Copper Transporters? Glutathione Reactivity of Products of Cu–Aβ Digestion by Neprilysin

Aβ4–42 is the major subspecies of Aβ peptides characterized by avid Cu(II) binding via the ATCUN/NTS motif. It is thought to be produced in vivo proteolytically by neprilysin, but in vitro experiments in the presence of Cu(II) ions indicated preferable formation of C-terminally truncated ATCUN/NTS species including CuIIAβ4–16, CuIIAβ4–9, and also CuIIAβ12–16, all with nearly femtomolar affinities at neutral pH. Such small complexes may serve as shuttles for copper clearance from extracellular brain spaces, on condition they could survive intracellular conditions upon crossing biological barriers. In order to ascertain such possibility, we studied the reactions of CuIIAβ4–16, CuIIAβ4–9, CuIIAβ12–16, and CuIIAβ1–16 with reduced glutathione (GSH) under aerobic and anaerobic conditions using absorption spectroscopy and mass spectrometry. We found CuIIAβ4–16 and CuIIAβ4–9 to be strongly resistant to reduction and concomitant formation of Cu(I)–GSH complexes, with reaction times ∼10 h, while CuIIAβ12–16 was reduced within minutes and CuIIAβ1–16 within seconds of incubation. Upon GSH exhaustion by molecular oxygen, the CuIIAβ complexes were reformed with no concomitant oxidative damage to peptides. These finding reinforce the concept of Aβ4–x peptides as physiological trafficking partners of brain copper.

A β peptides are products of extracellular hydrolysis of amyloid precursor protein (APP) present in neuronal synaptic membranes. 1−3 Overproduction or excessive aggregation of Aβ has been long implicated as an upstream cause of neuronal death in Alzheimer's disease (AD). 4,5 This concept gained recent reinforcement when direct deleterious action of Aβ dimers on neuronal glutamate receptors was demonstrated. 6 Aβ peptides are a heterogeneous peptide family, generated by a number of primary (acting on APP) and secondary (acting on Aβ) proteases. The most studied members of the Aβ family are Aβ 1−40 and Aβ 1−42 , but current analytical studies demonstrated a high abundance of the Ntruncated Aβ 4−42 peptide in both healthy and AD human brains. 7,8 Aβ 1−x peptides (x denotes naturally occurring 42 and 40 species, as well as model peptides 28 and 16) bind a Cu II ion avidly with K d about 100 pM. The resulting complexes can be easily activated by ascorbate to catalyze the production of reactive oxygen species (ROS). 9−12 Supported by reports on deranged copper metabolism in AD brains and colocalization of copper and aggregated Aβ peptides in amyloid plaques, these properties gave rise to a concept of Cu II −Aβ 1−x complexes as neurotoxic species in AD. 5,13,14 Remarkably, Aβ 4−42 and its C-terminally truncated analogs are Cu II chelators much more avid (3000 times at pH 7.4 for Aβ 4−16 vs Aβ 1−16 ) and specific than Aβ 1−x peptides. 15 This results from a specific character of their N-terminal sequence, Phe-Arg-His, belonging to the ATCUN/NTS family. 16 Furthermore, unlike Cu II −Aβ 1−x complexes, Cu II −Aβ 4−x did not generate ROS and could not be reduced electrochemically to Cu I species. 15 These findings suggest that Aβ 4−42 might actually serve as synaptic copper scavenger, helping restore the resting state of glutamatergic synapse, after the physiological Cu 2+ release during neurotransmission. 17,18 Digestion of Aβ peptides is considered as one of the major routes of their detoxification. They are thought to be cleaved down to oligopeptides that can cross the blood−brain barrier. 19,20 An Aβ-specific peptidase has not been found. Instead, a number of brain proteases with other known functions have been implicated in this process, including neprilysin (NEP), angiotensinogen converting enzyme (ACE), and insulin degrading enzyme (IDE). 21 NEP action on Aβ 1−x has also been indicated as the main source of Aβ 4−x in the brain. 22,23 A recent study of Aβ 1−16 and Aβ 1−40 cleavage by NEP in the presence and absence of Cu II ions did not quite reproduce such activity, however. Instead a significant extent of peptide fragmentation was observed. The fast digestion of the Gly9−Tyr10 bond yielded the Cu II -complexed short peptide Aβ 4−9 . Additionally, a complex of Aβ 12−16 was generated abundantly when Aβ 1−16 was used as a substrate and was also present as a minor species for Aβ 1−40 . 24 Aβ 4−9 and Aβ 12−16 are even stronger Cu II chelators than Aβ 4−16 , with K d of 6.6 fM and 9.5 fM vs 30 fM at pH 7.4. 25 This finding gave rise to an idea that such complexes might serve as shuttles for removing excess copper from the brain.
