Selective Reversal of Cu-Amyloid Aggregation Monitored in Real Time by Fluorescence Anisotropy: Ni-Bme-Dach vs EDTA Benchmarks

Abstract

Metal dyshomeostasis, particularly involving Cu²⁺, is increasingly recognized as a key contributor to amyloid-β (Aβ) aggregation and neurotoxicity in Alzheimer’s disease, motivating the development of chelators capable of selectively disrupting pathogenic metal-Aβ interactions without perturbing essential biological metals. Here, we employ steady-state fluorescence anisotropy as a real-time probe of TAMRA-Aβ_1–42 rotational mobility to quantify metal-induced aggregation and its reversibility by two chelators with distinct selectivities: EDTA, a broad-spectrum benchmark, and Ni-bme-dach, a sulfur-rich metallodithiolate with high Cu affinity. Cu²⁺ induces the most significant increases in anisotropy, consistent with rapid formation of large nanoscale aggregates, while Fe³⁺ produces moderate aggregation and Zn²⁺ has minimal effect across pH 6.5 and 8.0. EDTA fully reverses Cu²⁺-induced aggregation but does so nonselectively, accompanied by pronounced fluorescence hyper-recovery indicative of broad metal stripping and fluorophore-environment perturbation. In contrast, Ni-bme-dach selectively extracts Cu²⁺, restoring monomer-like anisotropy at both pH values without hyper-recovery. UV–vis spectroscopy confirms formation of a discrete [Cu₂-(Ni-bme-dach)₃] complex, while TEM and AFM corroborate anisotropy trends and reveal a clear hierarchy of chelation responsiveness: Cu (fully reversible) > Fe (partially reversible) ≫ Zn (negligible). Together, these results establish fluorescence anisotropy as a sensitive kinetic platform for benchmarking chelator selectivity and demonstrate that Cu-driven Aβ aggregation is uniquely and selectively reversible. This work highlights metal-specific reversibility as a critical design principle for next-generation, Cu-targeted chelation strategies in Alzheimer’s disease.

Introduction

Alzheimer’s disease (AD) remains without effective disease-modifying therapies, despite decades of effort to target amyloid-β (Aβ) production, fibrillization, or clearance. A growing consensus recognizes that metal dysregulation, particularly involving Cu²⁺, Fe³⁺/Fe²⁺, and Zn²⁺, constitutes a central but underexploited therapeutic axis in AD. Elevated Cu and Zn concentrations (tens to hundreds of micromolar) have been reported in amyloid plaques and peri-plaque regions, − where they accelerate Aβ aggregation, alter oligomer structure, promote oxidative stress, and amplify neurotoxicity. ,− While Cu²⁺ is especially pathogenic, both Cu^I/II or Fe^III/II can undergo redox cycling in the presence of a reducing agent, generating reactive oxygen species (ROS) that impair synaptic signaling and lead to neuronal death. −

Chelation studies linking metal homeostasis to Aβ_1–42 aggregation and oxidative damage suggest that redissolving Aβ deposits may improve clinical outcomes in patients with AD. − Consequently, there is intense interest in metal-modulating therapeutics, including Fe and Cu-binding ligands such as clioquinol, PBT2, and deferiprone. − While these agents demonstrate some ability to redistribute metals and attenuate Aβ toxicity, ,, they often share a fundamental limitation: they are broad-spectrum chelators or have high affinities for multiple metals. For example, ethylenediamine tetraacetic acid (EDTA) (Scheme ) has a high affinity for Cu, Zn, and Fe without discrimination, complicating mechanistic interpretation and risking disruption of metal-dependent neurobiology, including Zn-regulated synaptic function and Fe-dependent metabolic pathways. , This lack of selectivity has hindered the development of therapeutics that precisely target pathogenic metal-Aβ interactions without destabilizing essential metalloprotein systems.

