Copper binding to the amyloid-beta (A (cid:1) ) peptide associated with Alzheimer’s disease. Folding, coordination geometry, pH dependence, stoichiometry and affinity of A (cid:1) (1-28); insights from a range of complementary spectroscopic techniques.

There is now direct evidence that copper is bound to amyloid (cid:4)(cid:1)(cid:3) peptide (A (cid:1) ) in senile plaque of Alzheimer’s disease (AD). Copper is also linked with the neurotoxicity of A (cid:1) and free radical damage, and Cu 2+ chelators represent a possible therapy for Alzheimer’s disease. We have therefore used a range of complementary spectroscopies to characterize the coordination of Cu 2+ to A (cid:1) in solution. The mode of copper binding is highly pH dependent. EPR spectroscopy indicates that both coppers have axial, Type II coordination geometry, square-planar or square-pyramidal, with nitrogen and oxygen ligands. Circular dichroism studies indicate that copper chelation causes a structural transition of A (cid:1) . Competition studies with glycine and L-histidine indicate that copper binds to A (cid:1)(cid:5)(cid:6)(cid:4)(cid:7)(cid:8)(cid:9) at pH 7.4 with an affinity of K a ~10 7 M -1 . 1 H NMR indicates that histidine residues are involved in Cu 2+ coordination but Tyr10 is not. Studies using analogues of A (cid:1) (1-28) in which each of the histidine residues have been replaced by alanine or in which the N-terminus is acetylated suggest that the N-terminus and His13 are crucial for Cu 2+ binding and that His6 and His14 are also implicated. Evidence for the link between AD and Cu 2+ is growing, and our studies have made a significant contribution to understanding the mode of Cu 2+ binding to A (cid:1) in solution.


INTRODUCTION
Alzheimer's disease (AD) is characterized by innumerable deposits of extracellular amyloid plaques. A small peptide, amyloid-b peptide (Ab), plays a critical role in the initial build up of these amyloid plaques and is the main constituent of the amyloid deposits (1,2). In addition genetic alterations underlying familial AD are associated with an increase in the production and/or the deposition of Ab in the brain (3-6). Amyloid-b peptide can be between 39 and 43 residues in length, of which Ab  and Ab     and Ab  are found in the cerebrospinal fluid (CSF) and blood plasma of all humans where Ab 40 has a concentration of 5 nM in CSF (7,8). It is yet to be established what triggers Ab to convert from its soluble form to an amyloidogenic form, but it has been shown that physiological levels of Cu 2+ and Zn 2+ cause marked aggregation of Ab. This process is thought to be the prelude to amyloid formation (9). Levels of these metals are elevated in amyloid plaque deposits; 0.4 mM and 1 mM for Cu 2+ and Zn 2+ respectively (10).
Cu 2+ induced aggregation of Ab occurs as the pH is lowered to 6.8. This mildly acidic environment mimics a feature of inflammation found in AD (11). Studies on cerebrospinal fluid indicate that zinc will cause the selective aggregation of endogenous soluble Ab peptide (12). Metal chelators specific to Cu 2+ and Zn 2+ will reverse this aggregation process (13,14).
In addition, the neurotoxicity of Ab is linked to metal induced oxidative damage and is a feature of the pathogenesis of AD (15). A dual role as a pro-and anti-oxidant in copper redox cycling in a Fenton-type reaction has been proposed for Ab (16)(17)(18)(19)(20)(21)(22).
There are two commonly expressed objections to the role of copper in AD. The first objection is that AD is not associated with elevated exposure to environmental copper. It is important to clarify this misconception. The total concentration of copper within the brain is potentially more than sufficient to be neurotoxic. As a consequence the brain has efficient homeostatic mechanisms in place to maintain compartmentalization of metal ions, which when compromised cause neurodegenerative diseases such as Wilson's and Menkes disease. There is evidence to suggest that homeostatic mechanisms for metal ions are impaired in AD patients and AD is characterized by altered metal ion-dependent processes and metal ion concentrations in the brain (23). The second objection is that the affinity of Ab for Cu 2+ is too low to bind these metals at their extracellular concentrations. This is also a misconception since extracellular levels of Cu 2+ may reach as high as 15 mM (10), while we show here that Ab affinity for Cu 2+ is at the sub-micromolar level and is reported to be much higher in amyloid plaques (24).
