0.5xα-crys α-syn+1xα-crys

7oC

5oC

0oC

200 nm

0 0

0 5

0 0

0 5

0 0

0 5

0 0

0 0

0 0

0

( e r u t a r e m e

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)1j-casein (closed symbols) and 2 mgÆmL)1a-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.

0.0746 nm on channel 3, the mass increased by

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 25C 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 25C 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 3 2 1

0

5

.

0

8

.

0

1

.

0

4

.

0

7

.

0

7

.

0

4 3 2 1

0

3

.

1

5

.

1

8

.

1

5

.

2

3

.

2

5

.

2

8

.

2

4 3 2 1

0

) s r u h ( e m i

α- y +αB - c r s

α- y u c l e i n

α- y +αB - c r s

α- y u c l e i n

α- y +αB - c r s

α- y u c l e i n

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)1and 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.

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 37C 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

time (hours)

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

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.

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

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

elution time (min)

k-cas a-crystallin cas reduced -cas -

- crys cas red crys

14.6 kD

67 kD

669 kD

2000 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.

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 (25C)

and decreased upon heating to 70C, 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)

1H-NMR spectroscopy

Cross-peaks from the NH to aliphatic proton regions

of1H 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 EA 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