1. Trang chủ
  2. » Luận Văn - Báo Cáo

Tài liệu Báo cáo khoa học: High affinity copper binding by stefin B (cystatin B) and its role in the inhibition of amyloid fibrillation docx

14 587 0
Tài liệu đã được kiểm tra trùng lặp

Đang tải... (xem toàn văn)

Tài liệu hạn chế xem trước, để xem đầy đủ mời bạn chọn Tải xuống

THÔNG TIN TÀI LIỆU

Thông tin cơ bản

Tiêu đề High affinity copper binding by stefin B (cystatin B) and its role in the inhibition of amyloid fibrillation
Tác giả Eva Žerovnik, Katja Škerget, Magda Tušek-Žnidarič, Corina Loeschner, Marcus W. Brazier, David R. Brown
Trường học Jožef Stefan Institute
Chuyên ngành Biochemistry and molecular biology
Thể loại Journal article
Năm xuất bản 2006
Thành phố Ljubljana
Định dạng
Số trang 14
Dung lượng 0,97 MB

Các công cụ chuyển đổi và chỉnh sửa cho tài liệu này

Nội dung

Keywords copper-binding proteins; cystatin; inhibition of amyloid fibril formation; oligomers; protein aggregation; stefin B Correspondence E.. There is no difference in the affinity of c

Trang 1

role in the inhibition of amyloid fibrillation

Eva Zˇerovnik1, Katja Sˇ kerget1

, Magda Tusˇek-Zˇnidaricˇ2, Corina Loeschner3, Marcus W Brazier3 and David R Brown3

1 Department of Biochemistry and Molecular Biology, Jozˇef Stefan Institute, Ljubljana, Slovenia

2 Department of Plant Physiology and Biotechnology, National Institute of Biology, Ljubljana, Slovenia

3 Department of Biology and Biochemistry, University of Bath, UK

Common features of many neurodegenerative diseases

are misfolding, aggregation and amyloid fibril

forma-tion of a pathological mutant (in inherited diseases), or

of a normal protein or its normal variant (in sporadic

cases) Amyloid fibril formation is regarded as a generic

property, common to most proteins [1,2], encouraging

the study of proteins not involved in any pathology

Amyloid-induced toxicity has also been proposed to be

a generic phenomenon [3], with prefibrillar oligomers as

the most likely toxic agent Sequestering of the fibrils

into intracellular inclusions or extracellular plaques

might actually be beneficial, as it is now believed that

soluble oligomers are the cause of the initial insult rather than the insoluble fibrous material

Environmental factors, among them metal ions, are believed to contribute to the onset of Parkinson’s, Alzheimer’s and prion disease The influence of metal ions on the underlying process of amyloid fibril forma-tion remains controversial In a number of cases copper, like other redox active metals, has been shown to pro-mote aggregation or polymerization However, there are recent reports that binding of Cu2+and Zn2+, but not Fe3+, to amyloid-b peptide retards amyloid fibril formation [4]

Keywords

copper-binding proteins; cystatin; inhibition

of amyloid fibril formation; oligomers;

protein aggregation; stefin B

Correspondence

E Zˇerovnik, Department of Biochemistry

and Molecular Biology, Jozˇef Stefan

Institute, Jamova 39, 1000 Ljubljana,

Slovenia

Fax: +386 1477 3984

Tel: +386 1477 3753 ⁄ 3900

E-mail: eva.zerovnik@ijs.si

David R Brown, Department of Biology and

Biochemistry, University of Bath, Claverton

Down, Bath, BA2 7AY, UK

Fax: +44 1225 386779

Tel: +44 1225 383133

E-mail: bssdrb@bath.ac.uk

(Received 3 May 2006, revised 16 July

2006, accepted 18 July 2006)

doi:10.1111/j.1742-4658.2006.05426.x

We show that human stefin B, a protease inhibitor from the family of cystatins, is a copper binding protein, unlike stefin A We have used isothermal titration calorimetry to directly monitor the binding event at

pH 7 and pH 5 At pH 7 stefin B shows a picomolar affinity for copper but at pH 5 the affinity is in the nanomolar range There is no difference

in the affinity of copper between the wildtype stefin B (E31 isoform) and a variant (Y31 isoform), whereas the mutant (P79S), which is tetrameric, does not bind copper The conformation of stefin B remains unaltered by copper binding It is known that below pH 5 stefin B undergoes a conform-ational change and amyloid fibril formation We show that copper binding inhibits the amyloid fibril formation and, to a lesser degree, the initial aggregation Similarities to and differences from other copper binding amy-loidogenic proteins are discussed

Abbreviations

AFM, atomic force microscopy; ITC, isothermal titration calorimetry; SEC, size exclusion chromatography; TEM, transmission electron microscopy; TFE, 2,2,2 trifluorethanol; ThT, thioflavin T.