A recent study using Raman spectroscopy has provided direct evidence that copper and zinc are bound via the histidine imidazole rings in isolated senile plaque cores (25). Perhaps one of the most significant pieces of evidence to link copper with AD is the observation that normally insoluble amyloid deposits of post-mortem brain tissue from AD patients can be solubilized in aqueous media by the presence of metal chelators specific to Cu 2+ (14). It has been shown that the use of copper chelators can markedly inhibit amyloid accumulation in AD transgenic mice and are in trials as potential drug therapies for AD (26,27). A recent study has shown that trace amounts of copper in the drinking water of rabbits induces betaamyloid plaque accumulation (28,29).
Despite an increasing body of evidence to link Cu 2+ with AD, the precise coordination geometry and the residues involved in Cu 2+ ligation are yet to be established. Both monomeric and dimeric species have been proposed (30). In addition, there are disagreements as to the affinity and stoichiometry of binding. For example, both attomolar affinities (24) and micro-molar affinities (31) for Cu-Ab(1-42) have been reported. In this study, we use a range of complementary spectroscopies to characterize the binding of Cu 2+ to Ab and the structural changes induced in Ab upon copper coordination. To facilitate solution spectroscopy methods we have used the more soluble fragment of Ab(1-28) which lacks the C-terminal third of the molecule. Residues 29-42 are highly hydrophobic and are not believed to be associated with direct coordination of the metal ion (30,41). In addition, we have studied a number of analogues of Ab  in which each of the three histidine residues have been replaced with an alanine.

EXPERIMENTAL
Peptide Synthesis and Purification:-Peptides representing various fragments of the amyloid-b peptide were synthesized by employing solid phase F-moc chemistry and produced by the ABC facility at Imperial College, London. After removal from the resin and de-protection, the samples were purified using reverse phase HPLC and characterized using mass spectrometry and 1 H NMR.
Titrations:-The pH was measured before and after each spectrum was recorded. N-Ethylmorpholine (EM) buffer was found not to interfere with Cu 2+ binding. Typically, 50 mM EM buffer was used for EPR studies while for 1 H NMR and CD studies, samples were prepared in ultra high quality (>18 W.cm resistivity) water and the pH was adjusted using small amounts of 0.1 M NaOH or HCl. The peptide concentrations were determined using the extinction coefficient of 1280 M -1 cm -1 (due to the single tyrosine residue) (32). Typically, the freeze-dried peptides contained 5 to 10 % moisture by weight. The addition of metal ions or competing ligands to the Ab peptides was performed using small aliquots from stock aqueous solutions.

Circular Dichroism (CD):-CD spectra were recorded on an AVIV Circular Dichroism
Model 202 spectrometer at 25° C. Typically a cell with a 0.1 cm path length was used for spectra recorded between 185 and 260 nm with sampling points every 0.5 nm. A 1 cm cell path-length was used for data between 240 and 800 nm with a 2 nm sampling interval. A minimum of three scans was recorded and baseline spectra were subtracted from each spectrum. AVIV software was used to smooth data when necessary. Data was processed using Kaleidagraph spreadsheet/graph package. Direct CD measurements (q, in mdeg) were converted to molar ellipticity, De (M -1 cm -1 ) using the relationship De = q /33,000 * c * l where c is the concentration and l the path length.
Absorption Spectroscopy (UV/Vis ):-UV/visible electronic absorption spectra were obtained with a Hitachi U-3010 double beam spectrophotometer, using a 1 cm path length quartz cuvette.
