Structure and Function of Stefin B Oligomers – Important Role in Amyloidogenesis

Many functional proteins act as oligomers. Oligomerization is a well-controlled and regulated process. Those proteins cannot adopt their physiological functions without oligomerization, while the protein misfolding and subsequent unphysiological oligomerization influence the primary protein functions and produce a “gain in toxic function” as the prefibrillar oligomers are toxic for the cells. Misfolding, oligomerization and aggregation are the reasons for the so called conformational diseases. Several of them are neurodegenerative, but some also affect other vital organs. The accumulation of intracellular protein aggregates (various inclusions) and of extracellular protein deposits cause severe cellular degeneration, such as neurodegen‐ eration of affected neurons. Different proteins form rather similar but not identical fibrillar structures, all showing cross-β-structure, where continuous β-sheets run perpendicular to the long axis of the fibrils.


Introduction
Many functional proteins act as oligomers. Oligomerization is a well-controlled and regulated process. Those proteins cannot adopt their physiological functions without oligomerization, while the protein misfolding and subsequent unphysiological oligomerization influence the primary protein functions and produce a "gain in toxic function" as the prefibrillar oligomers are toxic for the cells. Misfolding, oligomerization and aggregation are the reasons for the so called conformational diseases. Several of them are neurodegenerative, but some also affect other vital organs. The accumulation of intracellular protein aggregates (various inclusions) and of extracellular protein deposits cause severe cellular degeneration, such as neurodegeneration of affected neurons. Different proteins form rather similar but not identical fibrillar structures, all showing cross-β-structure, where continuous β-sheets run perpendicular to the long axis of the fibrils.
Neurodegenerative diseases like Alzheimer's disease, Parkinson's disease, Huntington's disease, prion diseases and many others have in common protein aggregation to amyloid fibrils. Type II diabetes is also an amyloid disease although it is not neurodegenerative. Nowadays it is believed that ordered prefibrillar oligomers or protofibrils may be responsible for cell death and that mature fibrils may even be neuroprotective. Most amyloid prone proteins form different oligomers in the lag phase of the amyloid fibril formation. Conformational diseases (Table 1) are difficult to diagnose in the early stage, because they are usually asymptomatic during their development. Even when they can be diagnosed in the early stage there is an ethical reason for avoiding diagnosis (i.e. to hesitate letting know the patient or the relatives), because there are still no therapies which would slow down or even stop the progression of this type of diseases. Many different proteins are being studied in order to understand the common molecular mechanism of the conformational disease and to develop Both the amyloid forming proteins involved in certain conformational disease and model proteins are used ( Table 2). Stefin B is a model protein for studying amyloid fibril formation.