Trang 2

Using human stefin B as a model for studying the

mechanism of amyloid fibril formation [5,6], it has

been found that fibrillation starts with a lag phase and

continues with a fibril growth reaction The lag phase,

in which granular (micellar-like) aggregate accumulates

[5], can be reduced by increasing the temperature, by

adding the organic solvent trifluoroethanol (TFE) or

by seeding [6] The time course for morphological

changes occurring during the amyloid fibril formation

by human stefin B is reminiscent of that described

for other amyloidogenic proteins, including amyloid-b

peptide [7] By following the kinetics of stefin B fibril

formation, conditions were defined where the protein

exists in the form of prefibrillar oligomers⁄ aggregates,

which persist during the lag phase The prefibrillar

forms were shown to be cytotoxic and to interact with

acidic phospholipids [8]

Human stefin B, officially termed cystatin B

(sub-family A, (sub-family I25 of cystatins following the

MEROPS classification [9]), is a cysteine protease

inhibitor [10,11] The structure and function of this

protein may be relevant to both amyloid fibril

forma-tion and metal binding Stefin B is homologous to a

closely related protein, stefin A Crystal structures of

stefin B in complex with papain [12] and of stefin A in

complex with cathepsin H [13] have been determined

The solution structure of free stefin A is also known

[14] Domain-swapped dimers have been shown for

ste-fin A and for cystatin C [15–17] Domain swapping

may have a role in amyloid fibril formation of this

family of proteins [16]

Stefin B is expressed widely in human tissue and is

thought to act as an inhibitor of the lysosomal

cathep-sins Alternative functions are possible, as the protein

was found as part of a multiprotein complex of

unknown function, specific to the cerebellum [18] It is

located not only in the lysosomes and in the

cyto-plasm, but also in the nucleus [19] Lack of expression

of stefin B is associated with signs of cerebellar

gran-ular cell apoptosis, ataxia and myoclonus as shown in

studies of stefin B deficient mice [20] Genes involved

in the activation of glial cells were overexpressed in

such mice [21] Stefin B (cystatin B gene) is also tightly

linked to epilepsy Mutations in this gene [22,23],

which lead mainly to lower protein expression, result

in progressive myoclonus epilepsy of the Unverricht–

Lundborg type The protein was reported to be

overex-pressed after seizures [24], implicating its

neuroprotec-tive role, similarly to that of cystatin C [25] Similarly

to cystatin C [26], it was found as a constituent of

senile plaques of different disease origin [27]

In the current study the ability of stefin B to bind

copper was assessed We used isothermal titration

calorimetry to monitor the binding event at pH 7 and

pH 5 It was found that the protein binds two Cu2+ atoms with high affinity whereas the mutant P79S, which is tetrameric, does not It also was shown that the presence of equimolar to three-fold molar excess of

Cu2+inhibits the fibrillation propensity of the protein This was demonstrated by thioflavin T fluorescence and electron microscopy

Results

Measurement of copper binding by isothermal titration calorimetry

One of the most widely accepted methods for deter-mining the affinity of a ligand for a protein is iso-thermal titration calorimetry (ITC) We used ITC to determine the affinity of copper for human stefins Recombinant stefin proteins were dissolved in 5 mm Mes buffer at either pH 7 or pH 5 The proteins ana-lyzed were stefin A, stefin B (E31 isoform), a variant

of the protein (variant 2) with a change from E to Y

at amino acid residue 31 (Fig 1), and a mutant form

of the variant (P79S) Copper was found to bind to stefin B at pH 7 but not to stefin A (Fig 2) The bind-ing isotherm data were fitted to sequential bindbind-ing site parameters and the best fit, producing the smallest v2 values, indicated two binding sites, both with affinities

in the picomolar range at pH 7 (Table 1) No optimal

fit was found for stefin A, indicating that the protein has no specific affinity for copper

Further analysis showed that stefin B, again unlike stefin A, also binds copper at pH 5 but that the affin-ity is by two orders of magnitude less, in the nano-molar range (Fig 3, Table 1) Additional ITC experiments were carried out with the variant form of stefin B and its mutant form, P79S The variant stefin B (Y31 isoform) binds Cu2+ with similar affinity to that

of the more common E31 isoform (Table 1) However, the mutant form of the variant, P79S, shows no specific copper binding at either pH (Fig 3, Table 1)