Fluorescence Spectroscopy:-Fluorescence spectra were collected using a Hitachi F-2500 fluorescence spectrophotometer. An excitation frequency of 280 nm was used and data collected over the range of 290 -400 nm. Samples were placed in a four-sided quartz fluorescence cuvette (Hellma) and data recorded at room temperature.
EPR:-X-band Electron Paramagnetic Resonance (EPR) data was recorded using a Bruker ELEXSYS 500 spectrometer, operating at a microwave frequency of ~ 9.3 GHz. All spectra were recorded using a microwave power of 0.63 mW across a sweep width of 2000 Gauss (centred at 3200 Gauss) with a modulation amplitude of 10 Gauss. Samples were frozen in quartz tubes and experiments carried out at between 20 and 120 K using a liquid Helium cryostat. A minimum of two scans were recorded per spectrum. All EPR spectra shown have been background subtracted from a water blank with subsequent baseline correction in the XEPR software package using 3 rd or 4 th order polynomial splines. In order to analyse EPR data using the method described by Peisach and Blumberg (34), it is necessary to convert values from Gauss (G) to milli-Kaisers (mK) by using the formula A II (mK) = 0.046686*g*DH where g = 2.0023 and DH is the A II splitting measured in Gauss.
NMR:-Proton NMR data was collected using a Varian UNITYplus-600 MHz 1 H frequency spectrometer. Data was processed using the VNMR software package. All spectra were recorded at 25°C in 100% D 2 O solution, typically at a peptide concentration of 1mM. Proton peak assignments were made by the analysis of two-dimensional TOCSY and ROESY spectra of the apo peptides. TOCSY and ROESY spectra were collected with typical mixing times of ~ 75 and ~ 300 ms respectively.  Figure 1 shows the EPR spectra of 0. 8  We have carried out complementary studies using CD spectroscopy. Figure 2 shows visible CD spectra for Ab(1-28) with one equivalent Cu 2+ at various pH values between 4.5 and 10.5. At pH 7.5 a visible absorption band is observed at 600 nm (e 600nm = ~50 M -1 cm -1 ) which is typical of a type II Cu 2+ d-d transition. At pH 7.5 and below, the associated CD band is extremely weak and is not detected, however the accompanying CD band at 312 nm assigned as an amide-to-copper charge transfer band is observed at pH 5.5 and above (35). Only as the pH is raised to above pH 8.5 is a negative CD band at 514 nm observed, arising from Cu 2+ d-d transitions. Inserts in Figure 2 show the change in the intensity of CD bands in the presence of 1 equivalent of Cu 2+ at 252 nm, 312 nm and 514 nm.

pH dependence of Cu 2+ binding:-
Relatively strong CD bands are often observed for d-d transitions of Cu 2+ tetragonal complexes (36,37). However these complexes involve main-chain amide coordination as well as histidine coordination via the imidazole ring. In these cases, the dominant contribution to optical activity observed is due to the vicinal contributions resulting when the asymmetric alpha carbon is held in a chelate ring between two chelating donor atoms, for example adjacent main-chain amides (38). At physiological pH and below, the lack of optical activity from the d-d transition of the Cu-Ab complex suggests that backbone amide coordination is not taking place. It is likely that raising the pH above 8 promotes amide deprotonation and copper coordination by the main chain, resulting in a CD band being observed at 514 nm. In summary, it is clear that Ab forms a Type II square-planar coordination geometry with Cu 2+ , and both EPR and CD measurements indicate that the coordinating ligands are highly pH dependent, a mixed species is present at physiological pH, and main-chain amide coordination is not present at lower pH values. sample, confirms that all the EPR signals for the Cu-Ab complex are observed and therefore rules out the possibility of EPR silent spin-coupled Cu 2+ . Addition of a further mole equivalent of copper (and up to five mole equivalents) results in no further increase in the intensity of the EPR spectra. Cu 2+ ions in water will give a largely EPR silent signal at pH > 7 (39). We have confirmed this observation by obtaining EPR spectra of CuCl 2 in EM buffer at pH 7.8; copper EPR signals are drastically attenuated relative to Cu 2+ bound to Ab. Spin integration of the EPR spectra has been plotted versus copper addition as shown as an insert in Figure 3. It is clear from the EPR data that Ab(1-28) binds two Cu 2+ ions sequentially.

