Effect of C-Terminal Residues of Aβ on Copper Binding Affinity, Structural Conversion and Aggregation

Many properties of Aβ such as toxicity, aggregation and ROS formation are modulated by Cu2+. Previously, the coordination configuration and interaction of Cu2+ with the Aβ N-terminus has been extensively studied. However, the effect of Aβ C-terminal residues on related properties is still unclear. In the present study, several C-terminus-truncated Aβ peptides, including Aβ1-40, Aβ1-35, Aβ1-29, Aβ1-24 and Aβ1-16, were synthesized to characterize the effect of Aβ C-terminal residues on Cu2+ binding affinity, structure, aggregation ability and ROS formation. Results show that the Aβ C-terminal residues have effect on Cu2+ binding affinity, aggregation ability and inhibitory ability of ROS formation. Compared to the key residues responsible for Aβ aggregation and structure in the absence of Cu2+, it is more likely that residues 36–40, rather than residues 17–21 and 30–35, play a key role on the related properties of Aβ in the presence of Cu2+.


Introduction
Alzheimer's disease (AD) is a neurodegenerative disorder that destroys neuronal cells in the human brain [1,2]. Numerous reports have shown that one of the pathological hallmarks in the brain of AD patients is the cerebral senile plaques [1,2]. Senile plaques contain 90% of b-amyloid peptide (Ab), including Ab1-40 and Ab1-42, which is a proteolytic product of amyloid precursor protein (APP) [3][4][5]. The others remained in senile plaques include apolipoproteins E, lipids from membranes of degenerated portions of neuron, and abnormally high concentration of metal ions such as Cu 2+ , Zn 2+ , or Fe 2+ [6,7].
In the amyloid cascade hypothesis, the Ab aggregates are proposed to be the main toxic species and the cause of AD [8,9]. Ab adopts a b-sheet conformation in the aggregated state, and the amyloid aggregates can induce free radical formation and subsequently cause neuronal death [8,9]. Among the Ab aggregates, oligomer and protofibril rather than mature amyloid fibril have been demonstrated to be the most toxic species to neurons [10][11][12]. The aggregation and toxicity of Ab is well correlated with its sequence and structure [13][14][15]. Previous studies using different Ab fragments or truncated Ab peptides reported that residues 17-21 and 30-35 are the most important regions for aggregation and neurotoxicity [14,15].
The deposition of Ab has been shown to be modulated by metal ions [6,7], particularly Cu 2+ . Abnormally high concentrations of Cu 2+ have been found in cerebral amyloid-deposits of AD patients [6]. It has been shown that Cu 2+ is bound to Ab [6,16]. Either one or two Cu 2+ bound to Ab peptide has been proposed. The binding site is mainly located at the N-terminus of Ab, particularly the three histidine residues (His6, His13 and His14), and forms a 3N1O coordination configuration [16,17]. The reported Cu 2+ binding affinities for monomeric Ab vary widely between micromolar and nanomolar [18]. The effect of Cu 2+ ion on Ab has been shown to be twofold, the first is to accelerate the aggregation of Ab, [19,20], and the second is to induce the formation of reactive oxygen species (ROS) [18,21].
The role of Ab coordinated with Cu 2+ on the free radical has still been under debate. Both pro-oxidant and antioxidant roles for Ab associated with the ROS produced by Cu 2+ have been suggested [21][22][23]. Although early studies suggested that Ab peptides can spontaneously produce free radicals [14,18,19,22], several studies have shown that Ab required the presence of Cu 2+ to produce ROS [18,21,22]. The possible mechanism of ROS formation may be through a series of electron transfer reactions when Cu 2+ binds to Ab [18,21]. The ROS induced by Ab/Cu 2+ aggregates is reduced by addition of other antioxidants or Cuselective chelators [17,[23][24][25]. As opposed to the pro-oxidant role, other studies have proposed an antioxidant activity of Ab [26][27][28]. In particular, monomeric Ab1-40 has been shown to inhibit neuronal death caused by Cu 2+ induced oxidative damage [27,28]. Furthermore, Viles group has demonstrated that Ab does not silence the redox reaction of Cu 2+ via chelation but react with the hydroxyl radicals produced by Cu2 + /ascorbate and quench the harmful oxidative species [26].
The effect of Ab sequence on structure, aggregation ability and ROS formation in the absence of Cu 2+ has been extensively studied [14,15]. In general, residues 17-21 and 30-35 are identified as the key region responsible for aggregation and neurotoxicity [14,15]. In the presence of Cu 2+ , the interaction and coordination configuration of Ab/Cu 2+ complex have been characterized [16][17][18]. However, most of these studies have focus on the elucidation of interaction and coordination configuration of the Ab N-terminus with Cu 2+ . So far, the effect of Ab C-terminal residues on Cu 2+ binding affinity, aggregation ability and ROS formation in the presence of Cu 2+ has yet been studied.
In the present study, we investigated the effect of Ab C-terminal residues on Cu 2+ binding affinity, structure, aggregation ability and ROS formation. Full length Ab1-40 and several C-terminustruncated Ab peptides, including Ab1-35, Ab1-29, Ab1-24 and Ab1-16 were synthesized and used to characterize these subjects. Our results indicated that, though the major Cu 2+ binding site is located at N-terminus of Ab, the C-terminal residues of Ab, particularly residues 36-40, have a significant effect on the binding affinity of Cu 2+ , conformation, aggregation ability and the inhibitory ability of ROS driven by Cu 2+ .