Conformation and stability in the presence and absence of copper

Stefin B is a predominantly b-sheet protein with five strands wrapping around an a-helix The far UV CD spectra in Fig 4A reveal small differences in intensity and shape between stefins A and B and the P79S mutant The two isoforms of stefin B have exactly the same far UV CD Regardless of the sequence differences (as highlighted in Fig 1), the secondary and tertiary structures of stefins A and B are the same, as determined

Trang 3

from the 3D structures [12,14] Therefore, the

differ-ences in the far UV CD must be due to the contribution

of tyrosines to this region of the spectrum [28]

Figure 4B shows the near UV CD spectra of

stefin A, stefin B, and the P79S mutant It can be seen

that spectra of stefin B and the P79S mutant are very

similar in shape whereas that of stefin A is different

This is accounted for by the different aromatic amino

acid content (Fig 1) The similar shapes of the stefin B

and P79S spectra provides evidence for similar 3D

structure and correct folding of the mutant The lower

intensity may arise from partitioning of the protein

into an aggregated state

The effect of copper binding on the secondary

struc-ture was determined using CD in the far UV To

prepare proteins without Cu2+, part of each protein

solution was exchanged by ultrafiltration with the chelating buffer at pH 7 and then diluted to the appro-priate concentration The other part was diluted directly into buffer with 50 lm CuSO4 The far UV

CD spectra of stefin B (E31 isoform) were the same in the presence and absence of copper (Fig 5A) This indicates that copper binding has no apparent effect

on the structure of the protein Similarly, the presence

of Cu2+had no effect on the CD spectrum of variant

2 (Fig 5B) For comparison, the spectra of the P79S mutant and of stefin A were recorded with and with-out copper (Fig 5C,D) The latter two proteins do not bind Cu2+ Being aware of the pitfalls of such an ana-lysis for proteins with unusual aromatic contribution

to the far UV CD, the secondary structure estimates were calculated from the far UV CD spectra using

-6

-4

-2

0

2

0 10 20 30 40 50 60 70

Time (min)

-0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

-2

0

Molar Ratio

kcal/mole of injectant kcal/mole of injectant

-2 -1 0

0 10 20 30 40 50 60 70

Time (min)

-0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 -4

-2 0

Molar Ratio

Fig 2 Comparison of copper binding by ste-fin A and steste-fin B using ITC Protein sam-ples were prepared at 35 l M in 5 m M Mes

pH 7 buffer The ligand used for injection was Cu2SO4equilibrated with a four molar excess of glycine Cu 2+ was applied up to a five molar excess The top panel represents the data from the titration as a series of peaks corresponding to the heat change (lcalÆs)1) with each injection The bottom panel is a plot of heat change on ligand addi-tion (kcalÆmole)1) against the ligand ⁄ stefin molar ratio The background heat change from the Cu 2+ ⁄ Gly mixture injected in the Mes buffer was subtracted from the raw data Data of one representative experiment each is shown.

Wt mmsgapsatq pataetqhia dqvrsqleek enkkfpvfka vsfksqvvag tnyfikvhvg dedfvhlrvf qslphenkpl

Var2 mmsgapsatq pataetqhia dqvrsqleek ynkkfpvfka vsfksqvvag tnyfikvhvg dedfvhlrvf qslphenkpl

P79S mmsgapsatq pataetqhia dqvrsqleek ynkkfpvfka vsfksqvvag tnyfikvhvg dedfvhlrvf qslphenksl

StA mipgglseak patpeiqeiv dkvkpqleek tnetygklea vqyktqvvag tnyyikvrag dnkymhlkvf kslpgqnedl

Wt tlsnyqtnka khdeltyf

Var2 tlsnyqtnka khdeltyf

P79S tlsnyqtnka khdeltyf

StA vltgyqvdkn kddeltgf

Fig 1 Comparison of stefin sequences Shown are the primary amino acid sequences of the three stefin B proteins studied and that of stefin A The potential copper binding site with four histidine residues is shown in the boxes Differences between the wildtype stefin B, variant

2 and the P79S mutant of the variant are shown by the bold, underlined letters All four proteins of 98 amino acids are approximately 11 kDa.

