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

High affinity copper binding by stefin B (cystatin B) and its
role in the inhibition of amyloid fibrillation
Eva Z
ˇ
erovnik
1
, Katja S
ˇ
kerget
1
, Magda Tus
ˇ
ek-Z
ˇ
nidaric
ˇ
2
, Corina Loeschner
3
, Marcus W. Brazier
3
and David R. Brown
3
1 Department of Biochemistry and Molecular Biology, Joz
ˇ
ef 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. I n 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 Cu
2+
and Zn
2+
, but
not Fe
3+
, 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. Z
ˇ
erovnik, Department of Biochemistry
and Molecular Biology, Joz
ˇ
ef Stefan
Institute, Jamova 39, 1000 Ljubljana,
Slovenia
Fax: +386 1477 3984
Tel: +386 1477 3753 ⁄ 3900
E-mail:
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:
(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.
4250 FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS
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, 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 Cu
2+
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
Cu
2+
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 binding site
parameters and the best fit, producing the smallest v
2

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 Cu
2+
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
E. Z
ˇ
erovnik et al. Copper binds to cystatin B
FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS 4251

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 Cu
2+
, 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 CuSO
4
. 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 Cu
2+

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 Cu
2+
. 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
-2
-4
-2
0
2
010203040506070
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
010203040506070
Time (min)
µcal/sec

µcal/sec
-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 A pH7
Stefin B pH7
Fig. 2. Comparison of copper binding by ste-
fin A and stefin B using ITC. Protein sam-
ples were prepared at 35 l
M in 5 mM Mes
pH 7 buffer. The ligand used for injection
was Cu
2
SO
4
equilibrated 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 y
nkkfpvfka vsfksqvvag tnyfikvhvg dedfvhlrvf qslphenkpl
P79S mmsgapsatq pataetqhia dqvrsqleek y
nkkfpvfka vsfksqvvag tnyfikvhvg dedfvhlrvf qslphenksl
StA mipgglseak patpeiqeiv dkvkpqleek tnetygklea vqyktqvvag tnyyikvrag dnkymhlkvf kslpgqnedl
Wt tlsnyqtnka khdeltyf
Var2 tlsnyqtnka khdeltyf
P79S tlsnyqtnka khdeltyf
StA vlt
gyq
vdkn kddelt
g
f
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.
Copper binds to cystatin B E. Z
ˇ
erovnik et al.
4252 FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS
dichroweb online software [29,30]. There was no dif-

ference when comparing the effects of Cu
2+
, 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 Cu
2+
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 Cu
2+
(Fig. 5E) may, similarly to
P79S (Fig. 4B), arise from enhanced protein aggrega-
tion rather than a conformational change.
To assess protein stability, thermal denaturation of
the protein in presence of Cu
2+
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
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
5m

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
site Stefin A Stefin B
Stefin B
(Variant 2)
Stefin B

(P79Sb)
pH 7
Site 1 nsd 7.2 · 10
10
8.0 · 10
10
nsd
Site 2 – 1.0 · 10
10
7.6 · 10
9
–
v
2
– 2821.11 2953.4 –
pH 5
Site 1 nsd 6.1 · 10
8
1.8 · 10
8
nsd
Site 2 – 1.4 · 10
8
1.1 · 10
8
–
v
2
– 5760.01 13065.2 –
E. Z

ˇ
erovnik et al. Copper binds to cystatin B
FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS 4253
that the protein aggregates heavily in the presence of
Cu
2+
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 Cu
2+
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 Cu
2+
, and after longer times (1 week at
room temperature and 4 days at 36 °C, 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 36 °C 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 Cu
2+