Synthesis and purification of Ab peptides
The synthesis of Ab peptides, including Ab1-40, Ab1-35, Ab1-29, Ab1-24 and Ab1-16, were performed in a solid-phase peptide synthesizer (PS3, Protein Technologies, Inc., AZ) using the FMOC protocol with HMP resin. After cleavage from the resin with a mixture of trifluoroacetic acid/H 2 O/ethanedithiol/thioanisole/ phenol, the peptides were extracted with 1:1 (v:v) ether: H 2 O containing 0.1% 2-mercaptothanol. The synthesized Ab peptides were purified using a C18 reverse-phase column with a linear gradient from 0% to 78%. Peptide purity was over 95% as identified by MALDI-TOF mass spectrometer. One mg of purified Ab peptides was dissolved in 1 ml trifluoroethanol, and centrifuged (20,0006g) to sediment the insoluble particles. This Ab solution was then dried under N 2 gas and resuspended in 1 ml phosphate buffer, pH 7.4, to provide a stock solution, and stored at 280uC until used.

Copper binding affinity assay
Tyrosine fluorescence spectroscopy was used to characterize the binding affinity of Cu to Ab [28]. Before measurements, the stock solution containing the different C-terminal truncated Ab peptides was diluted in Dulbecco's PBS, pH 7.0 to a final peptide concentration of 10 mM with different molar ratios of CuCl 2 . Spectra were collected on a microplate reader (FlexStation 3, MD). The excitation and emission wavelength was 278 and 305 nm, respectively. The intensity change at 305 nm was used to calculate the binding constant. Previously, the number of Cu ion bound to Ab has been debated. Either one or two Cu ion has been proposed to bind to Ab [29], and there is no two-Cu/Ab complex structure available. Two-degenerate scheme for either one-or two-Cu binding modes was hence considered and applied to calculate the binding constant.
For the one-Cu 2+ binding mode, the general equation for Cu 2+ binding is as follows:

AbzCu 2z
AbCu 2z K a~½ AbCu 2z ½Ab½Cu 2z , the degree of saturation, Y, can be written as , where I o and I x are the fluorescence intensity in Cu-free and Cubound state, respectively. I ' is the fluorescence intensity at saturation state, [Cu 2+ ] is the copper concentration, n is the copper binding number and K a is the association constant.
For the two-Cu 2+ binding mode, the two Cu 2+ ions are bound to Ab located at the N-terminal His-pocket. The general equation for Cu 2+ binding is as follows: The tyrosine fluorescence spectrum at any concentration is the net combination of the Cu-free and Cu-bound forms weighted by their concentrations. Two general models based on linked two-site binding are proposed.
The first model (dependent mode) is one in which the two Cu binding sites interact so that the second Cu 2+ ion binds with a different binding constant than the first. The degree of saturation for the dependent mode can be described as follows [30]: Y~K a1 ½Cu 2z z2K a1 K a2 ½Cu 2z 2 1zK a1 ½Cu 2z zK a1 K a2 ½Cu 2z 2 The second model (independent mode) assumes that the two binding sites, due to the different structure or accessibility, are independent with each other and should have equal binding constant (K a1 = K a2 ) for the two Cu 2+ ions [30]. The degree of saturation for the independent mode is described as follows: The related parameter was calculated using the nonlinear curve fitting function in the Origin6.0 program (Microcal Software, Inc., Nothampton, MA). This nonlinear fitting program uses the Levenberg-Marquardt nonlinear least-squares fitting algorithm. In the initial fitting stage, the Simplex method, which was set to 100 cycle runs, was used to calculate the initial parameter for further nonlinear curve fitting. A 0.95 confidence level was set to constrain the quality of curve fitting. The final fitting parameters were obtained when the value of x 2 was less than 0.05 and the parameters and errors for the parameters reached the convergent and steady state.