Stoichiometry of Cu
After two mole equivalents of Cu 2+ , Ab(1-28) becomes saturated with copper at physiological pH. Similar studies have been performed using CD, both in the UV and visible regions.
Monitoring of the band at 205 nm reveals saturation of Ab(1-16) and Ab(1-28) at two mole equivalent of Cu 2+ at pH 9. In summary both EPR and CD derived Cu 2+ binding curves indicate Cu 2+ saturation of binding sites on Ab(1-28) after two mole equivalents.

Folding of the backbone in the presence of Cu 2+ :-It is believed that amyloid formation in
AD and other amyloidogenic diseases such as prion disease are the result of protein or peptide misfolding. The Ab peptide has a structural transition associated with amyloid formation, with conversion from random coil to an extended beta sheet like conformation (40). CD spectra in the UV region can be used to monitor changes in the main-chain conformation. The additional 12 residues to the C-terminus have no effect on the copper-induced conformational transition. The changes in the spectra with Cu 2+ addition are complete by two mole equivalents of Cu 2+ , which agrees with the stoichiometries determined by EPR ( Figure   3). Isodichroic points are observed at 217 nm and 195 nm, and these are maintained between 0 and 1.0 equivalent of Cu 2+ . If a dimeric species were formed, the isodichroic point would be expected to shift at 0.5 equivalent Cu 2+ .
The chirality at 217 nm is largely invariant with the addition of copper between pH values of 5.5 -9.5. An increased negative CD signal at 217 nm is often attributed to b-sheet or extended conformation. The copper induced structuring of Ab can therefore not be directly attributed to an increase in b-sheet or extended structure. The appearance of a positive CD contribution at 205 nm is not characteristic of any defined secondary structure but does indicate increased ordering of the main-chain.
Affinity of copper binding:-Key to the physiological significance of Cu 2+ binding to Ab is its affinity. With this in mind, we have used the competitive effects of glycine and Lhistidine to measure Cu 2+ affinity for Ab by fluorescence spectroscopy. Figure 5a shows that the addition of Cu 2+ to Ab(1-28) causes marked quenching of the tyrosine fluorescence signal at 307 nm. As glycine is added, it competes with Ab for the Cu 2+ and the tyrosine fluorescence signal reappears, as shown in Figure 5b. Cu 2+ coordinates to glycine via the amino and carboxylate groups with an apparent (pH adjusted) of K a = 1.8 *10 6 M -1 , and two glycine residues will bind to a single Cu 2+ ion (33). It takes more than 100 mole equivalents of glycine to cause the tyrosine fluorescence signal to completely return to its maximal strength. Half the maximal quenching is achieved at ~18 (± 2) equivalents of glycine. Thus the affinity of Cu 2+ for Ab  is at least an order of magnitude higher than that of glycine, putting the dissociation constant K d in the sub mM range (K d << 0.5 mM). Similar experiments have been carried out using L-histidine as the competing ligand. In this case, tyrosine fluorescence returns to its maximal value with only 2.5 mol equivalents of Lhistidine, as shown in Figure 5c. Two molecules of histidine will bind a single Cu 2+ ion using the amino and imidazole nitrogens as ligands with an apparent K d at pH 7.8 of 1.5 nM. This indicates that copper will bind to Ab(1-28) with a lower affinity than L-histidine. This puts the affinity of Cu 2+ for Ab greater than 1.8*10 6 M -1 but less than 6.7*10 8 M -1 or a K d of <<500 nM but >1.5 nM, i.e. 10-100 nM.