Human stefin B
Stefins are endogenous cysteine protease inhibitors [3], which are ubiquitously expressed in human tissues [4]. They are specific for the papain-family of cysteine proteases and classified as the IH clan in the MEROPS scheme [5]. They are mainly intracellular inhibitors, although have also been found outside the cell in body fluids [6]. They do not have a signal peptide, and they bear a cystatin motif QXVXG, which is the main site involved in binding to target enzymes.
Human stefin B (also termed cystatin B) possesses 98 amino acid residues (Mr= 11 kDa) and no carbohydrate groups or disulphide bridges [7], although it contains one free cysteine [8]. It is an intracellular protein, which tightly and reversibly binds to papain-like cysteine proteases. We will simply call it stefin B from now on.
Stefin B main function is protection against inappropriate proteolysis of lysosomal cysteine proteases [9]. It is an inhibitor of cathepsins B, L, H and S [7,10,11]. However, it exerts some additional functions. It was found to interact with five known non-protease proteins (neurofilament light chain (NFL), brain β-spectrin, RACK-1, human myotubularin related protein 8 (Mtrp) and human T-cell activation protein (Tcrp)) [12]. NFL and β-spectrin are specific to the nervous system. Stefin B (cystatin B) knock-out mice show neurological disorder (loss of the cerebellar granule cells, because of apoptotic bodies, chromatin condensation and some other changes). This suggests that stefin B has an essential anti apoptotic role in the cerebellum [12].
Stefin B is overexpressed in patients with hepatocellular carcinoma and is therefore in combination with some other proteins used as a marker for this disease (it is elevated in the serum already at the early stage of the disease development and therefore easy to detect) [13].
In different types of cancer (human colorectal cancer [14], gastric cancer [15], esophageal carcinoma [16] Stefin B functions are summarized in Table 3. Loss of functions because of alterations in the cystatin B gene (dodecameric repeat expansions in the promoter region or point mutations in the coding gene) is the cause of progressive myoclonus epilepsy of type 1 also known as Unverricht-Lundborg disease (EPM1) [25,28]. EPM1 mutans are polymeric and aggregate prone in vivo [29]. In vitro and ex vivo G4R, R68X, [30,31], G50E and Q71P [32,33] have been studied. Only G4R mutant folds like the wild type protein [31], all others lack tertiary structure and are partially unfolded [31,33]. G4R and R68X also form amyloid fibrils [31]. Both G50E and Q71P mutants are more susceptible to cleavages by proteases, because of the partially unfolded structure [33], which likely contributes to loss of function in cells. Stefin B deficiency triggers neurodegeneration by impaired redox homeostasis [26]. G4R and R68X do not have any inhibitory activity, while G50E and Q71P are much less active than the wild type protein [32]. All three missense mutants, except for G4R form rather large but diffuse aggregates in cells when over-expressed, while G4R forms small aggregates similar to those of the wild type protein [32].
Cystatin C, a secreted and extracellular protein, is another cysteine protease inhibitor and mutations in its gene are the cause of hereditary cerebral amyloid angiopathy [34]. Human cystatin C is a risk factor for late onset Alzheimer's disease. The protein co-deposits with amyloid β peptide (Aβ peptide) amyloid plaques in patients with Alzheimer's disease [35].
Moreover, cystatin C binds to both the whole amyloid precursor protein (APP) and to Aβ peptide and it inhibits Aβ peptide amyloid fibril formation in vitro [36].

Stefin B functions
inhibition of cathepsins B, L, H and S [7,10,11]; interaction with five non protease proteins: neurofilament light chain (NFL), brain β-spectrin, RACK-1, human myotubularin related protein 8 (Mtrp) and human T-cell activation protein (Tcrp) (NFL and β-spectrin are specific to the nervous system) -essential anti apoptotic role in the cerebellum [12]; involved in immune-response to bacterial challenge of the leech

Stefin B oligomers in vitro and in cells
Stefin B can adopt different oligomeric states in vitro and in cells. On the size-exclusion chromatography (SEC) the wild type protein elutes as a set of well-defined oligomers apart from monomers, dimers, tetramers and even higher oligomers ( Figure 1) [39]. Y31 isoform is predominantly dimeric [31], while Y31 P79S mutant is tetrameric [40]. All oligomers can be isolated as separate peaks by SEC and stay stable for weeks at pH 7.0 and 4 °C [38].
SEC results have been confirmed up to decamers by electrospray-ionization mass spectrometry (ESI-MS) [41]. In cells stefin B is present both in monomeric and oligomeric forms. Oligomers size ranges between 10 and 250 kDa. The higher oligomeric species are resistant to 1% SDS and 8 M urea and partially resistant to reducing agents (DTT treatment). The low molecular species comprise monomers, dimers, trimers and pentamers. Stefin B polymers in vivo seem to grow by monomer addition and not by domain-swapped dimer addition. The protein binds to many different proteins of various sizes [29]. Already the endogenous stefin B forms small punctate aggregates in cells and after overexpression the aggregates amount increases [30].
Oligomers have been observed for many other proteins, especially the amyloidogenic ones. Oligomeric species appear in the lag phase of the reaction and are usually more stable under non-amyloidogenic conditions. Size and morphology of those oligomers can be determined by SEC, mass spectrometry, dynamic light scattering and analytical ultracentrifugation. Some examples are given below.
Serpins are serine protease inhibitors also used as a model for studies of amyloid fibril formation. They have key regulatory functions in the inflammatory, complement, coagulation and fibrinolytic cascades [2]. The mechanism of their fibril formation is known in crystallographic details [42,43]. They form amyloid fibrils through domain-swapping [44]. Serpins preferably oligomerize than interact with other proteins [45]. They form both smaller oligomers and condensed longer polymers [46]. Monomers of some serpins are meta-stable with the rate limiting step representing the transition to the dimer. Once dimers are formed, they can connect to each other to form tetramers or recruit monomers to form trimers and also much longer oligomers [47].
Aβ peptide forms a set of oligomers from monomer to hexadecamer; all even and odd numbered oligomers have been detected by ESI MS [48]. Also for many other amylod forming proteins the whole set of oligomeric species has been observed (insulin [49], β 2 -microglobulin [50,51], SH3 domain [52]…).

