Neurodegenerative diseases remain perplexing and problematic for modern research. Those associated with amyloidogenic proteins have often been lumped together simply because those proteins aggregate. However, research has identified a more logical reason to group some of these diseases together. The associated proteins not only aggregate, but also bind copper. The APP (amyloid precursor protein) binds copper in an N-terminal region. Binding of copper has been suggested to influence generation of β-amyloid from the protein. PrP (prion protein) binds copper, and this appears to be necessary for its normal function and might also reduce its probability of conversion into an infectious prion. α-Synuclein, a protein associated with Parkinson's disease, also binds copper, but, in this case, it potentially increases the rate at which the protein aggregates. The similarities between these proteins, in terms of metal binding, has allowed us to investigate them using similar approaches. In the present review, we discuss some of these approaches.
Two of the most common neurodegenerative disorders are Parkinson's disease and Alzheimer's disease. Different proteins have been associated with each disease. However, it is commonly recognized that the generation of Aβ (amyloid β-peptide), especially in the longer form of Aβ-(1–42) is the most likely candidate for the cause . This protein fragment is derived from APP (amyloid precursor protein) by secretase cleavage . Another protein originally associated with the non-amyloid component of plaques in Alzheimer's disease is now more commonly associated with Parkinson's disease . α-Synuclein is the main component of Lewy bodies in both Parkinson's disease and other diseases such as DLB (dementia with Lewy bodies). There has only been one protein associated with the TSEs (transmissible spongiform encephalopathies). PrP (prion protein) is now accepted as being essential for prion diseases or TSEs . In this case, the protein is transformed into an abnormal isoform rich in β-sheet structure. All three proteins bind metals.
APP has been shown to be a copper-binding protein . The copper binding is associated with histidine residues in the N-terminal domain. Binding of copper prevents dimerization of the protein , which potentially reduces the rate of cleavage by the enzyme BACE1 (β-site APP-cleaving enzyme 1), the β-secretase that plays a role in the generation of Aβ. Additionally, there is evidence that zinc can bind to APP . Once cleaved from the full-length protein, Aβ has also demonstrated copper- and zinc-binding properties that are believed to influence its aggregation . It is unclear whether the cleavage of APP is necessary for the metal-binding capacity of the Aβ domain.
PrP has been shown to bind a number of metals, including copper, manganese and zinc . Of these, copper has been shown repeatedly to be the metal bound to the normal isoform. A variety of functions have been suggested for this copper binding. These include copper uptake and/or sequestration  and possible antioxidant activity . Additionally, there is good evidence that the copper binding can inhibit protein conversion into the abnormal isoform  and alter the incubation period of the disease in animals . In contrast, manganese binding is associated with protein conversion into an abnormal isoform . Protein purified from the brains of animals and patients with a prion disease shows the presence of manganese. There are structural differences in PrP when it binds either copper or manganese , although it can bind both metals at the same time .
α-Synuclein is a more recent addition to this family, although there has been evidence for some time that copper influences its aggregation. However, more recent data suggest that it can potentially bind at two different sites in the protein . Parkinson's disease is more closely associated with changes in iron metabolism than copper. It has also recently been suggested that α-synuclein can bind iron . It is currently unclear whether either metal is associated with the protein in vivo in the absence of disease or which metal is a potential ‘appropriate’ binding partner. However, the conclusion that copper binding could initiate protein aggregation suggests that the interaction is aberrant.
Studies of the metal-binding affinities for proteins are necessary to verify that they would bind a metal in a physiological context. Many different techniques have been employed to establish these values, ranging from NMR to equilibrium dialysis. Our preference has been for ITC (isothermal titration calorimetry). This technique allows real-time measurements that can give values for both association and dissociation of a metal and stochiometries based on enthalpy changes associated with binding events. Our research has mostly focused on PrP, but current research is determining these values for both α-synuclein and APP.
Initial studies of metals binding to PrP suggested fairly low affinity values for copper binding [18,19]. Although the octameric repeat region was clearly responsible for binding four atoms of copper, there was also a suggestion that a fifth site located in the N-terminus could also bind copper. Later studies have shown that this fifth site is associated with two histidine residues (His95 and His110 in mouse) and could potentially represent a higher-affinity site . Values for the dissociation range from femtomolar down to micromolar. Studies of copper uptake by cells clearly show that PrP is involved in copper entry into cells with Km values in a nanomolar range. This clearly supports a fairly high affinity, possibly in the low nanomolar range . Our own studies of copper binding suggest that the highest-affinity site has a femtomolar dissociation constant, with the other sites being much lower . Studies of copper-depleted cells indicate that PrP cannot be purified without at least one copper ion bound, supporting further the notion of a single high-affinity site .
