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

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

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].

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

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

) 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-

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

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

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+

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-

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

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

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+

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

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+

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.

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-

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-

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

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).

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

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

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.

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

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 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,

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.

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,

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)