Crossing the blood−brain barrier (BBB) is a complicated and not fully elucidated process, involving passage through the layer of epithelial cells forming the blood vessel walls. 26 Therefore, the transferred molecule could be exposed for a certain amount of time to intracellular conditions, including millimolar (0.5−10 mM) concentrations of reduced glutathione (GSH). 27 GSH is the main reducing agent for Cu II species entering the cell interior and is also implicated in the intracellular Cu I transport. 28−30 It is also present extracellularly in the brain, serving as neuromodulator. 31 GSH facilitated the otherwise very sluggish reductive copper transfer from Aβ 4−16 to metallothionein-3 (MT-3), indicating that it could reduce the Aβ 4−x -bound Cu(II) to Cu(I) despite the electrochemical resistance of the parent complex to such reaction. 32 We therefore decided to follow the reaction of Aβ 4−x peptides with GSH in more detail, using Aβ 4−16 as a suitable soluble, nonaggregating substitute of Aβ 4−42 . We also tested Aβ 4−9 and Aβ 12−16 . Our experiments were performed under aerobic (21% O 2 ) and anaerobic (<1% O 2 ) conditions in order to gain insight into the relation of the studied reaction to oxidative stress conditions. The differential kinetic resistance of the studied complexes to reduction supports their possible roles in Cu II transport in the brain.
In initial experiments, 0.315 mM Cu 2+ ions were reacted for 24 h with 1.75 mM GSH in 20 mM ammonium acetate at 25°C under aerobic conditions, with and without 0.35 mM Aβ 4−16 (0.9/5/1 and 0.9/5 molar ratios, Figure 1A and Supporting Information Figure S1, respectively). In control experiments 1.75 mM glutathione disulfide (GSSG) was used instead of GSH ( Figure S2), and Cu 2+ ions were omitted from the reaction of GSH with Aβ 4−16 ( Figure S3). These experiments allowed us to identify and assign the spectral changes occurring in the course of reactions of Cu II (Aβ 4−16 ) with GSH. New bands in the near-UV range between 315 and 265 nm ( Figure S4) appeared gradually in the presence of Aβ 4−16 , at the expense of the Cu(II) band of the 4N complex at 525 nm. In the absence of the peptide, the same bands emerged rapidly. They could be assigned to the Cu(I) complex of GSH, reported previously by others. 33,34 This Cu II reduction phase lasted for about 9 h and reproducibly reached about 65% Cu II conversion at 25°C, as calculated from the intensity of the Cu II (Aβ 4−16 ) d−d band at 525 nm (Table 1). It was followed by the shorter reoxidation phase, which led to a practically full restoration of Cu II (Aβ 4−16 ). In the absence of Aβ 4−16 , the Cu II GSSG complex absorbing at 625 nm was the final reaction product ( Figure S1). It was not formed in the presence of Cu II (Aβ 4−16 ), because of the log K difference at pH 7.4 in favor of the latter, 10.37 vs 13.53. 15,35 The reaction rates increased with temperature ( Figure S5). The ESI-MS analysis of reaction products indicated the absence of covalent oxidative modification of Aβ 4−16 (Figures S6 and S7). The only change in its mass spectrum was due to partial detachment of bound Cu II ion resulting from its capture by GSH. The mass deficit of 2 Da, seen only in the copper-containing species, indicated the native ATCUN/NTS complex with two deprotonated, Cu IIbound amide nitrogens. 15 A transient spectral feature at 390− 405 nm accompanied the Cu I reoxidation phase. The same feature was present during the reoxidation phase of Cu II /GSH reaction in the absence of Aβ 4−16 ( Figure S8); hence it involves neither Aβ 4−16 nor its Cu II complex. A similar band was seen previously in a study of Cu I complexes in MT-3 and interpreted to originate from Cu I −Cu I interactions in the Cu 4 -thiolate cluster. 