Circular dichroism (CD) spectroscopy
Thirty mM of fresh peptide samples, diluted from the stock solution in phosphate buffer, pH 7.0, in the presence or absence of 30 mM Cu 2+ were used for CD measurements. CD spectra were recorded, within 1 hr after samples prepared, using either an Aviv 420 spectropolarimeter or synchrotron radiation CD (04B1) in the national synchrotron radiation center, Taiwan. All measurements were performed in a quartz cell with pathlength of 0.1 cm. Spectra were collected at the wavelengths from 190 to 260 nm in 0.5 nm increments. Reported CD spectra were the average from three repeats of samples. The reported CD spectra were corrected for baseline using the solution of PBS buffer, pH 7.0 and Cu 2+ ions. The secondary structure analysis was calculated using CDSSTR program in Dicroweb website [31].
The b-sheet propensity is defined as represent the percentage of b-sheet content in Cu-free and saturated Cu-bound state, respectively. C o and C ' are the concentrations of Cu 2+ in Cu-free and saturated Cu-bound state, respectively.

Aggregation assay
The aggregation process of Ab peptides in the presence or absence of Cu 2+ was assessed by the turbidity assay. Thirty mM of Ab peptides were placed in a 96-well plate and incubated in the presence or absence of 30 mM CuCl 2 at 37uC. Turbidity was measured using a microplate reader (FlexStation 3, MD) at a wavelength of 450 nm.

ROS assay
ROS (H 2 O 2 ) level induced by Ab/Cu 2+ was analyzed using the dichlorofluoresein diacetate (DCFH-DA) assay [17]. Dichlorofluorescein diacetate was dissolved in 100% dimethyl sulfoxide (DMSO), deacetylated with 1:1 (v/v) 4 M NaOH for 30 min, and then neutralized (pH 7.2) to a final concentration of 200 mM as stock solution. This stock solution was kept on ice and in the dark until use. The reaction was carried out in a 96-well plate (100 ml/ well) in Dulbecco's PBS, pH 7.2, containing the designed concentrations of Ab peptides, 30 mM of CuCl 2 , 20 mM deacylated DCF and 5 mM horseradish peroxidase, and incubated at 37uC for 1 hr. Measurements were performed on the day of sample prepared. Fluorescence readings were recorded on the microplate reader (Flexstation3, MD). The excitation and emission wavelengths were 485 and 530 nm, respectively.

Electron paramagnetic resonance (EPR) spectroscopy
Samples containing 300 mM of Ab peptides and Cu 2+ ions in 30% glycerol phosphate buffer, pH 7.2, freshly prepared from peptide stock solution were employed for EPR spectroscopic measurements. EPR spectra were obtained at X-band using a Bruker EMX ER073 spectrometer equipped with a Bruker TE102 cavity and an advanced research system continuous-flow cryostat (4.2-300 K). During EPR experiments, the sample temperature was maintained at 10 K. The microwave frequency was measured with a Hewlett-Packard 5246L electronic counter.