Studies of other metals binding to PrP have suggested possible binding of zinc, nickel and manganese . Although there is little evidence for zinc and nickel binding to PrP in vivo, manganese is associated with PrPSc (abnormal disease-specific conformation of PrP) isolated from the brains of patients with CJD (Creutzfeld–Jakob disease) or mice with scrapie [23,24]. Structural studies of PrP have shown that both copper and manganese can alter the structure of the protein . Binding of manganese to PrP results in a change in conformation to a form that aggregates, is neurotoxic  and has increase protease resistance, similar to PrPSc . Studies of affinity of PrP for manganese show that the affinity is much lower than that of copper. However, this result is expected, as all manganese-binding proteins have low affinities. The values found for PrP are equivalent to those of other known manganese-binding proteins .
Copper and zinc binding to APP has been less well studied. The copper-binding domain is found in the N-terminus of APP within amino acid residues 135–175 . The copper-binding domain contains the typical His-Xaa-His motif of a type II site similar to that seen for superoxide dismutase and lysyl oxidase copper homoeostasis protein groups. APP reduces the bound copper to Cu(I), suggesting that the protein could have a copper reductase activity . This reduction leads to oxidation of Cys144 and Cys158, resulting in the formation of a new disulfide bridge . The dissociation constant for copper binding to APP has been suggested to be approx. 10 nM . Binding of the copper to this domain involves the imidazole rings of three histidine residues, supporting the notion that this site has higher specificity for copper then other type II sites .
The metallochemistry of Aβ has been investigated in some detail . Aβ can be rapidly precipitated by Zn2+ ions at low physiological concentrations, and it has been reported that other metal ions such as Cu2+ and Fe3+, unlike Zn2+, produced a greater aggregation of Aβ under weakly acidic conditions . Such a mildly acidic environment probably resembles conditions occurring in the brain. The significance of these in vitro studies with Aβ and metal ions is emphasized by other data showing that the homoeostasis of zinc, copper and iron are significantly altered in Alzheimer's disease brain . The affinities of the Zn2+-binding sites on Aβ-(1–40) were measured as 100 nM and 5 μM, indicating that they may be occupied under physiological conditions . The highest affinity Cu2+-binding site on Aβ-(1–42) has a measured association constant (Ka) of 10–15 aM . With such strong affinity for Cu2+, Aβ species such as Aβ-(1–42) are likely to bind Cu2+in vivo. The increase in Cu2+ affinity of Aβ-(1–42) over the normal APP is related to APP proteolysis. APP is a membrane-spanning protein, and the Cu2+-binding site is probably hidden within the protein, becoming exposed in the proteolysed Aβ fragment. Also, Aβ-(1–42) has a higher β-sheet content, and these structures are frequently found in the tertiary structure of Cu2+ catalytic sites.
Studies of metal affinity for α-synuclein are relatively new. Initial studies of copper binding to α-synuclein have their origin in the study of its aggregation . Copper-mediated aggregation seemed to be related to interaction of copper with the C-terminus of the protein. This was shown by limited proteolysis of α-synuclein that cleaved off part of the C-terminus either at residue 97 or 114. The shorter fragment showed no response to copper in terms of oligomerization, whereas the 1–114 fragment did produce a limited amount of oligomerization . Despite early studies suggesting that copper causes aggregation of the protein via the C-terminus, a more recent study suggested that copper can bind to His50. Further binding can occur at the C-terminus, but this is of much lower affinity than the N-terminal site . Binding at the high-affinity site was shown to be sufficient to drive oligomerization of the protein. The high-affinity site appears to be a type-II site with square planar co-ordination. However, the affinity for this site is suggested to be 0.1 μM. That for the second site was shown to be 50 μM, which is in line with previous findings . Further studies suggest involvement of the N-terminus in binding at least one copper ion . Metal binding to α-synuclein has been studied using ITC. The results suggest that the protein binds two atoms of copper, but with high-micromolar affinity . The same study also suggested that α-synuclein can bind iron with a similar low affinity. Further study is clearly needed to verify that α-synuclein binds these metals, but our own studies suggest that they do and with affinities that would be more likely to suggest metal binding in vivo.
As metal-binding proteins, APP, Aβ, PrP and α-synuclein are linked to redox-active metals that could potentially have damaging effects on cells or other proteins. Therefore studying the redox chemistry of the metal–protein complex could give valuable insight into potential disease mechanisms induced by these proteins as a result of metal binding.