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, 25 °C, in the presence of
Cu
2+
or with no Cu
2+
present. Two concentrations of
CuSO
4
in the buffer were used (50 and 150 lm) giving
1 : 1 and 1 : 3 protein to Cu
2+
ratios.
Figure 6 and Table 2 show the outcome of the fibril-
lation assays with and without Cu
2+
present in the
medium. Fibrillation of stefin B wildtype (E31 isoform)
was first recorded at 40 °C, at pH 7, where no fibrilla-
tion was observed, and at pH 5 at 40 °C (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, 25 °C), are
plotted in Fig. 6B–D. It can be seen that Cu
2+
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 Cu
2+
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 Cu
2+
concentration, and is higher at
1 : 3 protein to Cu
2+
molar ratio than at 1 : 1. The
P79S mutant, which does not bind Cu
2+
and is tetra-

meric, also does not fibrillate to the same extent as
wildtype or variant 2, but some inhibition by Cu
2+
is
still observed (Fig. 6D, Table 2). Regardless, this does
not seem to depend on Cu
2+
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 Cu
2+
(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
190 200 210 220 230 240 250
nm
stefin A
stefin B
P79S
deg·cm
2–1
·dmol
–1
deg·cm
2–1
·dmol
–1
near UV CD spectra
-40
-20
0
20
40
60
80
100
120
250 260 270 280 290 300 310 320
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.
Copper binds to cystatin B E. Z
ˇ
erovnik et al.
4254 FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS
Discussion
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
200 210 220 230 240 250
nm
deg·cm
2–1
·dmol
deg·cm
2–1
·dmol
deg·cm
2–1
·dmol
wt stB with Cu
wt stB without
-6000
-4000
-2000
0
2000
4000
6000
8000
10000
12000
200 210 220 230 240 250
nm
var2 stB with Cu
var2 stB without
-3000
-2000

-1000
0
1000
2000
3000
4000
5000
6000
200 210 220 230 240 250
nm
P79S with Cu
P79S without
AB
C
deg·cm
2–1
·dmol
-8000
-6000
-4000
-2000
0
2000
4000
6000
200 210 220 230 240 250
nm
stA with Cu
stA without
D

mdeg
-2
-1
0
1
2
3
4
5
250 260 270 280 290 300 310 320
nm
pH7
pH7Cu
E
mdeg
-2
-1
0
1
2
3
4
5
250 260 270 280 290 300 310 320
nm
pH5
pH5Cu
F
Fig. 5. Circular dichroism spectroscopy as a function of Cu
2+

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 Cu
2+
. (D) Far UV CD spectra of stefin A in the presence and without Cu
2+
. (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 Cu
2+
and without Cu
2+
at pH 5.
E. Z
ˇ
erovnik et al. Copper binds to cystatin B

FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS 4255
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
0 5000 10000 15000 20000 25000
Time (min)
wt stB without
wt stB with Cu
pH=5, 25
o
C, 10%TFE
-200
0
200
400
600
800
1000
0 10000 20000 30000 40000
Time (min)
ThT fluorescence / 480 nm
ThT fluorescence / 480 nm
wt stB without
wt stB with Cu
BA
ThT fluorescence / 480 nm
ThT fluorescence / 480 nm
pH=5, 25
o

C, 10% TFE
-200
0
200
400
600
800
1000
0 10000 20000 30000 40000
Time (min)
var2 stB without
var2 stB with Cu
pH=5, 25
o
C, 10% TFE
-200
0
200
400
600
800
1000
0 10000 20000 30000 40000
Time (min)
mut P79S without
mut P79S with Cu
C
D
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 Cu
2+
46 lM, leading to 1 : 1 of protein to Cu
2+
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 lM 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 Cu
2+
in the buffer. (D) P79S mutant of variant 2 at pH 5, 10% TFE, 25oC, 0 and 50 lM Cu
2+
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 Cu
2+
in solution were 46 lM and 138 lM, which gives 1 : 1