Transmission Electron Microscopy (TEM)
A transmission electron microscopy (JEM-2000 EXII, JEOL, Japan) with an accelerating voltage of 100 KeV was used to analyze the morphology of Ab peptides incubated with Cu 2+ . Ten microliters of sample with the different Ab peptides and Cu 2+ ions in 1:1 molar ratio used for the aggregation assay was used. Each peptide sample was placed onto a carbon-coated 200 mesh copper grid (Pelco, Ca, USA). Excess solution was wicked dry with tissue paper, and the sample was negatively stained with 5 ml of 2% uranyl acetate for 30 seconds. After TEM analyses, these copper grids coated with Ab samples used for TEM analyses were further treated with 50 mL 1 mM EDTA solution three times to strip off Cu 2+ ions and then incubated at 37uC for 24 hrs. These copper grids coated with Ab samples treated with EDTA were then conducted for TEM analyses to observe the morphology of Ab peptides in absence of Cu 2+ .

Correlation of Cu binding affinity and Ab sequence
The aggregation and toxicity of Ab has been demonstrated to be modulated by Cu 2+ [6,7,17,18,21]. The interaction of Cu 2+ with Ab N-terminus has been extensively studied [17,18,[32][33][34]. The number of Cu 2+ bound to Ab has been debated which either one or two Cu 2+ has been proposed [17,18,29]. On the other hand, the effect of C-terminal residues on Cu 2+ binding affinity and other properties has yet to be studied. In order to unveil the effect of C-terminal residues on Cu 2+ binding affinity and other properties, several Ab peptides, including Ab1-40, Ab1-35, Ab1-29, Ab1-24, Ab1-16 and Ab25-35, were synthesized and used to characterize the correlation with Cu 2+ binding affinity, structural changes and aggregation ability.
To characterize the Cu 2+ binding affinity, tyrosine fluorescence spectroscopy was used to determine the Cu 2+ binding constants. Figures 1 (A-E) show the tyrosine fluorescence titration curves as a function of tyrosine fluorescence intensity vs. Cu 2+ concentration for Ab1-40, Ab1-35, Ab1-29, Ab1-24, and Ab1-16, respectively. Both one-Cu and two-Cu binding modes were applied to estimate the Cu 2+ binding constants. As shown in Fig. 1, the two-Cu mode (solid line) shows to fit the titration curve better than the one-Cu mode (dot line) for all Ab peptides, indicating that the Cu 2+ binding site is more likely to locate two ions instead of one ion for all Ab peptides. For the two-Cu mode, we further tested if the two Cu 2+ ions bound to Ab are dependent or independent of each other. As shown in Figs. 1 (A-E), for all Ab peptides, the nonlinear fitting curves were only convergent by using the dependent mode, suggesting that the binding constant of two Cu 2+ ions should be different for each other.
The calculated binding constants are summarized in Table 1. The K a1 value was approximately hundredfold higher than the K a2 value for all Ab peptides, indicating that the first Cu binds to Ab much stronger that the second Cu does. The K a1 value was in the range of 0.06-0.13 mM, and the K a2 value was in the range of 0.0007-0.0013 mM. In general, both K a1 and K a2 were dependent on sequence. The value of K a1 was increased with an increase of Ab C-terminal residues, except of Ab1-35. The K a1 value of Ab1-40 was approximately twofold higher than that of Ab1-16. In contrast, the trend of K a2 value was opposite to that of K a1 value which the K a2 values of Ab1-24 and Ab1-16 were higher than those of Ab1-29, Ab1-35 and Ab1-40. The K a2 value of Ab1-24 was approximately twofold higher than that of Ab1-35. The K a1 value for Ab peptides was in the order of Ab1-40$ Ab1-29$ Ab1-35< Ab1-24. Ab1-16, whereas the K a2 value for Ab peptides was in the order of Ab1-24$ Ab1-16$ Ab1-29< Ab1-40< Ab1-35.