Inhibitory activity of stefin B oligomers
For a long time it has been thought that monomer is the only active form of stefin B. All other oligomeric forms would serve as a reservoir of monomers or possess additional functions. Now we have shown that all types of oligomers are active against cysteine protease papain ( Figure 2). Dimers which are presumably domain-swapped (discussed in the structural characterization part) show the lowest activity in the enzyme to inhibitor 1:2 ratio. They, as well as other oligomers -tetramers and higher, are fully active in the enzyme to inhibitor 1:4 ratio (Fig. 2). SEC-isolated oligomers are all >95% pure and the amount of monomer present in other oligomers samples has been tested to show that the inhibitory effect is not due to monomer contamination. Higher oligomers are a mixture of various oligomers, starting from hexamers. The highest peak represents octamers, which are the prevalent species and therefore we propose that the average size of the higher oligomer is an octamer (this was used for determining the molar ratio). The sample of the higher oligomers contains a small amount of all other oligomers and we suspect that some inhibitory effect could be due to this contamination, but possibly not the whole effect. The results are in agreement with the detection of different oligomers in cells [29] and explain that all detected species could retain inhibitory function.

Structural characterization of different oligomeric states of stefin B
Crystal or solution structure of stefin B monomer and oligomers have been so far characterized from monomer in complex with the target enzyme [54], dimer under amyloid forming conditions [55], tetramer [40] to model of amyloid fibrils [56]. Domain-swapped dimer of stefin A has also been determined by heteronuclear NMR [57].
Stefin B was first crystallized in complex with the cysteine protease papain ( Figure 3A) [54]. The structure of the monomeric stefin B consists of a five stranded β-sheet wrapped around a five turn α-helix and with an additional carboxy terminal strand running along the convex side of the β-sheet. This type of the tertiary structure is conserved throughout the cystatin family (Fig. 3A) High resolution structure of stefin B dimer was determined by heteronuclear NMR [55]. When the proline at position 79 was mutated into serine, the tetrameric form of the protein became favourable [40]. This tetramer has been crystallized and its structure determined ( Figure 3B) [40]. The tetramer consists of two domain-swapped dimers connected with the socalled "hand-shaking" mechanism to each other. Hand-shaking occurs along with trans to cis isomerization of the proline at position 74. This proline is conserved throughout the cystatin family. Both domains preserve the fold of the monomer (the same as observed in the complex with papain); each domain is composed of the N-terminal part from one chain (strand 1, the α-helix and strand 2) and C-terminal part from the other chain (strands 3, 4 and 5), between both domains there is a linker region (two peptides, belonging each to one chain).
The structured core of human stefin B amyloid fibrils has been determined ( Figure 3C) [56]. Based on the H/D exchange rates the structure could be divided onto protected region (inside the fibril core) and unprotected region (striking out of the fibril). Strands 2, 3, 4 and 5 are protected, while strand 1, α-helix and loops between strands 3 and 4 and 4 and 5 are unprotected. The loop between strands 2 and 3 is the same as the one involved in the domainswapping mechanism. The fibril core would therefore be made of domain-swapped dimers but with the loop between strands 4 and 5 in position of a tetramer.