and 1 : 3 protein to Cu
2+
molar ratios, respectively.
Protein ⁄ variant [Cu
2+
](lM) Solvent composition % of inhibition
Stefin B E31 50 10% TFE, pH 5 63 ± 12
150 10% TFE, pH 5 80 ± 2
Stefin B Y31 50 10% TFE, pH 5 41 ± 10
150 10% TFE, pH 5 66 ± 2
P79S mutant 50 10% TFE, pH 5 58 ± 10
150 10% TFE, pH 5 62 ± 3
Stefin B E31 50 pH 5 32 ± 10
50 pH 7 0
Stefin B Y31 50 pH 5 50.5 ± 10
50 pH 7 0
Copper binds to cystatin B E. Z
ˇ
erovnik et al.
4256 FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS
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 number
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 lM 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 Cu
2+
. (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.
E. Z
ˇ
erovnik et al. Copper binds to cystatin B
FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS 4257
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
Cu
2+
of amyloid fibril formation of human stefin B
could result from loss of correct Cu
2+
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
Cu
2+
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 Cu
2+
(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 Cu
2+
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 Cu
2+
and Zn
2+
bind-
ing to amyloid-beta (1–40) peptide, in distinction to
Fe
3+
, retards amyloid fibril formation [4]. There were
also reports that prefibrillar aggregation was promoted
Copper binds to cystatin B E. Z
ˇ
erovnik et al.
4258 FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS
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 Cu
2+
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 Cu
2+
(but not Ni
2+
). 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 Cu
2+
, 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 Cu
2+
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 coli and 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 K
a
, 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
E. Z
ˇ
erovnik et al. Copper binds to cystatin B
FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS 4259
25 °C. All experiments were repeated at least three times.
Data were analyzed with the origin 5.0 software package
from MicroCal (Northampton, MA, USA). A baseline cor-
rection was applied to each experiment by subtraction of

data from a series of injections of copper chelate solution
into a buffer blank correlating to the heat of dilution of the
copper complex. After subtraction of the blank data a non-
linear least squares method was used to minimize v
2
values
and obtain best fit parameters for the association constants,
K
a
, and the change in enthalpies, H°. In all cases best fit
parameters were obtained from the sequential binding sites
model, whereby the user defines the number of binding sites
to be fitted in a sequential manner. Attempts to fit data to
anything other than one or two identical site models gave
unsatisfactory results. The affinity of copper to glycine was
measured in water, pH 7, with identical conditions to the
protein experiments. The best fit model for the coordination
of two glycine molecules to copper, corresponding to the
CuGly
2
complex results in K
1
¼ 4.0 · 10
5
m
)1
and K
2
¼
1.7 · 10

4
m
)1
with b2 ¼ 6.8 · 10
9
m
)1
(data not shown) in
agreement with the pH adjusted literature values of K
1
¼
3.0 · 10
5
m
)1
and b2 ¼ 8.7 · 10
9
m
)1
[55]. All subsequent
affinities reported for stefin protein binding are the product
of the measured Cu-stefin interaction and 4.0 · 10
5
m
)1
for
pH 7 or 2.9 · 10
3
m
)1

for pH 5 (K
1
of CuGly
2
).
As an independent method to confirm copper binding to
the protein, saturation experiments were undertaken.
Namely, 10 lm of stefin B was exposed to 10, 20, 50 and
100 lm of Cu(II) as Cu
2+
-Gly chelate. The protein was dia-
lysed and the amount of Cu
2+
determined per molecule of
stefin. The assay used was that published previously [53]. At
10 lm Cu
2+
1.3 ± 0.1 atoms per molecule bound. At 20, 50
and 100 lm amount was 2.1 ± 0.1 atoms per molecule. Ste-
fin A did not bind any copper under these conditions.
CD spectroscopy
CD spectra were measured using an Aviv model 62A DS
CD spectropolarimeter equipped with a thermoelectric sam-
ple holder for temperature control in the cell (AVIV Biome-
dical, Lakewood, NJ, USA). Temperature was set at 25 °C.
Data were collected every 1 nm and bandwidth was 1 nm
in the far UV while data were collected every 0.5 nm at a
bandwidth of 0.5 nm in the near UV. When using 1 mm
rectangular cell, protein concentrations were around
0.2 mgÆmL

