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: 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. 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 [...]... subtraction of the blank data a nonlinear least squares method was used to minimize v2 values and obtain best fit parameters for the association constants, Ka, and the change in enthalpies, H° In all cases best fit parameters were obtained from the sequential binding < /b> sites model, whereby the user defines the number of binding < /b> sites to be fitted in a sequential manner Attempts to fit data to anything other than... subsequent affinities reported for stefin < /b> protein binding < /b> are the product of the measured Cu -stefin < /b> interaction and 4.0 · 105 m)1 for pH 7 or 2.9 · 103 m)1 for pH 5 (K1 of CuGly2) As an independent method to confirm copper < /b> binding < /b> to the protein, saturation experiments were undertaken Namely, 10 lm of stefin < /b> B was exposed to 10, 20, 50 and 100 lm of Cu(II) as Cu2+-Gly chelate The protein was dialysed and. .. Characterization of copper < /b> interactions with alzheimer amyloid beta peptides: identification of an attomolaraffinity copper < /b> binding < /b> site on amyloid beta1–42 J Neurochem 75, 1219–1233 Borchardt T, Camakaris J, Cappai R, Masters CL, Beyreuther K & Multhaup G (1999) Copper < /b> inhibits beta -amyloid production and stimulates the non-amyloidogenic pathway of amyloid- precursor-protein secretion Biochem J 344, 461–467... recombinant human stefin < /b> B in complex with the cysteine proteinase papain: a novel type of proteinase inhibitor interaction EMBO J 9, 1939–1947 ˇ 13 Jenko S, Dolenc I, Guncar G, Dobersˇ ek A, Podobnik M & Turk D (2003) Crystal structure of stefin < /b> A in complex with cathepsin H: N-terminal residues of inhibitors can adapt to the active sites of endo- and exopeptidases J Mol Biol 326, 875–885 ˇ 14 Martin JR, Craven... 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 < /b> of cystatin C to Alzheimer’s amyloid beta inhibits in vitro amyloid fibril formation Neurobiol Aging 25, 1033–1043 34 Sanders... 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 lm or 150 lm Cu2SO4 dissolved in water was added to the buffers to saturate the protein, which was 45 lm After mixing the solutions, this gave protein to Cu2+ ratio of 1 : 1 and 1 : 3 Fibrillation in the presence of copper < /b> and without... followed using two isoforms of stefin < /b> 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 different 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. .. al Copper < /b> binds to cystatin B 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 correction was applied to each experiment by < /b> subtraction of data from a series of injections of copper < /b> chelate solution into a buffer blank correlating to the heat of dilution of the copper < /b> complex After subtraction... 39 Rabzelj S, Turk V & Zerovnik E (2005) In vitro study of stability and amyloid- fibril formation of two mutants of human stefin < /b> 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, SchulzSchaeffer W, Giese A, Westaway D & Kretzschmar H (1997) The cellular prion protein binds copper < /b> in vivo... & Fernandez CO (2005) Structural characterization of copper < /b> (II) binding < /b> to alpha-synuclein: Insights into the bioinorganic chemistry of Parkinson’s disease Proc Natl Acad Sci USA 102, 4294–4299 Hesse L, Beher D, Masters CL & Multhaup G (1994) The beta A4 amyloid precursor protein binding < /b> to copper < /b> FEBS Lett 349, 109–116 Atwood CS, Scarpa RC, Huang X, Moir RD, Jones WD, Fairlie DP, Tanzi RE & Bush AI . High affinity copper binding by stefin B (cystatin B) and its role in the inhibition of amyloid fibrillation Eva Z ˇ erovnik 1 ,. 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