1. Chelator Ligands for Metal Sequestration.

Another critical challenge is that the distinct aggregation pathways induced by Cu²⁺, Fe³⁺, and Zn²⁺, and their reversibility, remain poorly defined. Classical amyloid assays, such as Thioflavin T and Congo Red, primarily detect β-sheet-rich fibrils and do not report on early oligomers or subtle conformational changes induced by metal binding. − Consequently, it is unknown whether all metal-induced Aβ aggregates are equally accessible to chelation, or whether specific metals create reversible vs irreversible aggregation trajectories. Establishing this distinction is fundamental for designing selective chelators that disrupt only pathogenic metalated-Aβ species. To overcome these limitations, sensitive, real-time methods are needed to monitor metal-induced aggregation at low peptide concentrations. Fluorescence anisotropy provides this capability: by reporting peptide rotational mobility and hydrodynamic radius with nanosecond resolution, it can distinguish among monomers, small oligomers, and large supramolecular assemblies, an ability not afforded by intensity-only or fibril-specific assays. − Fluorescence anisotropy has previously been applied to monitor Aβ aggregation using long-lifetime transition-metal probes, particularly Ru²⁺ complexes, , and its physical basis is well established for amyloid-binding organic dyes such as Thioflavin T and Congo Red, − whose binding-induced rotational restriction underlies changes in fluorescence polarization. Such studies demonstrated that rotational constraints increase upon oligomer formation and that excited-state lifetimes strongly influence the accessible correlation-time regime. Fluorophore labeled-Aβ_1–42 is particularly advantageous, as the fluorophore preserves native metal binding while sensitively reporting structural changes. − N-terminal TAMRA labeling provides a convenient optical handle while preserving the metal-binding domain of Aβ. Studies of Aβ_1–42 peptides conjugated to TAMRA at the N-terminal aspartate residue have shown that this labeling does not disrupt intrinsic aggregation pathways or metal-binding behavior. While minor interactions between TAMRA and the nearby Tyr10 residue have been shown to modulate fluorescence. The label remains sufficiently distant from key coordination residues and effectively tracks metal-induced conformational transitions, oligomerization, and aggregation. Still, as with any fluorophore modification, subtle effects on aggregation behavior are possible, and interpretation of TAMRA-based readouts is within this context. Building on this foundation, the present work used anisotropy not as a standalone sizing tool but as part of an integrated platform that directly compared how Cu²⁺, Fe³⁺, and Zn²⁺ induced nanoscale aggregates respond to broad-spectrum versus Cu-selective chelation. When combined with UV–vis spectroscopy and nanoscale imaging (transmission electron microscopy) (TEM) and atomic force microscopy (AFM), this platform offers a comprehensive, multilength-scale view of metal-specific aggregation and chelation response.

Within a therapeutic context, selective Cu extraction represents a compelling strategy for neutralizing pathogenic Cu–Aβ complexes while avoiding the biological costs associated with broad metal removal. Metallodithiolate ligands such as nickel-N,N′-bis-2-mercaptoethyl-N,N′-diazacycloheptane (Ni-bme-dach) (Scheme ) embody this approach. Inspired by N₂S₂ motifs in metalloenzymes such as acetyl-CoA synthase, − Ni-bme-dach offers a biologically inspired alternative with inherent selectivity for soft, thiophilic metal centers, particularly Cu⁺/Cu²⁺. − The N₂S₂ motif closely mimics metal-binding sites in enzymes such as acetyl-CoA synthase, and its sulfur-rich donor set preferentially stabilizes Cu over harder Lewis-acidic metals like Fe³⁺ or over biologically abundant cations such as Ca²⁺ and Mg²⁺. , Formation constants further emphasize these distinctions: EDTA exhibits log K values of 19.9 for Cu²⁺ and 16.4 for Zn²⁺, whereas Ni-bme-dach forms even stronger complexes with Cu (log K > 30) but much weaker complexes with Zn (log K ≈ 2). These thermodynamic preferences indicate that Ni-bme-dach is well positioned to selectively extract Cu²⁺ from metalated Aβ_1–42 while avoiding essential metals and minimizing off-target chelation. Consequently, Ni-bme-dach is expected to outperform EDTA as a selective and mechanistically precise capture agent capable of reversing Cu-mediated Aβ aggregation without perturbing normal metal homeostasis.