)1
(18 lm), with exact readings from A
280
. For
the spectra acquired to 190 nm, 0.2 mm mountable cell was
used with protein concentration five times higher. When
recording in the near UV, a 10 mm rectangular microcell
was used at a protein concentration of around
1–1.5 mgÆmL
)1
(90–136 lm). Far UV CD spectra were
expressed in the units of the mean residue ellipticity
[Q]
MRW
(degÆcm
2
Ædmol
)1
). Thermal denaturation scans were
performed by the software provided, raising the tempera-
ture in 1 °C steps and collecting data for 15 s at each
degree. Recording was done at 222 and 210 nm but only
the signal at 210 nm changed upon heating.
Additional near UV CD spectra measurements were per-
formed using Applied Photophysics Chirascan (Applied
Photophysics, Leatherhead, UK). A 10 mm rectangular
microcell was used with a protein concentration of
1.1 mgÆmL
)1
(100 lm). The spectra were recorded at 20 °C

at a bandwidth of 0.5, averaging the signal for 3 s every
0.5 nm. These were left in the units of mdeg.
CD spectra analysis
CD spectra recorded to 190 nm were analyzed using online
Circular Dichroism Analysis software, dichroweb [29,30]
( />Fibrillation assays
The buffers used were 0.015 m acetate, 0.15 m NaCl,
pH 5.0, NaCl ⁄ Pi pH 7.3 (or 0.01 m phosphate buffer
pH 7.0, 0.15 m NaCl) and the pH 5 buffer with added TFE
to get a final 10% (v ⁄ v) concentration of TFE. Chelating
buffers were prepared using chelex medium. The protein
solutions were exchanged with the chelating buffer of pH 7
prior to fibrillation assays. Either 50 l m or 150 lm Cu
2
SO
4
dissolved in water was added to the buffers to saturate the
protein, which was 45 l m. After mixing the solutions, this
gave protein to Cu
2+
ratio of 1 : 1 and 1 : 3.
Fibrillation in the presence of copper and without this
ion was followed using two isoforms of stefin B, one with E
at site 31 (wildtype) and the other with Y at site 31 (variant
2), and the mutant P79S of the variant, in the three differ-
ent buffers as described above. Samples in NaCl ⁄ Pi buffer
(pH 7.3) and in acetic buffer (pH ¼ 5.0) were thermostated
at 40 °C, while samples in acetic buffer with 12% (v ⁄ v)
TFE (10% TFE final concentration) were incubated at
25 °C. The process of fibrillation was followed for 14 days.

ThT fluorescence measurements
ThT dye was used to determine the presence of amyloid-
like fibrils. ThT was dissolved in phosphate buffer (25 mm,
0.1 m NaCl, pH 7.5) at A
416
¼ 0.66.
Assay solutions contained 10 lL of protein at concentra-
tion 45 lm and 114 lL of ThT solution. A volume of
124 lL of the mixture was pipetted into a well of a 96-well
plate (white plastic). The plate was loaded into a floures-
cence plate reader (Tecan, Safire, Austria) and the program
xflour4 was used to obtain data. The fluorescence was
measured with excitation at 440 nm and emission at
480 nm, with an integration time of 40 ls. Relative fluores-
cence units were plotted vs. time. Each experiment was
repeated at least three times. Control experiment with ThT
buffer alone with and without the same amount of Cu
2+
Copper binds to cystatin B E. Z
ˇ
erovnik et al.
4260 FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS
added was also carried out and there was no quenching of
ThT fluorescence observed.
Transmission electron microscopy
Protein samples (20 lLof34lm protein solution, diluted if
appropriate) were applied on a Formvar and carbon coated
grid. After 5 min the sample was soaked away and stained
with 1% (v ⁄ v) uranyl acetate. Samples were observed with
a Philips (Amsterdam, the Netherlands) CM 100 transmis-