We have obtained similar Cu 2+ affinities for Ab(1-28) using CD spectroscopy. CD was used to directly measure copper-Ab associated absorption band at ~ 314 nm. Using the competitive effects of glycine, copper absorption bands become CD silent when bound to non-chiral glycine (37). We find using the CD band at ~ 314 nm that similar amounts of glycine are In contrast, the CD spectra of Ab(Ac10-16) and Ab(1-11) are very different from Ab (1-16) and Ab , as shown in Figure 6. In particular, at pH 7.5, Ab(1-16) and Ab(1-28) show almost no CD bands associated with d-d transitions above 500 nm while Ab(1-11) has a negative CD band at 590 nm and Ab(Ac10-16) has a strong positive CD band at 560 nm. The CD spectra of Ab(Ac10-16) suggests that the N-terminus and/or His6 are involved in copper coordination. In addition, Ab(1-11) CD spectra indicate that His13 and/or His14 are key residues in coordinating the Cu 2+ ion. A difference in the EPR spectra between Cu-Ab(Ac10-16) and copper complexes of the larger fragments are observed but are less pronounced, as Ab(Ac10-16) still gives rise to an axial Type II Cu 2+ spectrum (data not shown). The pH dependence of binding is similar for Ab(1-16) and Ab(1-28) although EPR data (see insert of An additional peptide has also been synthesized in which the N-terminus is blocked by acetylation. It is believed that nitrogens from these four loci are the most likely candidates for copper coordination. The EPR spectra of all four analogues studied indicate axial Cu 2+ coordination containing N and O ligands. At pH 6 differences in the spectra between all the analogues and Ab  are slight, at pH 9 the Cu-EPR spectra of the analogues are largely unchanged. In contrast, as shown in Figure 1 there is a strong pH dependence to the Cu-EPR spectra of wild-type Ab(1-28). The EPR data suggest that the complex formed at the higher pH requires the presence of all the His residues and the N-terminal amino group. However we have used a number of further spectroscopic techniques described below to support this assertion. For example, when coordination geometry and ligands are similar, visible CD can potentially be much more sensitive to the coordination sphere around the Cu 2+ ion compared to EPR.
Visible CD spectra are very sensitive to the relative position of coordinating ligands with changes in the intensity and sign of CD bands (38). The visible CD spectra of the wild-type peptide and CD spectra for the four Ab(1-28) mutants are shown in Figure 7. Spectra have been obtained at pH 5.5, 7.5 and 9.5 for all four analogues as well as wild-type Ab(1-28). In addition, spectra have been obtained with both one and three mole equivalents of Cu 2+ present. CD bands associated with the Cu 2+ ion are observed at key positions of 514 nm, (assigned to Cu d-d transitions), 360 nm (assigned to N im > Cu 2+ charge transfer), 312 nm (assigned to N amide to Cu 2+ charge transfer) and 280 nm (assigned to N im p2> Cu 2+ charge transfer) (35). Taking each analogue in turn with one equiv Cu(II) at pH 9.5, the N-terminally blocked Ab(Ac1-28) gives a very different trace to wild-type Ab  implying the N-terminus is essential for Cu(II) binding. In particular, only Ab(Ac1-28) gives a CD band at 360 nm and the band at 514 nm increases by ~ 35 % relative to the wild-type spectrum.
Furthermore, only Ab(Ac1-28) gives a positive chirality above 580 nm. The spectra of Ab(H13A) are also perturbed relative to the wild-type spectra. In particular there is a ~ 40 % drop in signal at 514 nm but with an increase in the positive charge transfer band below 280 nm. The copper loaded spectrum of Ab(H14A) has some differences relative to the wildtype, with small reduction in intensity (~ 17%) of the band at 514 nm. It is not clear if the small drop in intensity for the band at 514 nm is significant, or due to reduced solubility or small variation in pH. Similarly, the differences between Ab(H6A) and the wild-type spectrum are slight, with a very small reduction (~ 4 %) in the band at 514 nm. Taking these observations together, it appears that at one equivalent of Cu 2+ , His13 and the N-terminus certainly coordinate the copper ion. Cu 2+ coordination to His14 and His6 is less apparent from the visible CD data but cannot be ruled out. Indeed it is apparent from the CD data in the UV region as shown in Figure 8, and described below, that His14 and His6 are involved in coordination.