Herein, this study investigates how Cu²⁺, Fe³⁺, and Zn²⁺ differentially trigger Aβ_1–42 aggregation and whether the resulting assemblies are structurally and mechanistically reversible. Experiments were conducted at pH 6.5 and 8.0 to model mildly acidic endosomal and near-physiological extracellular conditions commonly used in metal-Aβ studies. The novelty lies in a direct comparison between a broad-spectrum chelator (EDTA) and a Cu-selective metallodithiolate (Ni-bme-dach) using an integrated platform of fluorescence anisotropy, fluorescence intensity, UV–vis spectroscopy, TEM, and AFM. This approach addresses key questions: Which metals drive distinct aggregation pathways? Are these aggregates equally susceptible to chelation-driven disassembly? Can Ni-bme-dach selectively extract Cu compared to EDTA’s nonselective behavior? How do nanoscale morphological changes correlate with real-time optical signatures of metal binding and aggregation? By focusing on early-stage, metal-induced Aβ_1–42 aggregation under controlled in vitro conditions, this work establishes a mechanistic framework for evaluating metal-specific aggregation and chelation. Integrated analyses reveal a clear hierarchy of chelation response: complete reversibility for Cu-induced aggregates, partial for Fe, and minimal for Zn, a level of detail not previously mapped. Unlike fibril-binding dyes such as ThT, steady-state fluorescence anisotropy using TAMRA-Aβ_1–42 reports real-time changes in hydrodynamic radius, enabling detection of early stage, metal-specific oligomerization events that precede β-sheet formation.

Experimental Procedure

Materials

Hexafluoroisopropanol (HFIP) and EDTA were purchased from Sigma-Aldrich. Sodium phosphate monobasic monohydrate, sodium phosphate dibasic heptahydrate, and copper sulfate were from BDH Chemicals. Ferrous chloride and ferric chloride were from Fisher Scientific, while zinc bromide was from Matheson Coleman and Bell. 4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) was from OmniPur. The metallodithiolate (NiN₂S₂) complex, Ni-bme-dach, was prepared according to published procedures and was provided by Marcetta Y. Darensbourg of Texas A&M University. TAMRA-labeled-Aβ_1–42 was purchased from CPC Scientific as a lyophilized powder and stored at −20 °C in HFIP. TAMRA-Aβ_1–42 was prepared with a trifluoroacetate counterion and modified at the N-terminal end of the aspartic amino acid residue of the Aβ_1–42 sequence: Asp–Ala–Glu–Phe–Arg–His–Asp–Ser–Gly–Tyr–Glu–Val–His–His–Gln–Lys–Leu–Val–Phe–Phe–Ala–Glu–Asp–Val–Gly–Ser–Asn–Lys–Gly–Ala–Ile–Ile–Gly–Leu–Met–Val–Gly–Gly–Val–Val–Ile-Ala – OH. All reagents were used as received. All experiments employed 100% TAMRA-labeled Aβ_1–42, with no unlabeled peptide added.

Preparation of TAMRA-Aβ_1–42

A 0.05 mg of lyophilized TAMRA-Aβ_1–42 was suspended in 500 μL anhydrous HFIP and sonicated for 1 min to disperse any initial aggregates, then further diluted with an additional 500 μL HFIP. HFIP is a fluorinated alcohol that is miscible with water and well-suited for dissolving aggregation-prone peptides. The concentration of this solution was determined with UV–vis using the absorbance of the TAMRA dye at 545 nm and molar absorptivity of 90,000 cm^–1 M^–1. Stock solutions were stored in 200 mL aliquots at −20 °C to minimize freeze–thaw cycles of the initial stock. For all subsequent studies, stocks were diluted with HFIP to a final concentration of 5.52 μM before use and were stored at −20 °C.