sion electron microscope at 80 kV, with magnifications
from 10 000· to 130 000·. Images were recorded by Bio-
scan CCD camera using digital micrograph software,
Gatan Inc. (Washington, DC, USA).
Size-exclusion chromatography
Oligomers present in the samples of the wildtype protein
were determined by the size exclusion chromatography
using Superdex 75 (Amersham Pharmacia Biotech, Piscat-
away, NJ, USA) column on an AKTA FPLC system
(Amersham Pharmacia Biotech). First, monomeric stefin B
wildtype (E31 isoform) was prepared by collecting the cor-
responding peak from the Superdex 75 column, then, the
monomer was incubated (see above) in the presence of
Cu
2+
and without Cu
2+
, at pH 5 and pH 7. The mobile
phase was 0.01 m phosphate buffer, containing 0.12 m
NaCl at pH 7, flow rate was set at 0.7 mLÆmin
)1
and elu-
tion peaks were detected by UV absorbance at 280 nm.
Acknowledgements
We thank Louise Kroon Z
ˇ
itko and Manca Kenig for
cloning and isolating the recombinant proteins. We
also are grateful to Sabina Rabzelj and Sas
ˇ

a Jenko
Kokalj for performing certain SEC experiments and
for activity measurements. We are grateful to Professor
Roger H Pain for suggesting improvements in the lan-
guage usage. This work was partially funded by the
grant OB14PO4SK (proteolysis and regulation) from
the Slovenian Ministry for Science and Technology.
References
1 Guijarro JI, Sunde M, Jones JA, Campbell ID &
Dobson CM (1998) Amyloid fibril formation by an SH3
domain. Proc Natl Acad Sci USA 95, 4224–4228.
2 Dobson CM (1999) Protein misfolding, evolution and
disease. Trends Biochem Sci 24, 329–332.
3 Bucciantini M, Giannoni E, Chiti F, Baroni F, Formigli
L, Zurdo J, Taddei N, Ramponi G, Dobson CM &
Stefani M (2002) Inherent toxicity of aggregates implies
a common mechanism for protein misfolding diseases.
Nature 416, 507–511.
4 Raman B, Ban T, Yamaguchi K, Sakai M, Kawai T,
Naiki H & Goto Y (2005) Metal ion-dependent effects
of clioquinol on the fibril growth of an amyloid {beta}
peptide. J Biol Chem 280, 16157–16162.
5Z
ˇ
erovnik E, Pompe-Novak M, S
ˇ
karabot M, Ravnikar
M, Mus
ˇ
evic

ˇ
I & Turk V (2002) Human stefin B readily
forms amyloid fibrils in vitro. Biochim Biophys Acta
1594, 1–5.
6 Jenko S, S
ˇ
karabot M, Kenig M, Gunc
ˇ
ar G, Mus
ˇ
evic
ˇ
I,
Turk D & Z
ˇ
erovnik E (2004) Different propensity to
form amyloid fibrils by two homologous proteins-
Human stefins A and B: searching for an explanation.
Proteins 55, 417–425.
7 Matsunaga Y, Z
ˇ
erovnik E, Yamada T & Turk V (2002)
Conformational changes preceding amyloid-fibril forma-
tion of amyloid-beta and stefin B; parallels in pH
dependence. Curr Med Chem 9, 1717–1724.
8 Anderluh G, Gutierrez-Aguirre I, Rabzelj S, C
ˇ
eru S,
Kopitar-Jerala N, Mac
ˇ

ek P, Turk V & Z
ˇ
erovnik E
(2005) Interaction of human stefin B in the prefibrillar
oligomeric form with membranes. Correlation with cel-
lular toxicity. FEBS J 272, 3042–3051.
9 Rawlings ND, Tolle DP & Barrett AJ (2004) MEROPS:
the peptidase database. Nucleic Acids Res 32, D160–
D164.
10 Turk V & Bode W (1991) The cystatins: protein inhibi-
tors of cysteine proteinases. FEBS Lett 285, 213–219.
11 Turk V, Turk B & Turk D (2001) Lysosomal cysteine
proteinases: facts and opportunities. EMBO J 20, 4629–
4633.
12 Stubbs MT, Laber B, Bode W, Huber R, Jerala R,
Lenarc
ˇ
ic
ˇ
B & Turk V (1990) The refined 2.4 A
˚
X-ray
crystal structure of recombinant human stefin B in
complex with the cysteine proteinase papain: a novel
type of proteinase inhibitor interaction. EMBO J 9,
1939–1947.
13 Jenko S, Dolenc I, Gunc
ˇ
ar G, Dobers
ˇ