EPR spectra of Ab/Cu 2+ complexes
As we showed that the C-terminal residues of Ab can affect the Cu 2+ binding affinity, it is of interest to examine if the interaction of the C-terminal residues with Cu 2+ alters the coordination configuration of Ab/Cu 2+ . In order to characterize the coordination configuration, EPR spectroscopy was used to determine the coordination configuration of Ab/Cu 2+ for the different Cterminus-truncated Ab peptides. Figure 2 shows the EPR spectra for Cu 2+ with Ab1-40, Ab1-35, Ab1-29, Ab1-24 and Ab1-16. The EPR parameters of g ) , g | | and A are listed in Table 2. It can be seen that the hyperfine peaks of EPR spectra for the different Ab peptides showed a similar pattern. The estimated g ) , g | | and A parameters for the different Ab peptides were similar and very close to literature report in aqueous condition except of Ab1-16 [33,34]. The values of g ) , g | |   [17,18,29,32,33], and the main coordination configuration of Ab/Cu 2+ , located at the N-terminus was not significantly altered by the association of C-terminus of Ab.
Secondary structure of Ab peptides in the presence of Cu 2+ Previous results show that the increase of C-terminal residues increased the Cu 2+ binding affinity but did not cause any significant change of Ab-Cu 2+ coordination configuration. However, several studies have shown that the binding of Cu 2+ to Ab can induce a conformational conversion from either helix or random coil into b-sheet [19,20]. Therefore, the effect of Cterminal residues on the secondary structure of Ab peptides in the presence of Cu 2+ was examined by using CD spectroscopy. Figures 3 (A) and (B) show the CD spectra for the different Ab peptides in the absence or presence of Cu 2+ ions, respectively. Table 3 summarizes the estimated content of secondary structure. In general, in the absence of Cu 2+ , all Ab peptides adopt a high percentage of random coil. Ab1-16 contained the highest percentage of random coil (72%) and the lowest percentage of b-sheet (24%), whereas other Ab peptides contained a similar secondary structure content, 30-34% of b-sheet, 61-64% of random coil and 4-5% of a-helix. In the presence of Cu 2+ , the secondary structure content for Ab1-35, Ab1-29, and Ab1-24 peptides was similar to that obtained in the absence of Cu 2+ . For Ab1-16, the b-sheet percentage was slightly increased (27%), and the random coil percentage was slightly decreased (70%). In contrast to other C-terminus-truncated Ab peptides, the secondary structure of Ab1-40 showed a dramatic change while adding the Cu 2+ ions. The b-sheet content of Ab1-40 increased from 34% to 47%, and the random coil percentage of Ab1-40 decreased from 61% to 50% in the presence of Cu 2+ .
We further analyzed the correlation between secondary structure and Cu 2+ concentration. The plot of secondary structural content (b-sheet, random coil and a-helix) vs. Cu 2+ concentration for Ab peptides is depicted in Figures 4 (A-C), respectively. In general, the contents of b-sheet and random coil for Ab1-40, Ab1-35, Ab1-29, and Ab1-24 were dependent on Cu 2+ concentration, whereas the secondary structure content of Ab1-16 was independent of Cu 2+ concentration. For Ab1-40, Ab1-35, Ab1-29, and Ab1-24, the content of b-sheet structure increased with an increase of Cu 2+ concentration ( fig. 4 (A)), whereas the content of random coil decreased with an increase of Cu 2+ concentration ( fig. 4 (B)). The a-helix content showed no obvious change with the increase of Cu 2+ concentration for all Ab peptides ( fig. 4 (C)). The change of b-sheet content for Ab1-40 was more significant than those for other Ab peptides. Furthermore, the relationship between b-sheet propensity and Ab sequence in the presence of Cu 2+ was also correlated by plotting the b-sheet propensity in the presence of Cu 2+ vs. Ab sequence. As it can be seen that has the b-sheet propensity of Ab1-40 is significantly higher than those for other Ab peptides. This is generally agreement with the result obtained from K a1 binding constant which Ab1-40 has the higher Cubinding affinity.   Aggregation ability for Ab peptides It has been shown that the binding of Cu 2+ can modulate the aggregation mechanism of Ab [6,14,33]. As demonstrated by the present study, the conformational conversion into b-strand structure was dependent on Ab C-terminal residues in the presence of Cu 2+ . In order to further characterize the effect of C-terminal residues on Ab aggregative ability, we analyzed the aggregation profiles for the different C-terminus-truncated Ab peptides in the presence of Cu 2+ . Figures 5 (A) and (B) show the aggregation profiles for Ab1-40, Ab1-35, Ab1-29, Ab1-24, and Ab1-16 in the absence and presence of Cu 2+ , respectively. As shown in figure 5 (A), only Ab1-40 was able to form aggregates in the absence of Cu 2+ , whereas the other Ab peptides remained at nucleation state in the absence of Cu 2+ . In the presence of Cu 2+ , Ab1-40, Ab1-35 and Ab1-29 were able to aggregate, whereas Ab1-24 and Ab1-16 remained at nucleation state ( fig. 5 (B)), indicating that the Cterminal residues of Ab, particularly residues 25-40, have effect on the aggregation in the presence of Cu 2+ . The aggregation rate for Ab1-40 in the presence of Cu 2+ was faster than the rates for Ab1-35 and Ab1-29. This result further suggests that the C-terminal residues, particularly 36-40, may play an important role on the aggregation mechanism of Ab driven by Cu 2+ . In general, the