Metal Binding, Aggregation, and Chelation Studies with TAMRA-Aβ_1–42

For studies using TAMRA-Aβ_1–42, fluorescence intensity and anisotropy data were collected at multiple time points and pH values. For metal-binding and chelation studies with EDTA, 10 mL of 5.52 μM TAMRA-Aβ_1–42 was added to 1000 μL of 10 mM HEPES buffer at pH 6.5 or 8.0, yielding a final concentration of 55 nM. This is followed by the addition of 2 μL of 33 mM of CuSO₄, FeCl₃, or ZnBr₂ in DI water, a 10 min incubation period, and stirring before data collection. Lastly, EDTA (5 μL of 0.5 M in DI water) was incubated with the metalated-TAMRA-Aβ_1–42 sample for 5 min prior to data collection. For studies using the Ni-bme-dach, 20 mL of 5.52 μM TAMRA-Aβ_1–42 was added to 1000 μL of 10 mM HEPES buffer at pH 6.5 or 8.0, yielding a final concentration of 110 nM. This is followed by the addition of 1 μL of 33 mM CuSO₄, FeCl₃, or ZnBr₂ in DI water, with a 10 min incubation period and stirring. Lastly, Ni-bme-dach (20 μL of 0.018 M in DI water) was incubated with the metalated-TAMRA-Aβ_1–42 sample for 5 min before data collection. Fluorescence intensity counts and anisotropy data were recorded at each sequential step of metal ion or chelator additions to the TAMRA-Aβ_1–42 described above. Triplicate studies were performed, and data were expressed as means and standard deviations. Sample preparation was the same for the UV–vis metal binding and chelation studies with Ni-bme-dach and EDTA. Control samples with TAMRA-Aβ_1–42 only, Ni-bme-dach or EDTA with each metal species, and metal ions only were compared at similar concentrations described here.

Fluorescence and UV–Vis Spectroscopy

Fluorescence measurements were performed on a HORIBA Fluorolog-QM PTI spectrophotometer using Felix32 software. Fluorescence intensity and anisotropy were measured using a 1.0 cm path length quartz cell with an excitation wavelength of 545 nm and an emission range of 550–650 nm. For fluorescence anisotropy measurements, emission was collected at 600 nm with a slith width of 6 nm. This wavelength was selected because it is near the λ_max of TAMRA emission; however, it was shifted to reduce interference from scattering. A SCHOTT Colored Glass Longpass Filter with a cut-on wavelength of 570 nm was placed in the sample compartment in front of the emission detector to filter out wavelengths below 570 nm and ensure that polarized light from the excitation monochromator did not reach the detector. All anisotropy measurements were collected in steady-state mode. Because time-resolved anisotropy decays were not acquired, rotational correlation times and hydrodynamic radii are not extracted from these data. Instead, anisotropy is used as a relative indicator of changes in rotational mobility accompanying aggregation or chelation. The steady-state anisotropy (r) was derived from eq

$$r=\frac{I_{\mathrm{V}\mathrm{V}}-{GI}_{\mathrm{V}\mathrm{H}}}{I_{\mathrm{V}\mathrm{V}}+{2GI}_{\mathrm{V}\mathrm{H}}}$$ 1

where I _VV and I _VH are the fluorescence intensities, the subscripts indicate the orientation of the excitation and emission polarizers, and G = I _HV/I _HH is the wavelength-dependent sensitivity of the instrument. Reported delta anisotropy (Δr) values are the average of three independent samples; each averaged over 60 s unless otherwise stated. There was a 10 min incubation period after the addition of the metal and chelator before data were collected. For anisotropy studies, Absorbance measurements were obtained using an Agilent Technologies Cary 5000 UV–Vis–NIR Spectrophotometer with a 1.0 cm path length quartz cuvette. These measurements capture early stage aggregation behavior, as metal-treated samples were incubated for approximately 10 min at room temperature.