Trang 4

dichroweb online software [29,30] There was no

dif-ference when comparing the effects of Cu2+, which

supports the conclusion that copper binding does not

alter the secondary structure of the protein

Near UV CD spectra are a better measure of protein tertiary structure and thus the solution conformation There are cases where the tertiary structure can denature with no substantial change in the secondary structure, leading to intermediate states Similar obser-vations were made for stefin B previously [31] Near

UV CD spectra of the protein at pH 7 and pH 5 with

no metal bound and in its presence were recorded (Fig 5E,F) The spectra show that stefin B is sensitive

to Cu2+ at pH 7 where a significant decrease in ellip-ticity is detected, whereas at pH 5 this does not seem

to be the case Lower intensity of the CD signal at

pH 7 in presence of Cu2+ (Fig 5E) may, similarly to P79S (Fig 4B), arise from enhanced protein aggrega-tion rather than a conformaaggrega-tional change

To assess protein stability, thermal denaturation of the protein in presence of Cu2+ or in its absence was recorded at 210 nm (at a protein concentration of around 20 lm; not shown) and there was no difference

in the temperature of half-denaturation Performing thermal denaturation at 277 nm (which is only possible

at around 100 lm protein concentration) has shown

Molar Ratio -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

Molar Ratio

kcal/mole of injectant kcal/mole of injectant

kcal/mole of injectant kcal/mole of injectant

-2

0

-1

-0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

Stefin A pH5 -2

0

-1

-0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 -0.20

-0.15 -0.10 -0.05 0.00 0.05 0.10

Molar Ratio

P79S pH7

-0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 -0.20

-0.15 -0.10 -0.05 0.00 0.05 0.10

Molar Ratio P79S pH5

Variant 2 pH7 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 -4

-2 0

Molar Ratio

Stefin B pH5

Molar Ratio

Variant 2 pH5 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 -4

-2 0

Fig 3 Effects of pH and primary sequence

on stefin copper binding Further analysis of

stefin proteins (35 l M ) with ITC to assess

affinity of copper binding Experiments were

carried out at either pH 7 or pH 5 using

5 m M Mes as a buffer Shown are stefin A

and stefin B at pH 5 and variant 2 of

ste-fin B and the P79S mutant of the variant at

both pH 7 and pH 5 Each panel is a plot of

heat change on ligand addition (kcalÆmole)1)

against the ligand ⁄ stefin molar ratio for one

of the proteins analysed The background

heat change from the Cu 2+ ⁄ Gly mixture

injected in the Mes buffer was subtracted

from the raw data Data of one

representa-tive experiment each is shown.

Table 1 Copper affinity for stefin proteins as determined by ITC.

Values shown for affinity are those from the best fit of two sites.

nsd ¼ no site detected Values are in units of M )1 and are the

averages of three measurements The standard error was less than

5% of the value for each measurement Shown are the two values

for a sequential two site fit Fitting for one or three sites resulted in

two values at least two orders of magnitude higher.

Binding

Stefin B (Variant 2)

Stefin B (P79Sb)

pH 7

7.6 · 10 9

pH 5

1.8 · 10 8

nsd

Trang 5

that the protein aggregates heavily in the presence of

Cu2+ at pH 7 and to a lower extent at pH 5

There-fore, no comparison of the stability of the tertiary

structure to the effect of Cu2+could be made

Influence of copper on oligomerization,

aggregation and amyloid fibril formation

First, we probed the effect of copper binding on

oligomer formation The wildtype stefin B has

well-defined oligomers, which can be separated by size

exclusion chromatography (SEC) The isolated

mono-mer was incubated at pH 7 or at pH 5, in the presence

or absence of Cu2+, and after longer times (1 week at

room temperature and 4 days at 36C, respectively),

samples were taken for the SEC analysis The results

can be described as follows: at a lower temperature no

difference in the ratio between the oligomers is seen,

whereas after incubation at 36C there is a marked

shift from the monomer towards the dimer (and some

tetramer) at pH 7 At pH 5 the protein undergoes

complete dimerization even with no Cu2+ present A

conclusion can be made that copper binding facilitates

dimerization already at neutral pH (which would be expected to promote further oligomerization⁄ aggrega-tion)

It has been shown previously that 10% 2,2,2 trifluor-ethanol (TFE) is the optimal concentration needed to accelerate fibril growth by stefin B at pH 5 [5,6] In the present series of experiments we compared fibrillation

of stefin B wildtype (E31 isoform), of stefin B variant

2 (Y31 isoform) and the mutant P79S of the variant, all at pH 5.0 with 10% TFE, 25C, in the presence of

Cu2+or with no Cu2+present Two concentrations of CuSO4 in the buffer were used (50 and 150 lm) giving

1 : 1 and 1 : 3 protein to Cu2+ratios

Figure 6 and Table 2 show the outcome of the fibril-lation assays with and without Cu2+ present in the medium Fibrillation of stefin B wildtype (E31 isoform) was first recorded at 40C, at pH 7, where no fibrilla-tion was observed, and at pH 5 at 40C (Fig 6A) The thioflavin T (ThT) intensity increased to some extent under these latter conditions, reflecting fibril growth, but much less than when TFE was added (Fig 6B) Fibrillation of all the three proteins (stefin B wildtype, variant 2 and the P79S mutant) at the stand-ard assay conditions (pH 5, 10% TFE, 25C), are plotted in Fig 6B–D It can be seen that Cu2+ inhib-ited fibril growth in all cases: with stefin B wildtype (Fig 6B), with stefin B variant 2 (Fig 6C) and even with the P79S mutant (Fig 6D) A very similar overall picture was obtained with three-fold Cu2+ excess (not shown)

