N-truncation of Aβ peptides and Cu(II) binding: affinity, structures, reactivity

© 2018 Dr. Mariusz MitalCompleted under a Cotutelle arrangement between the University of Melbourne and the Polish Academy of Sciences, Warsaw, PolandAlzheimer's disease is a neurodegenerative condition characterized by progressive cognitive decline and cerebral deposition of fibrillar plaques comprised of the β–amyloid (Aβ) peptide. According to the current state of knowledge, aggregation of Aβ is responsible for the onset and development of symptoms of AD. The first protein sequencing studies of the plaque core of AD patients identified a significant proportion of Aβ peptides with ragged N–termini, with Aβ4–x and Aβ12–x (x – different length of amino acid chain e.g. 16, 28, 40, 42) isoforms accounting for more than 60% and 8% of brain amyloid, respectively. This finding was largely overlooked until recently, when mass spectroscopic studies identified Aβ4–42 and Aβ1–42 as the dominant isoforms present in the hippocampus and cortex of sporadic AD patients, as well as in healthy brains. In the same papers, another minor form was identified – Aβ11–42. Unlike the commonly studied Aβ1–42 and Aβ1–40, all of these peptides contain the amino–terminal (H2N–Xaa–Yaa–His–) copper binding motif (ATCUN) characterised by the His residue at the third position from the N–terminus which enables high affinity Cu(II) binding. Moreover, Aβ12–x and Aβ11–x contain another His residue at the second or fourth position, respectively, which can affect their coordination properties. Additionally, Aβ11–x contains a glutamic acid residue at the first position, which can spontaneously cyclise into pyroglutamic acid. Aβ5–x is another N–terminally truncated form present in Aβ deposits in brains of sporadic and familial AD patients. It contains His at the second position from the N–terminus (Xaa–His–) which enables another type of high–affinity Cu(II) binding site. Studies presented by Lovell et al. in 1998 demonstrated increased levels of Cu, Fe and Zn ions in amyloid plaques ("Metal Ions Hypothesis"). However, these results were discordant with previous results reported by Deibel et al., where a significant decrease in copper in AD hippocampus and amygdala was reported. In addition, the meta–analysis presented by Schrag et al. showed no evidence for change in neocortical iron but a significant decrease in neocortical copper in AD in comparison with age–matched control tissue. Also, no clear evidence for copper imbalance has been presented using animal models. Therefore, the hypothesis that the total amount of Cu(II) ions present in the brains of AD patients and healthy persons is the same but the extracellular–intracellular copper ratio is increased, has gained more support. Cell–free studies have demonstrated that Cu(II) significantly accelerates Aβ aggregation. The Cu(II) interaction with Aβ peptides has also been linked with the production of free radicals in the presence of natural reductants like ascorbic acid (Asc) or hydrogen peroxide (H2O2). The metal hypothesis proposes that these processes contribute to disease progression and that “therapeutic chelation” could prevent it. Nowadays, it is thought that Aβ aggregation is not the only reason for the development of AD but also a balance between the production of Aβ and their degradation by enzymes is crucial. One of these degrading enzymes is neprilysin (NEP), a zinc–dependent endopeptidase associated with the membrane that cleaves peptides at the amino side of hydrophobic residues. A deep understanding of Cu(II) binding properties of N–truncated Aβ peptides vs. Aβ1–x, together with the knowledge how they interact with other Cu(II) binding partners, such as small molecule chelators, is required to assess their relevance to brain copper homeostasis in health and AD. Besides the Cu(II) coordination properties, it is important to understand the redox activity of these Cu(II) complexes in order to establish their putative role in