TEM Studies of Aβ-Aggregation in the Presence and Absence of Metal Ions and Chelators

The following samples were prepared for TEM imaging: TAMRA-Aβ_1–42 only, TAMRA-Aβ_1–42 in the presence of metal ions, and TAMRA-Aβ_1–42 in the presence of metal ions and EDTA or Ni-bme-dach. Briefly, 10 μL of 5.52 μM TAMRA-Aβ_1–42 in HFIP was added to 100 μL of 10 mM PB (pH 8.0) to yield a final concentration of 0.55 mM, which was then deposited onto a TEM grid. For metal-binding studies, 2 μL of 33 mM CuSO₄, FeCl₃ or ZnBr₂ in H₂O was added to 10 μL of 5.52 μM TAMRA-Aβ_1–42 in HFIP diluted with 100 μL of 10 mM PB pH 8.0 followed by incubation for 10 min before deposition onto a TEM grid. For chelation studies, the metal-TAMRA-Aβ_1–42 sample, as described above, was incubated with 5 μL of 0.5 M EDTA in H₂O for 10 min before deposition onto a TEM grid. For studies with Ni-bme-dach, samples were prepared similarly for deposition except with 20 μL of 5.52 μM Aβ_1–42 in 100 μL of 10 mM PB pH 8.0 to yield a final concentration of 1.1 mM, with additions of freshly prepared 1 μL of 33 mM CuSO₄, FeCl₃, or ZnBr₂, and 20 μL 0.018 M Ni-bme-dach in H₂O. Each sample was drop-cast onto carbon-coated (300 Å) Formvar films on copper grids (Ted Pella) and air-dried overnight before imaging. Samples were stained with 2% uranyl acetate, and images were obtained of the samples after they were dried overnight. TEM imaging was acquired on a Tecnai F-20 FEI microscope using a CCD detector at an acceleration voltage of 200 kV.

AFM Studies of Aβ-Aggregation in the Presence and Absence of Metal Ions and Chelators

The following samples were prepared for AFM imaging: TAMRA-Aβ_1–42 only, TAMRA-Aβ_1–42 in the presence of metal ions, and TAMRA-Aβ_1–42 in the presence of metal ions and EDTA or Ni-bme-dach. Briefly, 10 μL of 10.3 μM TAMRA-Aβ_1–42 was dried into a film in an Eppendorf vial to remove HFIP, resuspended in 100 μL of 10 mM PB, pH 8.0, and sonicated for 30 s to disperse any aggregates formed during drying. For the TAMRA-Aβ_1–42-only sample, 5 μL of a 103 nM solution in 10 mM PB (pH 8.0) was deposited onto freshly cleaved mica, dried under a clean stream of N₂, rinsed with nanopure water, and dried again under N₂. To study the change in aggregate size upon metal binding, the remaining 95 μL of 103 nM TAMRA-Aβ_1–42 solution was incubated with 1 μL of 33 mM CuSO₄, FeCl₃, or ZnBr₂ in H₂O for 15 min, then deposited and dried onto a fresh mica surface in the same manner. To study the effect of chelation, 5 μL of 0.5 M EDTA in H₂O or 20 μL of 0.018 M Ni-bme-dach in H₂O was added to the resuspended TAMRA-Aβ_1–42 and metal solution, incubated for 5 min, and then 5 μL of this solution was deposited onto and dried onto another freshly cleaved mica surface. Imaging was performed using a Digital Instruments Veeco AFM/LFM Instrument (Veeco Metrology group). Rotated monolithic, uncoated silicon AFM probes with a 125 μm tip length, a 300 kHz resonant frequency, and a 40 N/m spring constant (model Tap300G, Ted Pella) were used. The machine was operated in tapping mode, with the drive amplitude kept to a minimum, collecting scans of 2 μm × 2 μm at a slow scan frequency of 1.5–2.5 Hz and 512 × 512 scans per line. AFM images were analyzed using Digital Instruments software. TAMRA-Aβ_1–42 peptide size was measured using cross-section analysis to create a size distribution histogram for each image.