The results were normalized in such a way that the maximal value of ThT fluorescence intensity was taken

as 100% From these, for each reading of ThT fluores-cence the percentage of inhibition of the fibril growth was obtained The percentage of inhibition (Table 2) is correlated with Cu2+ concentration, and is higher at

1 : 3 protein to Cu2+ molar ratio than at 1 : 1 The P79S mutant, which does not bind Cu2+ and is tetra-meric, also does not fibrillate to the same extent as wildtype or variant 2, but some inhibition by Cu2+ is still observed (Fig 6D, Table 2) Regardless, this does not seem to depend on Cu2+concentration

Transmission electron microscopy (TEM) data were collected at 9000 min of fibrillation, which is the time where maximal change in ThT fluorescence occurs (Fig 6B,C) The TEM images (Fig 7) confirm that a three-fold molar excess of Cu2+ (Fig 7B,D) markedly reduces the amount of fibrils in comparison to granular aggregate in stefin B and in the variant In comparison with earlier studies it seems that even aggregation is inhibited and not only fibril formation This will be discussed in view of a proposed higher toxicity of the aggregates in comparison to the mature fibrils

far UV CD spectra

-8000

-6000

-4000

-2000

0

2000

4000

6000

8000

nm

stefin A stefin B P79S

near UV CD spectra

-40

-20

0

20

40

60

80

100

120

nm

stefin A stefin B P79S

B

A

Fig 4 Circular dichrosim spectra of stefin B and of stefin A (A) Far

UV CD spectra of stefin A, stefin B and the P79S mutant (B) Near

UV CD spectra of stefin A, stefin B and the P79S mutant.

Trang 6

Cystatins and neurodegenerative disease

Cystatin C is a well-known amyloidogenic protein The

L68Q variant is associated with a hereditary form

of cerebral amyloid angiopathy that results in a

fatal brain hemorrhage [32] Wildtype cystatin C has

been found as a component of amyloid plaques in Alzheimer’s disease [26] and shown to inhibit amyloid fibril formation of amyloid-beta [33] In searching the literature we found a report of stefins A and B together with some cathepsins being found in the core

of senile plaques of different disease origin [27] Cystatins have been found to be important in neuro-degeneration and in neuroregeneration Cystatin C

-8000

-6000

-4000

-2000

0

2000

4000

6000

8000

nm

wt stB with Cu

wt stB without

-6000 -4000 -2000 0 2000 4000 6000 8000 10000 12000

nm

var2 stB with Cu var2 stB without

-3000

-2000

-1000

0

1000

2000

3000

4000

5000

6000

nm

P79S with Cu P79S without

C

-8000 -6000 -4000 -2000 0 2000 4000 6000

nm

stA with Cu stA without

D

-2

-1

0

1

2

3

4

5

nm

pH7 pH7Cu

E

-2 -1 0 1 2 3 4 5

nm

pH5 pH5Cu

F

Fig 5 Circular dichroism spectroscopy as a function of Cu2+concentration (A) Far UV CD spectra of stefin B wildtype (E31 isoform) in pres-ence of Cu 2+ and without Cu 2+ Measurements in the far UV were carried out at 25 C at pH 7.3 (NaCl ⁄ P i buffer) (1 mm rectangular cell, bandwidth 1 nm, each 1 nm for 5 s) (B) Far UV CD spectra of stefin B variant 2 (Y31 isoform) in the presence and without Cu 2+ (C) Far UV

CD spectra of the P79S mutant in the presence and without Cu2+ (D) Far UV CD spectra of stefin A in the presence and without Cu2+ (E) Near UV CD spectra of stefin B wildtype in the presence of Cu 2+ and without Cu 2+ at pH 7 Measurements in the near UV were collected

at 20 C using a 10 mm rectangular microcell, bandwidth 0.5 nm, collecting data each 0.5 nm for 3 s (F) Near UV CD spectra of stefin B wildtype in the presence of Cu2+and without Cu2+at pH 5.