TEM morphology of Ab peptides
As shown in the previous sections that the C-terminal residues of Ab have impact on Cu 2+ binding affinity, secondary structure and aggregative ability. Therefore, we wondered if the C-terminal residues have any effect on the morphologies of fibrils formed by the Ab peptides. To examine the effect of C-terminal residues on the morphologies of Ab fibrils in the presence of Cu 2+ , transmission electronic microscopy was used to observe the fibril morphologies. Figures 6 (A), (C), (E) and (G) show the morphologies for Ab1-40, Ab1-36, Ab1-29 and Ab1-16 in the presence of Cu 2+ (molar ratio Ab/Cu = 1), respectively. It can be seen that most Ab peptides formed a non-amyloid-like morphology in the presence of Cu 2+ , except of Ab1-16 which did not form any amyloid fibrils. The same Ab peptides/Cu 2+ samples of the same spots were then treated with 1mM EDTA to strip off the Cu 2+ ions and further incubated at 37uC and 24 hrs. After Cu 2+ ions were depleted by EDTA, the morphology of Ab peptides was then analyzed using TEM. Figures 6 (B), (D), (F) and (H) show the TEM images for morphologies of Ab1-40, Ab1-36, Ab1-29 and Ab1-16, respectively. The morphologies for Ab1-40, Ab1-35 and Ab1-29 aggregates in the presence of Cu 2+ are obviously very different from the morphologies of these peptides with Cu 2+ stripped off by EDTA. In figure 6 (B), the fibril of Ab1-36 was a typical amyloidogenic and network-like morphology after Cu 2+ was stripped off by EDTA. On the other hand, the fibrils of both Ab1-29 ( fig. 6(E)) and Ab1-40 ( fig. 6(A)) with Cu 2+ stripped off by EDTA were shorter and non-network-like morphology. For both Ab1-16 and Ab1-24, they did not form any fibril in the presence or absence of Cu 2+ .

Correlation of H 2 O 2 formation and Ab sequence in the presence of Cu 2+
The role of Ab/Cu 2+ on the formation of ROS is controversial. Both antioxidant and pro-oxidant roles for Ab on the ROS formation in the presence of Cu 2+ have been proposed [20][21][22][23]. In order to elucidate the effect of Ab C-terminal residues on either antioxidant or pro-oxidant role, a DCF assay which usually detects the formation of H 2 O 2 was used to measure the level of ROS for the different Ab peptides in the presence of Cu 2+ . Figure 7 shows the plot of DCF fluorescence intensity vs. Ab concentration. For most C-terminus-truncated Ab peptides, the DCF fluorescence intensity was decreased with an increase of Ab concentration, indicating that the formation of H 2 O 2 was inhibited by most Ab peptides, except of Ab25-35. For Ab25-35, which lacks the Cu 2+ binding site, did not show to inhibit the formation of H 2 O 2 . The H 2 O 2 level was equal to that of Cu 2+ only. For the comparison of inhibitory ability for these Ab peptides, only full-length Ab1-40 was able to completely inhibit the formation of H 2 O 2 at the molar ratio of Ab/Cu 2+ = 1, whereas for other peptides such as Ab1-35, Ab1-29, Ab1-24 and Ab1-16, the formation of H 2 O 2 was not completely inhibited at the molar ratio of Ab/Cu 2+ = 1. A higher peptide concentration was needed to completely reduce the H 2 O 2 level to zero for Cterminus-truncated Ab peptides.
Since the Cu 2+ binding affinity was showed to be proportional with the length of C-terminal residues, taken together, our results further suggest that the Cu 2+ binding affinity may be the key factor for the inhibition of H 2 O 2 formation driven by Cu 2+ . In general, the inhibitory ability of ROS for these Ab peptides was proportional with the binding affinity of Cu 2+ and in the order of Ab1-40. Ab1-29. Ab1-35< Ab1-24< Ab1-16.. Ab25-35.