Statistical Analysis

Fluorescence anisotropy was measured in triplicate for Aβ-only controls, metal-treated samples (Cu²⁺, Fe³⁺, Zn²⁺), and those treated with EDTA or Ni-bme-dach at pH 6.5 and 8.0 to assess pH-dependent effects. Baseline anisotropy (r _Aβ) was calculated from Aβ controls, and Δr values were derived as Δr = r _condition – r _Aβ for group comparisons. One-way ANOVA assessed differences among Aβ, metal, and metal + chelator groups, with Levene’s test confirming variance homogeneity (p > 0.69). Bonferroni-corrected pairwise t tests, followed by significant effects (p < 0.05), and effect sizes were reported as partial eta squared (η²). EDTA results were used to evaluate the reversal of Cu-, Fe-, and Zn-induced anisotropy shifts, whereas Ni-bme-dach tests assessed chelator selectivity. Both chelators exhibited robust Cu²⁺-specific reversal, partial Fe²⁺-reversal (pH-dependent), and negligible effects on Zn²⁺. Statistical analyses were performed in Python (NumPy, SciPy) with a significance level of p < 0.05. A statistical workflow is in the Supporting Information.

Results and Discussion

The following sections systematically build a metal-selectivity hierarchy by integrating fluorescence intensity, anisotropy, UV–vis spectroscopy, and nanoscale imaging. Each technique contributes a distinct dimension to understanding Cu-, Fe-, and Zn-specific aggregation and chelation.

Cu²⁺ is the Dominant Driver of Aβ Aggregation Across pH Conditions

Fluorescence intensity was first used to evaluate the effects of Cu²⁺, Fe³⁺, and Zn²⁺ on the TAMRA-Aβ_1–42 behavior, thereby establishing a baseline for subsequent fluorescence anisotropy studies as a real-time probe of aggregation. All experiments employed 100% TAMRA-Aβ_1–42, in which the metal-binding region is preserved to enable sensitive anisotropy detection, consistent with prior reports. , Intensity measurements confirmed that Cu²⁺ is the dominant driver of TAMRA-Aβ_1–42 aggregation under both pH conditions, providing the foundation for anisotropy studies. For fluorescence intensity studies, TAMRA-Aβ_1–42 was diluted to 55 nM in 10 mM HEPES buffer, excited at 545 nm, and emission spectra collected from 550 to 650 nm using a 570 nm long-pass filter to minimize scattering. Concentrations were maintained at 55 nM to prevent TAMRA self-quenching, and control experiments showed minimal fluorescence drift in TAMRA-Aβ_1–42 alone, thereby confirming that the observed changes were metal- or chelator-induced rather than due to fluorophore interactions. Stock samples were stored in HFIP to maintain a stable, near-monomeric state. Metals were introduced by adding 2 μL of 33 mM CuSO₄, FeCl₃, or ZnBr₂ for a final concentration of 65 μM and a peptide/metal ratio of ≈1:1200, followed by incubation for 10 min with gentle stirring before data collection. While elevated Cu²⁺ and Zn²⁺ concentrations (tens to hundreds of micromolar) in amyloid plaques and peri-plaque regions, − , the metal/Aβ ratios employed here reflect conditions typical of plaque-associated environments.

In the representative spectra (Figure A), addition of Cu²⁺ produced a substantial quenching of TAMRA emission, 48% ± 12% at pH 8.0 (Figure A (a) and (b)) and 54% ± 24% at pH 6.5 (Figures B and S1). No similar quenching occurs when TAMRA-Aβ_1–42 is incubated with an equivalent buffer for 5 min (Figure S2). This rapid change contrasts with the gradual fluorescence decay observed without metal ions, which decreased by 52 and 74% after 24 and 72 h, respectively (Figure S3). Fe³⁺ induced a similar magnitude of quenching (42% ± 29% at pH 8.0; 55% ± 12% at pH 6.5) (Figures B and S1), albeit with higher variability due to its heterogeneous coordination chemistry and redox activity. In contrast, Zn²⁺ produced only modest decreases (12% ± 12% at pH 8.0; 8.5% ± 4.1% at pH 6.5, Figures B and S1), consistent with weaker or more localized structural perturbation of the peptide.