Trang 7

was reported to modulate neurodegeneration and

neurogenesis following status epilepticus in mouse [25]

Mutations in human stefin B (cystatin B gene; CSTB)

were identified as a cause of the progressive myoclonus

epilepsy of the Unverricht–Lundborg type In studies

of CSTB-deficient mice, lack of this inhibitor was found to be associated with signs of cerebellar granular cell apoptosis [20] The mice develop progressive ataxia and myoclonic seizures and undergo an extensive loss

of Purkinje cells They provide a reasonably good model for the disease The transcripts that were consis-tently increased in brain tissue from CSTB-deficient mice encode proteins involved in responding to neuronal damage [21], i.e., genes which code for increased proteolysis, apoptosis and glial cell activation Copper homeostasis is important in the brain, there-fore the role of copper binding or loss of its binding could be related to specific cerebellar function(s) of stefin B [18], which remains to be seen by more in vivo studies

Stefin B as a copper binding protein

We have demonstrated that human stefin B is a high affinity copper binding protein It exerts two high affinity biding sites in the picomolar range at pH 7

pH=5, 40 o C

-200

0

200

400

600

800

1000

Time (min)

wt stB without

wt stB with Cu

pH=5, 25 o C, 10%TFE

-200 0 200 400 600 800 1000

Time (min) ThT fluorescence / 480 nm ThT fluorescence / 480 nm

wt stB without

wt stB with Cu

B A

ThT fluorescence / 480 nm ThT fluorescence / 480 nm

pH=5, 25 o C, 10% TFE

-200

0

200

400

600

800

1000

Time (min)

var2 stB without var2 stB with Cu

pH=5, 25 o C, 10% TFE

-200 0 200 400 600 800 1000

Time (min)

mut P79S without mut P79S with Cu

Fig 6 Inhibition of fibrillation of stefin B by Cu 2+ as probed by ThT fluorescence For experimmental detail see Experimental procedures Final protein concentration was in all cases 45 l M and final concentration of Cu2+46 l M , leading to 1 : 1 of protein to Cu2+ratio Results for 1 : 3 pro-tein to Cu 2+ ratio have also been obtained (not shown) (A) Stefin B wildtype (E31 isoform) at pH 5, 40 C, 0 and 50 l M Cu 2+ in the buffer (B) Stefin B wildtype (E31 isoform) at pH 5, 10% TFE, 25 C, 0 and 50 l M Cu 2+ in the buffer (C) Stefin B variant 2 (Y31 isoform) at pH 5, 10% TFE, 25 C, 0 and 50 l M Cu2+in the buffer (D) P79S mutant of variant 2 at pH 5, 10% TFE, 25oC, 0 and 50 l M Cu2+in the buffer.

Table 2 Inhibition of fibrillation of stefin B proteins by Cu 2+

Con-centration of the protein was normally 45 l M while final

concentra-tions of Cu2+in solution were 46 l M and 138 l M , which gives 1 : 1

and 1 : 3 protein to Cu 2+ molar ratios, respectively.

Protein ⁄ variant [Cu 2+ ] (l M ) Solvent composition % of inhibition

Trang 8

The affinity for these sites is decreased with decreased

pH (Figs 2 and 3, Table 1) In comparison, human

stefin A, which has the same 3D structure, does not

bind copper Although the structure of the two

pro-teins is almost identical, the sequences differ in a

num-ber of places, but in particular, stefin B has a numnum-ber

of histidines in the C-terminus (box in Fig 1) at sites:

92, 75, 66 and 58 As histidine residues are central to

copper binding in many proteins they probably form

part of the copper binding sites in this protein

Although there are four histidines in the C-terminal,

another histidine is present at position 18, and given the folded state of the protein it is possible that this residue could play a role in copper binding No His residues are located at the homologous sites in human cystatin C (sequences were aligned), suggesting that copper binding might be specific to stefin B among the three human cystatins

Further support for the premise that the C-terminus could be the copper binding domain comes from stud-ies of the P79S mutant, which differs from the copper binding forms of stefin B by one amino acid residue

Fig 7 Inhibition of amyloid fibril growth as observed by TEM Samples at 34 l M protein concentration were incubated at 25 C in the stand-ard fibrillation buffer and after 9000 min (the plateau level of the reaction) they were prepared for TEM measurement (A) Stefin B variant 2 (Y31 isoform) prepared in chelated buffer pH 5, 10% TFE (B) Stefin B variant 2 (Y31 isoform) in the same buffer with 100 l M final Cu2+ (C) Stefin B wildtype (E31 isoform) prepared in chelated buffer pH 5, 10% TFE (D) Stefin B wildtype (E31 isoform) in the same buffer with

100 l M final Cu 2+ concentration.