Discussion
Amyloid cascade hypothesis proposes that the aggregated Ab species are toxic to neurons and the main cause of Alzheimer's disease [6]. The various forms of Ab, including monomer, oligomer and fibril, have been shown to coordinate with redox active transition metals, such as Cu 2+ and Fe 3+ , which induce the formation of ROS [17,18,20,21]. Although the interaction and coordination configuration of Ab/Cu 2+ complexes has been extensively studied [16][17][18]29,[32][33][34], the effect of Ab sequence, particularly C-terminal residues, on Cu 2+ binding affinity, structural property, aggregative ability and ROS formation still remains to be elucidated.
In the present study, results demonstrate that the C-terminal residues of Ab have significant effect on Cu 2+ binding affinity, structure, aggregation ability and inhibitory ability of ROS. For Cu 2+ binding affinity, the C-terminal residues of Ab, particularly residues 25-29 and 36-40, have a strong effect on Cu 2+ binding affinity as evidenced by the fact that the Cu 2+ binding constants for Ab1-40 and Ab1-29 are higher than those for other C-terminus- truncated Ab peptides. Even though the Cu 2+ binding affinity is dependent on C-terminal residues of Ab, the coordination configurations of Ab/Cu 2+ are not significantly altered by the interaction C-terminal residues with Cu 2+ as the hyperfine patterns and parameters obtained from EPR spectroscopy are similar for these different Ab peptides. The coordination configuration of Ab/Cu 2+ still adopt the a 3N1O mode, and His6, His13 and His14 residues are the main amino acid residues to interact with Cu 2+ even for C-terminus-truncated peptides [32,33]. For the Cu-binding mode, our present study shows that the Cu 2+ binding site for all Ab peptides is able to locate two Cu 2+ ions. These two Cu 2+ ions bound to Ab is dependent on the Cterminal residues. The binding constant for the first Cu 2+ ion is higher than that for the second Cu 2+ ion. This is consistent with a previous study [29]. The fold difference between the first Cu 2+binding constant and the second Cu 2+ -binding constant is also dependent on Ab C-terminal residues, ranged from 160 folds for Ab1-40 to 10 folds for Ab1-16, respectively. However, the binding constants obtained in this study are somehow lower than the previous report [29]. The possible reason may be two folds; the first reason may be due that the concentration of Ab peptides used is lower than the previous study, and the second reason may be caused by the different method applied.
It is interesting to note that the trend of K a1 and K a2 is generally opposite to each other. The K a1 values are higher for Ab peptides with residues 25-29 and 36-40, whereas the K a2 values for Ab peptides with residues 25-29 and 36-40 are generally lower compared to other C-terminus-truncated peptides. This indicates that the residues 25-29 and 36-40 possibly increase the binding affinity of the first Cu 2+ ion and decrease the binding of the second Cu 2+ ion. This result may provide an explanation for the previous observation that the second Cu site is only observed in the shorter truncated Ab peptides such as Ab1-16 [35], since the C-terminal residues, particularly residues 36-40, may impede the binding of the second Cu 2+ . However, the second Cu 2+ binding constant is rather small compared to the binding constant of the first Cu 2+ , Figure 6. The TEM images of Ab fibril morphologies. Images A, C, E and G represent the fibril morphologies for Ab1-40, Ab1-35, Ab1-29 and Ab1-16 with Cu 2+ stripped off by EDTA, respectively. Images B, D, F and H represent the morphologies for Ab1-40, Ab1-35, Ab1-29 and Ab1-16 in the presence of Cu 2+ , respectively. doi:10.1371/journal.pone.0090385.g006 thereby the role of second Cu 2+ ion also has little effect on the function of Ab such as coordination geometry. The effect of Cubinding on Ab function is mainly attributed from the binding of the first Cu 2+ ion.