1.

(A) Representative fluorescence spectra of (a) TAMRA-Aβ_1–42, (b) after the addition of CuSO₄, and (c) EDTA in 10 mM HEPES buffer pH 8.0. (B) Comparison of the % change (Δ) in fluorescence intensity (FLI) counts from monomeric TAMRA-Aβ_1–42 and in the presence of CuSO₄, FeCl₃, and ZnBr₂ and EDTA in 10 mM HEPES buffer pH 8.0 and pH 6.5. Final concentrations of the TAMRA-Aβ_1–42, metals, and EDTA are 55 nM, 65 μM, and 2.5 mM respectively. Bars represent mean ± SD (n = 3). Statistical significance was assessed using one-way ANOVA with Tukey’s post hoc test.

To confirm that fluorescence changes were due to metal interactions with TAMRA-Aβ_1–42, EDTA was used to remove Cu²⁺, Fe³⁺, or Zn²⁺ from the peptide. After adding EDTA (5 μL of 0.5 M, final concentration 2.5 mM; metal/chelator ratio 1:38), fluorescence intensity increased significantly (Figure A­(c)). Cu²⁺-Aβ_1–42 complexes showed the strongest recovery, 3558% at pH 8.0 and 2065% at pH 6.5, far surpassing the premetal baseline (Figure A). Zn²⁺ exhibited a moderate return (727% at pH 8.0; 174% at pH 6.5), whereas Fe³⁺ produced limited recovery (288% at pH 8.0; 58% at pH 6.5), consistent with incomplete Fe³⁺ chelation under mildly acidic conditions (Figures B and S1). Notably, EDTA addition to TAMRA-Aβ_1–42, even in the absence of metals, resulted in a substantial increase in fluorescence intensity or “hyper-recovery” (Figure S4). The significant increase in fluorescence intensity upon addition of EDTA is surprising and may reflect multiple mechanistic contributions. First, EDTA rapidly strips metals from their primary coordination sites on TAMRA-Aβ_1–42, thereby reversing the metal-induced conformational transitions that promote TAMRA quenching. Second, EDTA disrupts transient interactions between the TAMRA fluorophore and the nearby Tyr10 residue, an established quenching pathway, thereby generating a pronounced “quenching-release” or fluorescence “hyper-recovery” effect. , Third, EDTA relaxes peptide–peptide contacts, restoring a more extended, monomer-like conformation that is intrinsically more emissive. Fourth, trace counterions may remain after peptide purification and are chelated by EDTA, as the product sheet reports 95% purity by HPLC. However, the vendor (Anaspec.com) reports that no trace metals are present in the purchased peptide stock.

Statistical analysis of percent change in fluorescence confirms strong metal- and pH-dependent effects. Cu²⁺ produced the most pronounced response at both pH values, with a highly significant overall effect at pH 8.0 (F(2,6) = 39.41, p = 0.0004, η² = 0.93) and all pairwise comparisons remaining significant after correction (p _adj ≤0.010), consistent with robust quenching and EDTA-driven hyper-recovery. At pH 6.5, Cu²⁺ again showed a significant group effect (F(2,6) = 11.21, p = 0.0094, η² = 0.79), though pairwise comparisons did not survive correction, reflecting increased variability while preserving trend direction. Fe³⁺ displayed a significant effect at pH 8.0 (F(2,6) = 13.65, p = 0.0058, η² = 0.82) without significant corrected pairwise differences; while at pH 6.5, a borderline ANOVA (F(2,6) = 4.22, p = 0.0717) and selective significance between Fe³⁺ and TAMRA-Aβ_1–42 (p _adj = 0.0036) indicate partial EDTA-mediated recovery. Zn²⁺ induced more modest effects, with a significant but weakly resolved response at pH 8.0 (F(2,6) = 5.95, p = 0.0377, η² = 0.66) and a stronger, EDTA-responsive effect at pH 6.5 (F(2,6) = 21.33, p = 0.0019, η² = 0.88; p _adj ≈ 0.03).