the development of AD and whether chelators such as PBT2 can protect against the production of ROS. The relationship between various truncations, Cu(II) binding modes and enzymatic degradation of Aβ peptides may also be very important for designing therapies and understanding their outcome. This thesis aims to address many of these fundamental issues. Using spectroscopic methods (UV–Vis, CW–EPR and CD), the present work established the Cu(II) coordination properties of a series of model Aβ peptides (Aβ4–16; Aβ11–16; pAβ11–16; Aβ12–16; Aβ5–x, x = 9, 12, 16). CuAβ4–16 and CuAβ11–16 in the 1:1 peptide: metal ratio form four nitrogen (4N) Cu(II) complexes using their ATCUN motif, which is a high–affinity Cu(II) binding site. We have not observed any direct influence of the His at the fourth position present in Aβ11–16 on the Cu(II) coordination. The CuAβ12–16 complex at low pH (pH 3 – 4) is a 3N complex. At higher pH, it transforms into a 4N complex in the ATCUN motif (accounting for 100% of the Cu(II) speciation at pH ≥ 6.5). In addition, analogously to CuAβ12–16, we have shown that CuAβ5–x (x = 9, 12, 16) yield 3N complexes which are present from low to high pH, proving the important role of the His at the second position for Cu(II) binding in the absence of the ATCUN motif. Aβ5–x peptides are present in brains in low amounts, but are increased in the course of experimental AD treatments using protease inhibitors. We have also investigated the formation of 'internally–ternary' complexes between CuAβ5–x and the His or Tyr present in the same peptide chain. Based on the potentiometric technique, we have reported the binding affinities of these CuAβ complexes, placing them in the picomolar to femtomolar range. This is at least 3–orders of magnitude higher than the affinity of CuAβ1–x peptides considered so far to be key Cu(II) binding species. Further physico–chemical studies in this thesis concern mostly the interactions between Aβ4–16 and other Aβ peptides and a small molecule (a homologue of the experimental therapeutic PBT2). Based upon the results presented by Masters et al. and Portelius et al., we can propose that the 4–x N–terminally truncated Aβ peptides are very significant species in the brains of AD patients. PBT2 was reported to reduce neuropathological AD markers in transgenic animal models but had limited efficacy in human AD trials. The putative benefits of PBT2 were proposed to result from its ability to chelate and transport Cu(II), protecting against oxidative stress triggered by Cu/Aβ ROS production. Previous studies using the PBT2 homolog 2–[(dimethylamino)methyl]–8–hydroxyquinoline (L) showed that such 8–hydroxyquinolines can create a tridentate CuL complex with a conditional Kd value of 0.35 pM at pH 7.4. Moreover, L can form ternary copper complexes with His side chains of the Aβ1–16 peptide and other proteins. These observations suggest the ternary CuL(Aβ) complex is of relevance to any in vivo actions of PBT2 that involve Cu(II). To establish the relative affinity of Aβ4–16 to Cu(II) ions, competition experiments between Aβ4–16 and Aβ1–16 were performed using various spectroscopic methods (CW–EPR, UV–Vis and CD). In all cases, quantitative transfer of Cu(II) from Aβ1–16 to Aβ4–16 suggests the Cu(II) binding affinity of Aβ4–16 is more than an order of magnitude higher, which is consistent with potentiometric results. Given the relatively high binding affinity of Aβ4–16, it was anticipated that the Cu(II) transfer from CuL to Aβ4–16 would be immediate. However, the UV–Vis spectra showed a rapid exchange of Cu(II) from CuL to a ternary CuLAβ4–16 complex, which then slowly transformed into the CuAβ4–16 complex, taking approximately 90 min for complete Cu(II) transfer. Similar results were observed for a competition experiment involving Aβ5–9 and Aβ4–9 where transfer of Cu(II) ions from CuAβ5–9 to Aβ4–9 via a ternary Cu(Aβ5–9)NimAβ4-9 complex was complete only after approximately 23 hours. Using fluorescence spectroscopy, we