Besides the effect of C-terminal residues on Cu 2+ binding affinity, the C-terminal residues also show to have impact on the structural property of Ab in the presence of Cu 2+ . From the result of secondary structural analysis, Ab1-40 has the highest b-sheet propensity, indicating that residues 36-40 may play a key role on structural conversion of Ab from random coil into b-sheet driven by Cu 2+ . Previously, residues 17-21 and 30-35 have been shown to be the key regions on the conformation stability for Ab in the absence of Cu 2+ [14,15]. Our present results indicate that, in the presence of Cu 2+ , residues 36-40 may be more important than residues 17-21 and 30-35 for the conformational conversion of Ab driven by Cu 2+ . Recently, a solid-state NMR study showed that the hydrophobic core regions of residues 18-25 and 30-36 of fibril Ab/Cu 2+ complex have little structural change [34]. Our present result is consistent with their study.
A similar effect of C-terminal residues on Ab aggregation was obtained in the presence of Cu 2+ , since the aggregation ability of Ab is highly associated with the ability of structural conversion into b-sheet [14,15]. Previous studies showed that, in the absence of Cu 2+ , residues 17-21 and 30-35 are the most important regions for aggregation and neurotoxicity of Ab [14,15]. The present results show that the structural feature responsible for the aggregation in the presence of Cu 2+ is very different from that in the absence of Cu 2+ . In the presence of Cu 2+ , the residues 36-40, instead of residues 17-21 and 30-35, are the key amino acid residues responsible for the aggregation, as Ab1-40 has the fastest aggregation rate.
The role of Ab on the ROS production in the presence of Cu is still under debate. Both inhibition and production of ROS by Ab/ Cu complex have been proposed [21,22,27,28]. However, our results show that Ab inhibits the ROS production in the presence of Cu. Recently, Fang and his colleagues have reported that H 2 O 2 production is highly dependent on the state of Ab, which monomeric Ab tends to inhibit H 2 O 2 production in the presence of Cu, whereas Ab oligomer and fibril in the presence of Cu can induce H 2 O 2 production [27]. According to their finding, our result may indicate that Ab used in the present study may exist at monomeric state.
For the inhibition of ROS production, similar sequence-effect was also observed for the inhibitory ability of ROS formation. Results show that the inhibitory ability is also well correlated with the C-terminal residues of Ab. Ab1-40 with the C-terminal residues 36-40 is the only peptide which can completely inhibit the H 2 O 2 formation, whereas the other C-terminus-truncated peptides, lacking the residues 36-40, can only inhibit the level of H 2 O 2 to a less degree. Furthermore, the inhibitory ability is also dependent on Cu 2+ binding affinity, as the binding affinity is correlated with the length of C-terminal residues. Therefore, the stronger binding constant, the higher inhibitory ability of ROS formation is.
Previous studies showed that Ab peptides form non-amyloidogenic aggregates in the presence of Cu 2+ and amyloidogenic fibril in the absence of Cu 2+ [36][37][38]. Our present study also examined the morphologies for these Ab peptides in the presence of Cu 2+ . Results show a similar observation which all Ab peptides with Cu 2+ form non-amyloidogenic aggregates. After Cu 2+ ions stripped off by EDTA, only Ab1-40, Ab1-35 and Ab1-29 can form a typical amyloidogenic fibril, but both Ab1-24 and Ab1-16 do not form any fibril. It is of interest to note that the morphologies of amyloidogenic fibril formed by Ab1-40 and Ab1-29 and are slightly different from the morphology of Ab1-35 fibril. Both Ab1-40 and Ab1-29 form a short, rod and nonnetwork-like fibril, whereas morphology of Ab1-36 fibril is a typical thin and rod-and network-like fibril. The cause of the different morphologies between these peptides is unclear, but it is correlated with the sequence of Ab and Cu 2+ binding affinity, since the Cu 2+ binding affinity of Ab1-35 is weaker than those of Ab1-40 and Ab1-29.
In conclusion, our present results demonstrate i that the Cterminal residues of Ab have a significant effect on Cu 2+ binding affinity, structure, aggregation ability and inhibitory ability of ROS formation. Among the C-terminal residues, residues 36-40 play the most important key role on these properties. The involvement of C-terminal residues 36-40, instead of residues 17-21 and 30-35, on Cu 2+ binding affinity, b-sheet conversion, aggregation ability and inhibitory ability may provide a possible explanation for the different behavior of Ab in the presence of Cu 2+ .