ek A, Podobnik
M & Turk D (2003) Crystal structure of stefin A in
complex with cathepsin H: N-terminal residues of inhi-
bitors can adapt to the active sites of endo- and exo-
peptidases. J Mol Biol 326, 875–885.
14 Martin JR, Craven CJ, Jerala R, Kroon-Z
ˇ
itko L,
Z
ˇ
erovnik E, Turk V & Waltho JP (1995) The three-
dimensional solution structure of human stefin A. J Mol
Biol 246, 331–343.
15 Jerala R & Z
ˇ
erovnik E (1999) Accessing the global
minimum conformation of stefin A dimer by annealing
under partially denaturing conditions. J Mol Biol 291,
1079–1089.
16 Staniforth RA, Giannini S, Higgins LD, Conroy MJ,
Hounslow AM, Jerala R, Craven CJ & Waltho JP
(2001) Three-dimensional domain swapping in the
folded and molten-globule states of cystatins, an amy-
loid-forming structural superfamily. EMBO J 20, 4774–
4781.
E. Z
ˇ
erovnik et al. Copper binds to cystatin B
FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS 4261
17 Janowski R, Kozak M, Jankowska E, Grzonka Z,

Grubb A, Abrahamson A & Jaskolski M (2001) Human
cystatin C, an amyloidogenic protein, dimerizes through
three-dimensional domain swapping. Nat Struct Biol 8,
316–320.
18 Di Giamo R, Riccio M, Santi S, Galeotti C, Ambrosetti
DC & Melli M (2002) New insights into the molecular
basis of progressive myoclonus epilepsy: a multiprotein
complex with cystatin B. Hum Mol Genet 11, 2941–2950.
19 Riccio M, Di Giaimo R, Pianetti S, Palmieri PP, Melli
M & Santi S (2001) Nuclear localization of cystatin B,
the cathepsin inhibitor implicated in myoclonus epilepsy
(EPM1). Exp Cell Res 262, 84–94.
20 Pennacchio LA, Bouley DM, Higgins KM, Scott MP,
Noebels JL & Myers RM (1998) Progressive ataxia,
myoclonic epilepsy and cerebellar apoptosis in cystatin
B-deficient mice. Nat Genet 20, 251–258.
21 Lieuallen K, Pennacchio LA, Park M, Myers RM &
Lennon GG (2001) Cystatin B-deficient mice have
increased expression of apoptosis and glial activation
genes. Hum Mol Genet 10, 1867–1871.
22 Pennacchio LA, Lehesjoki AE, Stone NE, Willour VL,
Virtaneva K, Miao J, D’Amato E, Raminez L, Faham
M, Koskiniemi M et al. (1996) Mutations in the gene
encoding cystatin B in progressive myoclonus epilepsy
(EPM1). Science 271, 1731–1734.
23 Alakurtti K, Weber E, Rinne R, Theol G, Haan G,
Lindhout D, Salmikangas P, Saukko P, Lahtinen U &
Lehesjoki A (2004) Loss of lysosomal association of
cystatin b proteins representing progressive myoclonus
epilepsy, EPM1, mutations. Eur J Hum Genet 13, 208–

215.
24 D’Amato E, Kokaia Z, Nanobashvili A, Reeben M,
Lehesjoki AE, Saarma M & Lindvall O (2000) Seizures
induce widespread upregulation of cystatin B, the gene
mutated in progressive myoclonus epilepsy, in rat fore-
brain neurons. Eur J Neurosci 12, 1687–1695.
25 Pirttila
¨
TJ, Lukasiuk K, Ha
˚
kansson K, Grubb A,
Abrahamson M & Pitka
¨
nen A (2005) Cystatin C modu-
lates neurodegeneration and neurogenesis following
status epilepticus in mouse. Neurobiol Dis 20, 241–253.
26 Deng A, Irizarry MC, Nitsch RM, Growdon JH &
Rebeck GW (2001) Elevation of Cystatin C in Suscepti-
ble Neurons in Alzheimer’s Disease. Am J Pathol 159,
1061–1068.
27 Ii K, Hidehumi I, Kominami E & Hirano A (1993)
Abnormal distribution of cathepsin proteinases and
endogenous inhibitors (cystatins) in the hippocampus of
patients with Alzheimer’s disease, parkinsonism-dementia
complex on Guam, and senile dementia and in the aged.
Virchows Arch a Pathol Anat Histopathol 423, 185–194.
28 Manning MC & Woody RW (1989) Theoretical study
of the contribution of aromatic side chains to the circu-
lar dichroism of basic bovine pancreatic trypsin inhibi-
tor. Biochemistry 28, 8609–8613.