conducted Cu(II) competition experiments between Aβ12–16 and Aβ4–16. We expected either quenching of the tyrosine fluorescence from Aβ4–16 by transfer of Cu(II) from CuAβ12–16 to Aβ4–16, or a increase of the tyrosine signal by transfer of Cu(II) from CuAβ4–16 to Aβ12–16. We observed both of these processes depending on which form was available in excess, indicating that both of these forms have very similar Cu(II) binding affinities. Moreover, Aβ5–16 and Aβ4–16 can bind a second equivalent of Cu(II) ions via His13/14 but the Cu(II) binding affinity for this site is low and thus its biological relevance doubtful. Next, the redox activity of copper complexes formed in the presence of Aβ4–16 and the PBT2 homolog (L) was examined. ROS production was measured using APF (2–[6–(4’–amino)phenoxy–3H–xanthen–3–on–9–yl]benzoic acid), a fluorescent sensor of hydroxyl radicals. We observed that CuAβ4–16 was redox silent but Cu2Aβ4–16 was redox active. A model system comprising Cu(II)/L/imidazole in a 1:1:20 ratio was used for further characterization of the redox activity of the ternary CuLNim complex where Nim represented His–bearing peptides and proteins. Previous CW–EPR and UV–Vis studies have shown that under these conditions > 95% of Cu(II) is bound by a ternary CuLNim species with the electronic and geometric structure similar to that of CuLNimAβ [11]. In this experiment, we observed ROS production. These findings have important consequences for experimental AD therapeutics based on the structure of L, such as PBT2, since it is highly likely that any Cu(II) intermediate of PBT2 will result in ROS production by the ternary Cu(PBT2)Nim complex. In collaboration with Magdalena Wiloch (Department of Microbioanalysis, Faculty of Chemistry, Warsaw University of Technology) cyclic and square wave voltammetry experiments were performed. These revealed that the 1:1 Cu(II)/Aβ4–16 complex undergoes irreversible oxidation to a Cu(III)/Aβ4–16 species at a potential above 1V. Also, using UV–Vis spectroscopy, the characteristic signal for Cu(III) complexes was observed in the presence of ascorbate (Asc). Using fluorescence spectroscopy and various forms of Aβ (Aβ1–16; Aβ4–x, x = 8, 12, 16), we have investigated the formation of dityrosine in the presence of Asc and H2O2. For CuAβ4–16, we observed a signal characteristic for dityrosine only in the presence of the Asc but not in the presence of H2O2. Also, no dityrosine signal was observed for the sample containing CuAβ4–12 under both conditions. This suggests a crucial role of the His13/14 and Asc in formation of dityrosine, thus indicating that a huge copper overload would have to occur for the CuAβ4–x species to yield this toxic modification. The last part of this project was related to proteolysis of various Aβ peptides (Aβ1–16, Aβ4–16, Aβ5–16) and their Cu(II) complexes by NEP. Using HPLC with TOF ESI–MS, we observed formation of Aβ4–9 and Aβ12–16 fragments from Aβ1–16 and Aβ4–16 in the presence of NEP. Both these fragments contain ATCUN motifs. Moreover, the results obtained for CuAβ4–16 in the presence of NEP revealed the approximately 3–times faster proteolysis of the peptide in comparison to CuAβ1–16 and CuAβ5–16. NEP activity in the presence of metal ions was also measured using fluorescence spectroscopy and the non–fluorescent substrate Mca–RPPGFSAFK(Dnp), which yields a fluorescent product upon NEP cleavage. Results obtained for different substrate and Cu(II) concentrations were fitted to a non–competitive inhibition model with Ki ~ 1 μM. Taking into account that peak extracellular Cu(II) in glutamatergic synapses may be as high as 200 – 300 μM, this result suggests that NEP may also be inhibited by Cu(II) in vivo. Addition of excess Zn(II) did not show restoration of the enzyme activity, but also inhibited NEP. Similar to Cu(II) experiments, fluorescence assays were performed with different substrate and Zn(II) concentrations and fitted to a non–competitive inhibition model with Ki ~ 20 μM