Copper titrations on each of the analogues were also studied using CD in the UV region and compared to the wild-type spectra. Figure 8 shows the apo subtracted difference CD spectra of each of the 4 analogues and the wild-type spectra after 1 equivalent of Cu 2+ was added. Cu 2+ will broaden 1 H NMR signals when in close proximity to (< 7Å) or directly coordinated to the metal ion. Figure 9 clearly shows profound broadening of the He1 and Hd2 resonances for all three histidine residues of Ab(1-28) at pH 7.8 relative to the other aromatic side chains.
In particular Tyr10 which has been suggested as a potential ligand to Cu 2+ (30,41)  are also implicated, but not Tyr10.
Bridging Cu-His-Cu dimer:-It has previously been suggested that there is histidine bridging between two Cu 2+ ions in the Ab(1-28) Cu 2+ complex (30). This work was based on the observation that the EPR spectra broaden with Cu 2+ addition (at pH 6.8) and assumed this was due to electron coupling via a bridging histidine as is observed in EPR spectra of Cu 2 -SOD (42). However, we have not observed broadening of EPR spectra with increasing copper addition. EPR spectra in which electron coupling is taking place are very temperature dependent, and no such temperature dependence is observed in these spectra. Spectra have been obtained for Ab(1-28) with 0.8 mole equivalent Cu 2+ over a range of temperatures between 20 K and 110 K, but line shapes are not effect by this change in temperature. We suggest that the broadening observed in this previous study (30) were simply due to throughspace dipole coupling at high copper ion concentration rather than scalar through-bond coupling in a bridged complex for Cu-Ab-Cu.
Model Complex:-A crystal structure of a model compound containing a di-peptide analogue of two adjacent histidine residues has been reported (43). In this compound, copper chelates to the eN of both imidazole rings making a cis arrangement rather than trans coordination (90 o rather than 180 o ) to the copper ion in a square-planar arrangement. Sterically, this type of ligand arrangement could facilitate the coordination of the N-terminus and His6 without any obvious restrictions. Interestingly, coordination to copper from adjacent histidine imidazole rings makes it impossible for main-chain coordination, via the amide or carboxyl, for either Val12, His13, His14 or Gln15. The absence of main-chain coordination is supported by the CD spectra at neutral pH, which shows little chiral optical activity for d-d transitions. The lack of appreciable d-d transition CD bands suggest minimal vicinal effects (38), and this implies there is no main-chain co-ordination at physiological pH and below, as previous discussed. A study of Ab using Raman spectroscopy indicates Cu-eN at mildly acidic conditions and dN coordination at physiological pH and above (41). Imidazole coordination via the eN makes main-chain coordination less likely than the more stable sixmembered chelate ring between the dN and the His main-chain amide (44). This transition between eN and dN coordination with pH is supported by our CD studies that indicate mainchain coordination at higher pH values only. Figure 10 represents the possible ligands coordinating the Cu 2+ ion in a square-planar arrangement including the N-terminal amino group, and eN groups of His13 and His14. The fourth ligand is from His6 rather than Tyr10.

DISCUSSION
With increasing interest in the role of Cu 2+ in Alzheimer's disease (AD) (24,25,28,29,45) we have used a range of complementary spectroscopic techniques including CD, EPR, NMR and fluorescence to study Cu 2+ binding to the Ab peptide. In summary, Cu 2+ binds to the Nterminus of Ab and the three histidine imidazole rings but not Tyr10, as shown in Figure 10.