Initial oligomers on the way to fibril formation
That stefin B forms amyloid fibrils in vitro under relatively mild conditions was first shown in 2002 [58] and various solution conditions were probed, fibrils and protofibrils were imaged by AFM and TEM [59]. The fibril formation of stefin B was compared to that of stefin A, a much more stable homologue [60], to proline mutants [61], EPM1 mutants [62] and to chimeras between stefins A and B [63].
A model for the mechanism of stefin B amyloid fibril formation has been proposed based on temperature and concentration dependence of the kinetics [64]. Following changes during the lag phase of the reaction by SEC and ESI MS we were able to improve the model of amyloid fibril formation (Figure 4), further explaining the role of various off-pathway oligomers [41]. Dimers seem to be the building block from which fibril formation starts and the fibrils grow [56]. When the process was started from any kind of oligomers, these transformed into dimers [41]. NMR studies have shown that there are actually four types of dimers observed in the lag phase under amyloid forming conditions [55].

Oligomeric state related toxicity
Stefin B higher oligomers are equally toxic to cells as the molten globule state (mutant P74S Y31) [39]. Higher oligomers still have native like CD spectra and increased ANS binding, which would mean that they are properly folded. The toxicity seems to correlate with the ANS binding (ANS binding is indicator of exposed hydrophobic patches in the protein structure) [39]. MTS test has shown that isolated dimers and tetramers are not toxic ( Figure 5A), while higher oligomers show some toxicity to neuroblastoma cells (SH-SY5Y). In accordance, low ordered oligomers, monomers, dimers and tetramers do not increase caspase-3-like activity ( Figure 5B) [39]. All those results correlate with insertion into lipid membranes, where toxic species insert more effectively than not toxic. Monomers do not get internalized into the cytoplasm when added to the cell medium, while higher oligomers do [39].
Confirming the above results, in another study [65] it was demonstrated that a mutated P74S Y31 form of stefin B, which is molten globule at neutral pH and prefibrillar aggregates at pH 3, which are also in the molten globule conformation, are most toxic.
A growing body of evidence shows that the soluble oligomers formed during amyloid fibril formation exert toxicity and likely cause neurodegeneration [66]. Small soluble oligomers are the cause of synaptic disfunction, whereas large and insoluble aggregates are likely the reservoir of those toxic species [66]. In the case of Aβ peptide a significant correlation has been found between the levels of soluble oligomers and the degree of synaptic alteration, neurodegeneration and cognitive decline, while the correlation between insoluble deposits and those symptoms has not been demonstrated [67]. Non fibrilar dimers or oligomers of α-synuclein could play a major role in Parkinson's disease progression [68]. In accordance with a generic process of amyloid type of protein aggregation, not only oligomers of amyloidogenic proteins related to amyloid diseases exert toxicity, also HypF-N from E. coli, SH3 domain of PI3 kinase and some other proteins form oligomers that are toxic to fibroblasts and neurons, whereas amyloid fibrils of the same proteins show only low toxicity [69,70].
Conformation dependent antibodies distinguish between soluble oligomers [71][72][73] and amyloid fibrils [74,75]. Many of those antibodies recognize the generic epitopes regardless of the amino acid sequence and this suggests that different proteins form structurally similar oligomers and fibrils. It was shown that the antibody recognizing Aβ peptide fibrils, recognize also fibrils of transthyretin, islet amyloid polypeptide, β 2 -microglobulin and polyglutamine and that the same antibody does not recognize soluble oligomers of those proteins nor the native forms [75]. Antibodies recognizing soluble oligomers also inhibit their toxicity which suggests that soluble oligomers do not share only structural properties but also the mechanism of toxicity [72]. Figure 5. Viability of cells exposed to prefibrillar aggregates and oligomers of stefin B. (A) MTS test-viability of the SHSY5Y cells was measured after exposure for 16 hours to serum free medium, which contained prefibrillar oligomers of stefin B -StBpH3, StBpH5, StBpH3F (filtered), and separated dimers, tetramers and higher-order oligomers of stefin B at pH 7 -StBpH7; each at a final concentration of 44 µM. As a control for the effect of a non-toxic protein, cells were exposed for the corresponding length of time to 44 µM of soluble stefin A prepared at pH 5 (StApH5). All the results are relative to the cells alone, with no added substance or protein in the medium. (B) DEVD-ase activity after exposing SHSY5Y cells to prefibrillar aggregates. 200 µl of concentrated StBpH3 aggregates were added to 800 µL of cell medium to a final concentration of 44 µM (final pH was checked to be neutral). The cells were treated for four hours with 0.5 µM staurosporine (STS). Statistics-as described [39].
Both oligomers and amyloid fibrils are harmful for the cells; oligomers present mostly the toxic effect, while amyloid fibrils present steric obstacles.
• the disruption of the tissue architecture and functions promoted by the invasion of the extracellular space of organ by amyloid [76,77] • the destabilization of intracellular and extracellular membranes by oligomers [78,79] • the apoptotic cell death and receptor-mediated toxicity triggered by the oligomer interaction with various neuronal receptors [80] • the oligomer-mediated impairment of the P/Q-type calcium currents [81] • the impaired maturation of autophagosomes to lysosome mediated by the oligomer accumulation [82] • the dysfunction of autophagy, a lysosomal pathway for degrading organelles and proteins [83] • the oxidative damage-induced disruption of the cell viability promoted by the incorporation of redox metals into amyloid fibrils and subsequent generation of reactive oxygen species [84][85][86][87][88] • the general disorganization of cellular protein homeostasis associated with the exhaustion of the cell defence mechanisms, such as chaperone system [89,90] • proteasome inhibition [91] • the loss of crucial protein functions and/or gain of toxic functions