PrP has been suggested to be an antioxidant . Although other researchers have suggested that this is the case, their evidence has been based on experiments using high-affinity chelators that are likely to strip the protein of any bound copper [35,36]. Strong new evidence supports the notion that PrP can act as an antioxidant. Studies with cyclic voltammetry show that, when copper is bound to the protein, it can fully cycle between reductive and oxidative potentials (Figure 1) and has a midpoint potential close to 0 V . This clearly indicated that the electrochemical properties of PrP are equivalent to those of other known antioxidants. In contrast, manganese bound to PrP is predominantly oxidized, which suggests a non-cyclable behaviour .
Cyclic voltammograms recorded at a scan rate of 10mV ·s−1 for the reduction and reoxidation of copper refolded mPrP adsorbed on to a 3 mm diameter boron-doped diamond disc electrode and immersed in 5 mM Mes/Tris buffer solution at pH 7
APP can reduce Cu(II) to Cu(I) in a cell-free system, potentially leading to increased oxidative stress in neurons . The domain that contributes to such activities is the copper-binding domain  residing between residues 135 and 158 of APP, a region that shows strong homology with APLP (APP-like protein) 2, but not to APLP1. Potentially, APP–Cu(I) complexes are involved that reduce hydrogen peroxide to form an APP–Cu(II)–hydroxyl radical intermediate . APP residues 135–158 consisting of cysteine and copper-co-ordinating histidine residues can modulate copper-mediated lipid peroxidation and neurotoxicity in culture of APP-knockout (APP0/0) and wild-type neurons . Wild-type neurons were found to be more susceptible than APP0/0 neurons to physiological concentrations of copper, but not other metals.
A recent study has examined the electrochemistry of the Aβ–copper complex . It reveals that Cu(II) co-ordinates with Aβ in a 1:1 ratio. Independently of the methionine residue, the oxidation state of the copper centre in the complex is 2+. The data suggest that the presumed reduction reaction off the Cu(II) centre to Cu(I) cannot occur in vitro . The midpoint potential for the reduction of the copper centre was determined to be 0.08 V (compared with Ag/AgCl) . This group's result suggests that Met35 can be oxidized by the Cu(II) centre. The authors suggest that the copper–Aβ complex could interfere with electron transport of the mitochondria, thus altering cell viability .
Binding of metals to α-synuclein is a relatively new discovery. Therefore there has been no significant study of the electrochemistry resulting from this interaction. It has been established that copper can cause the oxidation of the protein, and this is associated with increased aggregation , but further work is still needed.
PrP, synuclein and Aβ are influenced by metals when it comes to their aggregation (Figure 2). Polymerization studies usually rely on recombinant protein or peptides, and the products of the reactions are analysed with fluorimetric assay such the thioflavin T assay, and other techniques such as electron microscopy, atomic force microscopy or various kinds of spectroscopy. In some cases, the presence of metals accelerates the aggregation process, as is known for α-synuclein . In the case of prion disease, copper has been shown to inhibit the potential of the protein to aggregate , although the picture is confusing as, in some cases, non-specific effects will accelerate aggregation . Binding of manganese to PrP generates a form of protein that can act as a seed and catalyse aggregation . However, manganese itself has little effect on this form of aggregation.
Schematic representation of protein aggregation
Of the three diseases discussed in the present article, only prion disease can be associated with an ‘infection’. This usually means experimental transmission of prion disease either from one animal to another or in cell culture. Just as metals play a role in the potential of the protein to aggregate, they also influence formation of the infectious agent. Copper increases prion infectivity in animal models , and chelators can alter the incubation period . Understanding how metals influence this process has been poorly explored and remains an interesting prospect for future study.
Expression of PrP is necessary for a mammal to be susceptible to prion disease infection . Increased PrP expression increases susceptibility . Aggregation of α-synuclein is associated with increased expression , whereas lowered expression of β-synuclein is also associated with Parkinson's disease and DLB . These findings suggest that the relative expression levels of the two proteins may be key to why α-synuclein aggregated. Formation of Aβ requires cleavage from APP by secretases. Copper is believed to influence this process, but there is also evidence that BACE1, the β-secretase, is a copper-dependent enzyme . The genes for these proteins have a common feature in that they all have MREs (metal-response elements) present in their promoters. In some cases, there is already evidence that copper or another metal can regulate their expression . However, significantly more information is needed in order to know how these genes are regulated. In the case of the PrP, there is evidence that intronic regions play a strong role in gene regulation . Further information along these lines could help to develop strategies that could result in treatments that decrease expression of amyloidogenic proteins which might switch off disease progress. Switching off the expression of PrP has already been shown to arrest prion disease .
Metal Metabolism: Transport, Development and Neurodegeneration: A Biochemical Society Focused Meeting held at Imperial College London, U.K., 9–10 July 2008. Organized and Edited by David Allsop (Lancaster, U.K.) and Harry McArdle (Rowett Research Institute, Aberdeen, U.K.).