Trang 9

change but lacks any copper binding capacity This

point mutation lies within the C-terminal region and

therefore could affect the structure of this part of the

protein However, it should not be dismissed that the

P79S mutant differs from the other stefin variants in

that it forms a tetramer Analysis of the far UV CD

spectra (Fig 4A) by dichroweb [29,30] suggests only

negligible change in the secondary structure between

the wildtype stefin B or variant stefin B and the P79S

mutant of the variant, consistent with tetramerization

Near UV CD spectra of stefin B and the P79S mutant

are also very similar (Fig 4B), consistent with proper

folding of the tetramer comparable to the wildtype

protein It is possible, however, that a new interface

formed in the tetramer would disrupt copper binding

Inhibition of amyloid fibril formation by stefin B

in presence of copper

The mechanism of amyloid fibril formation of cystatins

is being studied [5,6] It is proposed that domain

swapping is followed by tetramerization and further

oligomer formation [16,34], which accumulate into the

so-called ‘critical oligomers’ [35] and then grow into

protofibrils and mature fibrils Therefore, inhibition by

Cu2+ of amyloid fibril formation of human stefin B

could result from loss of correct Cu2+ coordination

before the stage of tetramerization Possibly, loss of

copper binding could still allow domain swapping to

occur It is of interest that a N-terminally truncated

pri-on protein, lacking the copper binding domain is

cap-able of domain swapping and forms a dimer as revealed

by crystal structure analysis [36] In our case, SEC data

collected for samples at pH 7 have confirmed that

cop-per binding shifts the equilibrium towards the dimer

and that the monomer remains monomeric in its

absence At pH 5, the protein is dimeric even with no

Cu2+bound

It has been shown that stefin B is very much prone

to undergo oligomerization and amyloid fibril

forma-tion [5,6,31] Therefore, we probed the effect of Cu2+

(loss of copper binding) on fibril formation of this

pro-tein Our results show that upon copper binding

amy-loid fibrillation gets less (this is judged by ThT

fluorescence intensity and TEM; see below) In

partic-ular, our data shows that copper binding inhibits

amy-loid fibril growth of the two stefin B isoforms but does

not affect the P79S mutant, which is purely tetrameric

This seems to suggest that Cu2+ could stabilize the

protein and thus inhibit amyloidogenesis before the

point of tetramerization This would suggest that

initial aggregation is critically dependent on domain

swapping, which is a step prior to tetramerization

[16,34] However, delay in tetramer formation could still lead to granular aggregate formation (Fig 7B,D)

It has been shown that prefibrillar oligomers may be more toxic than the fibrils themselves [37,38] In the case of stefin B (Fig 7) it seems that not only fibrilla-tion is diminished but also the amount of granular aggregate (Fig 7B,D) This is judged from our previ-ous observations of the lag phase granular aggregate obtained at the same protein concentration [5,39] Amyloid fibril formation of the N-terminal fragment

of stefin B up to residue 68, as observed in some patients with Unverricht–Lundborg type 1 progressive myoclonus epilepsy has shown an increased amyloido-genic potential, as reported by Rabzelj et al [39] Not-withstanding problems with folding of the fragment [39], which stays unfolded, this seems to support our premise that loss of copper binding (residues 92 and

75 are lost) contributes to the progress of amyloido-genesis

Effect of copper binding on amyloid formation

of other proteins

A number of other amyloidogenic proteins have been shown to be copper binding proteins Copper binding

is sometimes specific and at other times nonspecific In particular, copper like other redox active metals has been shown to promote aggregation or polymerization

in a number of cases Although the prion protein binds copper in its native conformation [40], the presence of copper has also been shown to accelerate aggregation

of the protein [41] or increase the infectivity of prion isolates However, this kind of interaction is nonspe-cific Other studies have suggested that specific binding

to the prion protein stabilizes its structure and pre-vents intermediate, partially unfolded states that could result in a major conformational change [42] Loss of appropriate metal binding and substitution with a dif-ferent metal could initiate such conformational chan-ges in the prion protein [43] Recently, it has been shown that copper binding to alpha synuclein causes aggregation and fibril formation [44] It is well known that alpha synuclein is one of the natively unfolded proteins Copper also binds to the amyloid precursor protein and to the cleavage product amyloid-beta [45,46] However, in this case, binding of copper to the amyloid precursor protein is thought to prevent clea-vage by beta-secretase [47] while interaction between copper and amyloid-beta initiates polymerization [48] Recent studies have shown that Cu2+and Zn2+ bind-ing to amyloid-beta (1–40) peptide, in distinction to

Fe3+, retards amyloid fibril formation [4] There were also reports that prefibrillar aggregation was promoted

Trang 10

by these ions [41] which would not be beneficial in the

light of higher toxicity of such aggregates [3]