Oligomeric state related membrane interactions
Amyloid forming proteins interact with biological membranes and even form pores, which are similar to those formed by the pore-forming proteins [1,93]. Interactions between stefin B and membranes have been extensively studied [65,94,95]. Three different proteins have been studied: wt protein, Y31 isoform and G4R mutant (mutant involved in EPM1 disease). Also some comparisons have been made between native and prefibrillar states of the protein (prepared with incubation at pH 3.3 or 4.8) ( Figure 6) [94]. Prefibrillar oligomers may be organized in such a way that they are more amphipatic than the native protein and therefore acquire a higher surface-seeking potential. All forms of the protein insert into acidic lipid membranes, cause permeabilization of unilamelar vesicles and destabilize the membranes [94,95]. Nevertheless there are some significant differences between them concerning pore formation. The mutant G4R does not form pores, but breaks the membrane very fast (in a few minutes), while both the wt protein and Y31 isoform form pores; wt pores are cation selective (and may not be deleterious for the cell) and Y31 pores are anion selective [95]. Wt sample was separated into different oligomers apart from monomers, into dimers, tetramers and higher oligomers. Monomers, dimers and tetramers insert into lipid membranes with approximately the same critical pressure, while the higher oligomers (average size is around octamers) insert more effectively [65].
The exact mechanism of amyloid induced toxicity is yet controversial; nevertheless the pore formation into plasma or mitochondrial membrane is now the leading theory of pathogenesis. Aβ peptide needs anionic phospholipids for binding and insertion into membranes [96][97][98][99][100][101]. Monomeric form does not interact with membranes. Amylin (islet amyloid polypeptide) readily forms pores into planar lipid bilayer, preferring negatively charged lipids and furthermore the channel activity is inversely proportional to the amount of negative surface charge in the membrane [102]. Native form of prion protein does not form pores, while the mutant prion protein forms irreversible pores in the lipid bilayer. They claim that the pore formation is the main cause of the prion disease [103]. α-synuclein, a natively unfolded protein, also forms pores into lipid membranes but in contrast to others described it is not known how does the oligomeric form increase membrane permeability without forming discrete channels [104]. In the literature we can find contradictory results what effects amyloid proteins exert on membranes. For the same proteins we can find reports that they affect membranes by a nonchannel mechanism and other reports that they affect membranes by the formation of ion channels. Two mechanisms of pore formation into membranes by the amyloid forming proteins are summarized in Table 5. Specific ion transport through ion channel followed by destabilized ionic homeostasis Aβ40 peptide [105] Aβ40 peptide [106,107] Aβ42 peptide [105] Aβ42 peptide [108] α-synuclein [105] Aβ22-35 peptide [109] IAPP or amylin [105,110] α-synuclein [111] Polyglutamine [105] IAPP or amylin [112] Prion (106-126) H1 [105] stefin B [95] SOD1 [113]