Ab is essentially a random coil peptide in aqueous medium, while the amyloid fibrillar form of the peptide is rich in b-sheet (40). We were therefore interested in the possibility that Cu We have shown that His13 is a critical residue in coordinating the Cu 2+ ion. Interestingly, rat-Ab lacks a histidine at position 13 and copper induced aggregation of rat-Ab is not observed (11). Furthermore, wild-type rats do not exhibit AD-like pathology, although transgenic rats containing human Ab can present with AD-like pathology. It has been shown that copper coordination has a profound influence on the solubility of Ab (11). The fourteen N-terminal amino acids of Ab contain four carboxylic acid groups and four basic groups, the three histidine residues and the N-terminal amino group. Cu 2+ coordination may cause the loss of all four positive charges. Thus copper binding will have a profound influence on the electrostatic surface of the Ab peptide with a reduction in charged groups on Ab and a change from an overall neutral charge to a highly negatively charged N-terminal tail. It is likely that this change in charged residues is the cause of the aggregation observed upon Cu 2+ binding.
We have shown that coordination of Cu 2+ to Ab is highly pH dependent. It is clear that there is a transition between two coordination geometries as the pH is raised. In the case of Ab(1- Ab  and Ab , with a high-affinity site at 0.5 mole equivalent of Cu 2+ (24). With no change in the appearance of spectra between 0.5 mole equivalent and 1 equivalent of Cu 2+ , both our EPR and CD data indicate that Ab(1-28) does not form a dimeric Cu(Ab) 2 species under these conditions. In addition, from our EPR data we observe no evidence of Ab  using bridged histidine coordination to form a dimeric species as previously suggested (30).
The very different affinities of Cu 2+ for Ab described by Atwood et al. (24) could reflect the differences in the complexes formed. We suggest that the Cu 2+ binding affinities described   in solution. Dimerization is a concentration dependent phenomenon. The strong tendency for Ab  to form aggregates may facilitate crosslinking and a Cu(Ab) 2 species. It is likely that soluble Ab(1-42) initially forms a one-to-one complex as is seen for Ab . The hydrophobic C-terminal residues 29-42 do not directly coordinate to copper. Indeed, apart from Met35, which has been ruled out by a study using Raman spectroscopy (41) in which a S-Cu band was not observed, residues 29-42 contain only aliphatic side-chains, which do not possess lone-pairs to facilitate Cu 2+ ion coordination.
Rather the C-terminal residues may promote aggregation that facilitates a cross-linked dimeric copper complex of Ab.
There have been relatively few structural studies of the Cu-Ab complex. The paramagnetic copper centre has hampered the determination of an NMR solution structure. Studies using Raman spectroscopy have suggested that Cu 2+ binds to the Nd of the histidine imidazole and will coordinate via the Ne nitrogen under mildly acidic conditions (41,47). Our CD studies support this pH dependent transition. In agreement with our CD and NMR data and glycine competition studies, the Raman spectroscopic study suggests that all three histidines coordinate to copper. However, in contrast with our observations, Tyr10 but not the Nterminus was implicated in coordination (41). In agreement with our work, potentiometric studies of Ab fragments (46) do implicate the N-terminus in binding Cu(II).
Recent studies now present strong evidence to link Cu 2+ with AD, for example, Raman microscopy has provided direct evidence that copper and zinc are bound via the histidine imidazole rings of Ab in isolated senile plaque cores (25). In addition, copper chelators can solubilize amyloid plaques and represent possible therapies for Alzheimer's disease, as copper chelators can inhibit amyloid accumulation in transgenic mice (26). Copper is also linked with the neurotoxicity of Ab and free radical damage associated with Alzheimer's disease (16). Finally, trace amounts of copper in the diet have been shown to induce betaamyloid plaques and learning deficits in a rabbit model of Alzheimer's disease (28). To conclude, we believe our studies have made a significant contribution to understanding Cu 2+ binding of Ab in solution by identifying the key ligands coordinating the Cu 2+ ion and the effect of pH on the binding mode and structuring of the main-chain of Ab.