29 Whitmore L & Wallace BA (2004) DICHROWEB, an
online server for protein secondary structure analyses
from circular dichroism spectroscopic data. Nucleic
Acids Res 32, W668–W673.
30 Lobley A, Whitmore L & Wallace BA (2002) DICHRO-
WEB: an interactive website for the analysis of protein
secondary structure from circular dichroism spectra.
Bioinformatics 18, 211–212.
31 Z
ˇ
erovnik E, Jerala R, Kroon-Z
ˇ
itko L, Pain RH & Turk
V (1992) Intermediates in denaturation of a small globu-
lar protein, recombinant human stefin B. J Biol Chem
267, 9041–9046.
32 Jensson O, Palsdottir A, Thorsteinsson L, Arnason A,
Abrahamson M, Olafsson I & Grubb A (1990) Cystatin
C mutation causing amyloid angiopathy and brain
hemorrhage. Biol Chem Hoppe Seyler 371
(Suppl.),
229–232.
33 Sastre M, Calero M, Pawlik M, Mathews PM, Kumar
A, Danilov V, Schmidt SD, Nixon RA, Frangione B &
Levy E (2004) Binding of cystatin C to Alzheimer’s
amyloid beta inhibits in vitro amyloid fibril formation.
Neurobiol Aging 25, 1033–1043.
34 Sanders A, Craven JC, Higgins LD, Giannini S, Conroy
MJ, Hounslow AM, Waltho JP & Staniforth RA (2004)
Cystatin forms a tetramer through structural rearrange-

ment of domain-swapped dimers prior to amyloidogen-
esis. J Mol Biol 336, 165–178.
35 Modler AJ, Gast K, Lutsch G & Damaschun G (2003)
Assembly of amyloid protofibrils via critical oligomers –
a novel pathway of amyloid formation. J Mol Biol 325,
135–148.
36 Knaus KJ, Morillas M, Swietnicki W, Malone M,
Surewicz WK & Yee VC (2001) Crystal structure of the
human prion protein reveals a mechanism for oligomeri-
zation. Nat Struct Biol 8, 770–774.
37 Walsh DM, Hartley DM, Kusumoto Y, Fezoui Y,
Condron MM, Lomakin A, Benedek GB, Selkoe DJ &
Teplow DB (1999) Amyloid beta-protein fibrillogenesis.
Structure and biological activity of protofibrillar inter-
mediates. J Biol Chem 274, 25945–25952.
38 Lashuel HA, Hartley D, Petre BM, Walz T & Lansbury
PT (2002) Amyloid pores pathogenic mutations. Nature
418, 291–291.
39 Rabzelj S, Turk V & Z
ˇ
erovnik E (2005) In vitro study
of stability and amyloid-fibril formation of two mutants
of human stefin B (cystatin B) occurring in patients with
EPM1. Protein Sci 14, 2713–2722.
40 Brown DR, Qin K, Herms JW, Madlung A, Manson J,
Strome R, Fraser PE, Kruck T, von Bohlen A, Schulz-
Schaeffer W, Giese A, Westaway D & Kretzschmar H
(1997) The cellular prion protein binds copper in vivo.
Nature 390, 684–687.
41 Qin K, Yang DS, Yang Y, Chishti MA, Meng LJ,