Copper binding to stefin B and its inhibition of amyloid fibril formation
Divalent metal ions (Cu (II), Zn (II), Fe (II)) are often observed to colocalize with amyloid plaques in vivo in much higher concentrations than usually present in the normal environment. This has led to the hypothesis that this metal ions bind to mature fibrils and influence the fibril formation reaction [114].
Indeed it was proven that human stefin B is a copper binding protein [114]. It shows a picomolar affinity to copper at pH 7 and a nanomolar affinity at pH 5. Both the wild type protein and the Y31 isoform bind copper, while the Y31 P79S, which is tetrameric, does not. Monomers and dimers are able to bind copper, while other oligomers are not. Copper binding does not change the conformation of the protein; however it inhibits amyloid fibril formation ( Figure 8). Of note it does not prevent aggregation to prefibrillar oligomers, which can be even more toxic to cells. The protein binds two Cu 2+ ions. Copper binding promotes protein dimerization at neutral pH, while at acidic pH protein undergoes dimerization, already without added copper [114].
Later-on, NMR was used to detect the three possible copper binding sites of stefin B. The first is in the α-helix facing away from the β-sheets together with the loop between strands 4 and 5. The second potential binding site is the C-terminal together with the loop between strands 3 and 4 (this binding site is more likely in the dimeric form of the protein). The third binding site is in the dimer only and is the loop between strands 2 and 3, which is the stretched loop in the domain-swapped dimer together with the few residues at the N-terminal [55].
It was confirmed that copper affects the stefin B fibril formation. In more details, it slows down the elongation phase. It was also shown that the presence of copper destabilizes the protein structure and therefore it was concluded that it preferably binds to the slightly unfolded state of the protein. The final fibril morphology does not differ if copper ions are present or not [55].
Metal ions have different effect on amyloid forming proteins. They can speed up the fibril formation, reduce the lag phase, retard the fibrillization or even stop the process completely [115][116][117].
Binding of metal ions to prion protein increases proteolysis resistance and structural changes which might play an important role in the conversion process [118].
SOD1 aggregation is enhanced and modulated by Ca 2+ ions; at physiological pH Ca 2+ induces conformational changes that increase β-sheet content, they can also divert the aggregation from amyloid fibrils to amorphous aggregates [119]. Cu 2+ and Zn 2+ accelerate deposition of Aβ40 peptide and Aβ42 peptide, which results in the amorphous aggregates [120,121]. Fe 3+ induces the deposition of fibrillar amyloid plaques of Aβ40 peptide and Aβ42 peptide [120]. Prion protein can directly influence neuronal zinc concentrations [122]. In prion plaques there was found a significant disregulation of copper and manganese, copper was depleted and manganese was enriched [123]. Fe 3+ and Al 3+ enhance both formation of mixed oligomers and recruitment of α-synuclein in pre-formed tau oligomers [124]. Copper and selenium inhibit amylin fibril formation, while aluminium and manganese promote it [125]. Copper amylin complex has an anti-aggregating and anti-apoptotic properties, quenching the metal catalysed ROS [126]. Zinc also inhibits amylin fibril formation, furthermore, it favours the formation of amylin hexamers and inhibits the formation of dimers [127].