The discovery that copper binding to stefin B

inhib-its fibril formation but does not prevent aggregation to

prefibrillar oligomers is quite significant Both these

facts are in accordance with amyloid-beta [4] and prion

studies [41], respectively We propose that in globular

proteins which bind Cu2+specifically, the metal

bind-ing can be protective against amyloid fibril formation

Therefore, maintaining correct metal ion protein

inter-actions might be key to whether such proteins are able

to enter an amyloidogenic pathway However, copper

binding, most likely nonspecific, does not always

pre-vent prefibrillar aggregate formation, which may be

even more toxic [3]

Recently, Miranker and coworkers [49] indicated

that b-2 microglobulin aggregated more heavily in the

presence of Cu2+ (but not Ni2+) They have shown

that, due to high affinity copper binding to a

conform-ationally changed monomer M*, equilibrium is shifted

to more oligomers Thus, in their case, specific copper

binding to oligomers accelerated amyloid aggregation

(measured by ThT fluorescence) In our case, the

monomer and dimer seem to bind Cu2+, whereas the

tetramer looses this ability, which makes it logical that

the fibrillation (which starts with oligomer formation)

would be inhibited

Conclusion

We show two per se interesting and novel facts: (a)

that human stefin B (cystatin B) is a high affinity

copper binding protein, and (b) that amyloid fibril

formation of this protein is diminished in presence

of Cu2+ ions Another interesting finding is that

copper binding gets less strong at pH 5, under fibril

promoting conditions and that a mutant, which was

shown to be tetrameric, does not bind copper

Per-haps all these facts are not related and have no

rele-vance to in vivo function of the protein and even

less to amyloid fibrillation As for the function

[10,11,18], this is open to more research The protein

decreases apoptosis not only by protease inhibition,

as shown by gene knockout studies [20,50] As

apop-tosis is highly connected to either oxidative stress

and⁄ or protein aggregation, alternative function

(mis-function) of this protein could be researched in those

directions

A broader implication for future research is that

understanding what causes the loss of appropriate

metal binding might be crucial for the understanding

of the role of amyloidogenic proteins in a number of

neurodegenerative disorders

Experimental procedures

Materials 2,2,2 Trifluorethanol was from Fluka (Buchs, Switzerland) and thioflavin T from Aldrich (St Louis, MO, USA) Other chemicals were from Sigma (St Louis, MO, USA), Carlo Erba (Milano, Italy), Serva (Westbury, NY, USA) and Merck (Darmstadt, Germany)

Recombinant proteins Recombinant human stefin B variants were produced in Escherichia coliand isolated as described [51,52]

Isothermal titration calorimetry measurements All measurements were made on a Microcal VP-Isothermal Titration Calorimeter instrument as previously described [53] Briefly, a time course of injections of a ligand to a macromolecule or vice versa were made in an enclosed reaction cell maintained at a constant temperature The instrument measured the heat generated or absorbed as the ligand-macromolecule reaction occured A binding isotherm was fitted to the data, expressed in terms of the heat change per mole of ligand against the ligand to macromolecule ratio From the binding isotherm values for the reaction stoichiometry, association constants Ka, the change in enthalpies H and change in entropies S were obtained All solutions were filtered through a 0.22 lm filter and degassed prior to use All measurements were made in a buffer consisting of 5 mm Mes at either pH 5 or pH 7 Solutions were treated with the chelex medium to remove trace metals, according to the manufacturer’s instructions (Sigma)

Direct titration of protein solutions with aqueous copper salts was avoided because of the nonphysiological nature of such interactions; these may be avoided by the use of a copper chelate Copper(II) forms a bis glycine complex, Cu(Gly)2, in the presence of excess glycine Therefore a copper⁄ glycine ratio of 1 : 4 was used by dissolving 3.0 mm copper(II)chloride and 12 mm glycine in chelex treated water The excess glycine ensured that titrated copper was either chelated to glycine or incorporated into the protein, avoiding aqueous copper in the reaction cell It also acted

as a competitor to nonspecific protein copper interactions

A control ITC experiment of titrating stefin B protein with glycine alone was performed and no binding was observed

In our hands, as others [54], the most reproducible data was obtained from injections of the metal into a solution of the protein Typically an initial injection of 2 lL copper chelate solution was followed by a further 29 injections of

4 lL of Cu(II) into the protein in the sample cell stirred

at 300 r.p.m Injections were separated by 120 s to allow equilibration and sample temperature was maintained at

Ngày đăng: 19/02/2014, 06:20

TỪ KHÓA LIÊN QUAN

TÀI LIỆU CÙNG NGƯỜI DÙNG

  • Đang cập nhật ...

TÀI LIỆU LIÊN QUAN

🧩 Sản phẩm bạn có thể quan tâm