Kretzschmar HA, Yip CM, Fraser PE & Westaway D
(2000) Copper (II)-induced conformational changes and
Copper binds to cystatin B E. Z
ˇ
erovnik et al.
4262 FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS
protease resistance in recombinant and cellular PrP.
Effect of protein age and deamidation. Biol Chem 275,
19121–19131.
42 Tsenkova RN, Iordanova IK, Toyoda K & Brown DR
(2004) Prion protein fate governed by metal binding.
Biochem Biophys Res Commun 325, 1005–1012.
43 Brown DR, Hafiz F, Glasssmith LL, Wong B-S, Jones
IM, Clive C & Haswell SJ (2000) Consequences of man-
ganese replacement of copper for prion protein function
and proteinase resistance. EMBO J 19, 1180–1186.
44 Rasia RM, Bertoncini CW, Marsh D, Hoyer W, Cherny
D, Zweckstetter M, Griesinger C, Jovin TM & Fernan-
dez CO (2005) Structural characterization of copper (II)
binding to alpha-synuclein: Insights into the bioinor-
ganic chemistry of Parkinson’s disease. Proc Natl Acad
Sci USA 102, 4294–4299.
45 Hesse L, Beher D, Masters CL & Multhaup G (1994)
The beta A4 amyloid precursor protein binding to cop-
per. FEBS Lett 349, 109–116.
46 Atwood CS, Scarpa RC, Huang X, Moir RD, Jones
WD, Fairlie DP, Tanzi RE & Bush AI (2000) Chara-
cterization of copper interactions with alzheimer
amyloid beta peptides: identification of an attomolar-
affinity copper binding site on amyloid beta1–42.

J Neurochem 75, 1219–1233.
47 Borchardt T, Camakaris J, Cappai R, Masters CL,
Beyreuther K & Multhaup G (1999) Copper inhibits
beta-amyloid production and stimulates the non-amyloi-
dogenic pathway of amyloid-precursor-protein secretion.
Biochem J 344, 461–467.
48 Atwood CS, Moir RD, Huang X, Scarpa RC, Bacarra
ME, Romano DM, Hartshorn MA, Tanzi RE & Bush
AI (1998) Dramatic aggregation of Alzheimer Abeta by
Cu
2+
(II) is induced by conditions representing physio-
logical acidosis. J Biol Chem 273, 12817–12826.
49 Eakin CM, Berman AJ & Miranker AD (2005) A native
to amyloidogenic transition regulated by a backbone
trigger. Nat Struct Biol 13, 202–208.
50 Houseweart MK, Pennacchio LA, Vilaythong A, Peters
C, Noebels JL & Myers RM (2003) Cathepsin B but
not cathepsins L or S contributes to the pathogenesis of
Unverricht-Lundborg progressive myoclonus epilepsy
(EPM1). J Neurobiol 56, 315–327.
51 Ritonja A, Machleidt W & Barrett AJ (1985) Amino
acid sequence of the intracellular cysteine proteinase
inhibitor cystatin B from human liver. Biochem Biophys
Res Commun 131, 1187–1192.
52 Jerala R, Trstenjak M, Lenarc
ˇ
ic
ˇ
B & Turk V (1988)

Cloning a synthetic gene for human stefin B and its
expression in E. coli. FEBS Lett 239, 41–44.
53 Thompsett AR, Abdelraheim SR, Daniels M & Brown
DR (2005) High affinity binding between copper and
full-length prion protein identified by two different tech-
niques. J Biol Chem 280, 42750–42758.
54 Zhang Y, Akilesh S & Wilcox DE (2000) Isothermal
titration calorimetry measurements of Ni (II) and Cu
2+
(II) binding to His, GlyGlyHis, HisGlyHis, and bovine
serum albumin: a critical evaluation. Inorg Chem 39,
3057–3064.
55 Dawson RMC, Elliot DC, Elliot DH & Jones KM
(1986) Data for Biochemical Research. 3rd edn. Oxford
University Press, Oxford.
E. Z
ˇ
erovnik et al. Copper binds to cystatin B
FEBS Journal 273 (2006) 4250–4263 ª 2006 The Authors Journal compilation ª 2006 FEBS 4263