Interaction between oligomers of stefin B and Aβ peptide in vitro and in cells-"amateur chaperones"
"Professional chaperones" (heat shock protein family) prevent protein misfolding and subsequent protein aggregation. "Amateur chaperones" bind amyloidogenic proteins and may affect their aggregation process. Both types of chaperones colocalize with pathological lesions of Alzheimer's disease, may be involved in Aβ peptide conformational changes, clearance of Aβ peptide from the brain. Both types of chaperones may be involved in the aggregation, accumulation, persistence and clearance of Aβ peptide and may therefore serve as a potential targets for medical treatment of those patients [128].
Aβ peptides are produced by proteolytic cleavage of the amyloid precursor protein (APP) by α-secretase, β-secretase and γ-secretase. Cathepsin B, which can be inhibited by stefin B, is likely contributor to β-secretase activity [129]. Cysteine protease inhibitors reduce both Aβ peptide level in the brain and β-secretase activity in vivo [129]. Stefin B has an effect on production of Aβ peptides and furthermore it also inhibits Aβ peptide fibril formation [38].
Two isoforms of stefin B (Y31 and wt E31) have been studied. Y31 isoform is predominantly a dimer [31] while the wild type protein exist as mixture of monomers, dimer, tetramers and even higher oligomers [39]. ThT fluorescence and transmission electron microscopy (TEM) have shown that Y31 isoform completely inhibits Aβ peptide fibril formation ( Figure 9I). The direct interaction between those two proteins has also been shown by SPR measurements, where concentration dependent interaction has been reported and by ESI MS where the complex between dimer stB and monomer Aβ peptide has been detected ( Figure 9II) [38]. Furthermore, isolated oligomers of the wild type protein have been studied and it was shown that only the tetramer inhibits Aβ peptide fibril formation and that the higher oligomers show only a weak inhibition. Stefin B also colocalizes with the Aβ peptide aggregates in cells (shown by confocal microscopy) and with the C-terminal fragment of APP (comprising Aβ peptide sequence) (shown by immunoprecipitation).
Molecule/protein Type of interaction stefin B Inhibits Aβ peptide fibril formation, interaction in vitro and in cells [38].
ferulic acid Inhibits Aβ peptide oligomer formation from monomers and at the same time accelerates fibril formation from already formed oligomers [136]. S14G humanin Inhibits aggregation into fibrils and disaggregates preformed fibrils, reduces cytotoxicity effect of Aβ peptide [137].
albumin Increases the lag phase and decreases the total amount of fibrils [138] crocetin Inhibits Aβ peptide fibril formation, destabilizes preformed fibrils, stabilize Aβ peptide oligomers and prevents their conversion into fibril [139].  In order to show if binding between domain-swapped dimers and Aβ peptide is a generic property, other cystatins have been used. It was shown that only cystatin C dimers inhibit Aβ peptide fibril formation (30% inhibition) while stefin A dimers exhibited no such effect [140].
Several molecules interacting with Aβ peptide have been reported by now (Table 6).

Conclusions and perspectives
Stefin B has been so far found as a good model system for studying amyloid fibril formation. It exhibits nearly all features shared with other amyloid forming proteins: it forms mature fibrils under mildly acidic conditions or even at neutral pH at somewhat higher temperature, forms membrane pores and therefore promotes membrane leaking, binds copper ions and its' oligomers are toxic. It is not a model protein only, but also could be termed an "amateur chaperone" affecting Aβ peptide fibril formation both in vitro and in vivo [38].
We are trying to extend our in vitro knowledge to cell cultures to contribute even more to the understanding of conformational disease. It is hoped that new knowledge of protein oligo-merization and aggregation on the molecular and cellular levels will contribute to the development of new therapeutic strategies for patients with various conformational diseases, including the neurodegenerative ones.
pression is related to lymph-node metastasis. International journal of cancer. Journal international du cancer 1998; 79 (2) [27] Ii K., H. Ito, E. Kominami, A. Hirano, 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