36 Indeed, Cu I preferentially forms a Cu 4 GSH 6 cluster at the molar excess of GSH. 33 However, the selective appearance of this low-energy band during oxidative decomposition of Cu 4 GSH 6 by molecular oxygen  Inorganic Chemistry pubs.acs.org/IC Communication suggests a contribution from partially oxidized species such as disulfide or Cu II . This issue will be investigated separately. The next series of reduction experiments was performed under the effectively anaerobic conditions, and indeed only the reduction phase of the reaction was observed during the 24 h incubation, leading to full Cu II reduction ( Figure 1B). Upon extending the incubation to 50 h, however, the reoxidation phase was observed after about 36 h of the incubation ( Figure  S9). This effect was due to ambient oxygen penetration of the samples residing in the spectrophotometer. The comparison of kinetic traces indicated the similarity of the early phase of the reduction process between the aerobic and anaerobic conditions ( Figure S10). These traces exhibited the mathematical form of first order kinetics for all conditions, only differing by the degree of Cu II reduction: ca. 65% under aerobic and nearly 100% under anaerobic conditions. However, as the actual reaction order was not determined, we compared the kinetics of individual reactions using initial velocities. The rate of Cu(II) reduction did not depend on the presence of ambient oxygen (Tables 1 and 2).  Table 2). The reoxidation phase occurred, however, similarly in all three cases (Figures 1 and 2). The reduction of CuAβ 1−16 under the same conditions was too fast for quantitation ( Figure  S11).
The kinetic, but not thermodynamic, resistance of Cu II Aβ 4−16 to reduction to Cu I species by thiols has been indicated in previous experiments. 32,37 Its kinetic character is reinforced by fast reduction of Cu II Aβ 12−16 , which has higher thermodynamic stability than Cu II Aβ 4−16 . 15,25 A clue for the mechanistic basis of this behavior is provided by the accelerating role of glutamic acid in both reductive copper transfer to MT-3 and nonreductive transfer to EDTA, along the affinity gradient. 38 This finding was interpreted in terms of assistance of copper transfer from the ATCUN/NTS motif via a putative partially coordinated intermediate species prone to form a ternary complex with a small transfer catalyst ligand. A similar mechanism was recently proposed in a study of Cu(II) reduction by GSH alone. 39 In the case of Cu II Aβ 12−16 , the fast Cu II reduction is most likely facilitated by the His residue in position two of the peptide chain, able to provide an alternatively coordinated, minor 3N species. 40,41 Such complexes are known to exchange Cu II ions rapidly. 42 They can also stabilize transient Cu I species. 38 Cu II Aβ 1−16 is known to facilitate Cu II reduction to Cu I by a number of mechanisms and is prone to ternary complex formation. 43,44 Summarizing, the results presented above indicate that Cu II Aβ 4−16 and especially Cu II Aβ 4−9 are sufficiently kinetically resistant to reduction by physiological concentrations of GSH to survive in the cell cytosol for hours without eliciting oxidative damage, while Cu II Aβ 12−16 and Cu II Aβ 1−16 do not have this ability. Therefore, Cu II Aβ 4−x complexes are good candidates to shuffle Cu II across the blood−brain barrier and in and out of the brain cells.
Experimental details, additional spectroscopic experiments, mass spectrometry data (PDF)