The fluorescence quenching and recovery observed in the presence of Cu²⁺ are consistent with prior ¹H and ¹⁵N NMR studies, which show that Cu²⁺ binding to the Aβ monomer backbone induces oligomerization and that this process is fully reversible upon EDTA treatment. , The differential recovery observed across metals further highlights distinct interaction modes. Zn²⁺’s modest quenching and recovery are consistent with 2D and 3D ¹⁵N NMR evidence showing weaker, more localized binding involving His6, His11, His13, Asp1, and residues near Asp23–Lys28. In contrast, Fe³⁺’s limited recovery reflects its more heterogeneous and pH-sensitive coordination environment, involving carboxylate residues including Glu3, Glu11, Glu22, Asp1, and Asp23. − Together, these metal-dependent fluorescence signatures confirm that each ion interacts distinctly with TAMRA-Aβ_1–42, with Cu²⁺ producing the most pronounced conformational and aggregation-related perturbations.

Fluorescence Anisotropy Enables Real-Time Monitoring of Metal-Induced Aggregation

Because fluorescence intensity studies showed rapid, metal-specific conformational changes in TAMRA-Aβ_1–42; steady-state fluorescence anisotropy was used to measure changes in hydrodynamic radius from early oligomer formation, with EDTA as a benchmark for reversibility. These measurements focused on short-time aggregation equilibria, with samples incubated for approximately 10 min at room temperature, in contrast to the longer times required for fibril maturation at 37 °C. Briefly, 55 nM TAMRA-Aβ_1–42 in 10 mM HEPES buffer was excited at λ_max 545 nm, and anisotropy was monitored at λ_max 600 nm. After establishing a baseline, 65 μM CuSO₄, FeCl₃, or ZnBr₂ (peptide/metal ≈1:1200) was added, and samples were gently stirred and incubated for 10 min before data acquisition for T = 1 min. Cu²⁺ produced the most significant anisotropy increase, rising from 0.12 ± 0.01 to 0.20 ± 0.01 at pH 8.0 (Δr = 0.09 ± 0.01) (Figure A). At pH 6.5, Cu²⁺ also promoted aggregation, though to a lesser extent (Δr = 0.03 ± 0.02), reflecting the known pH dependence of Cu-mediated Aβ coordination (Figure C). Fe³⁺ generated moderate anisotropy increases (Δr _avg = 0.03 ± 0.01 at pH 8.0; 0.03 ± 0.02 at pH 6.5) (Figure B,C), consistent with formation of smaller, less compact aggregates, whereas Zn²⁺ produced negligible changes (−0.01 ± 0.02 at pH 8.0; −0.01 ± 0.01 at pH 6.5), reflecting its weak aggregation propensity. The relative magnitude of Δr correlates with known coordination modes: Cu²⁺ binding to His6/His13/His14 yields large, rigid assemblies; Fe³⁺ interacts with Tyr10 and carboxylates to form smaller structures; − and Zn²⁺ coordinates histidines and carboxylates without promoting aggregation. Importantly, buffer-only controls produced no anisotropy change (Figure A, orange trace), confirming that Δr reflects metal coordination rather than spontaneous self-association. Addition of EDTA returned anisotropy values close to the metal-free baseline; however, accompanying fluorescence data indicate that EDTA also perturbs the local fluorophore environment beyond metal removal. These results verify that anisotropy changes originate from metal-induced aggregation with EDTA as an internal control for reversibility.

2.

(A) A representative fluorescence anisotropy plot of (a) TAMRA-Aβ_1–42 in 10 mM HEPES pH 8.0 with sequential additions of (b) 2 mL of CuSO₄ and (c) 5 mL EDTA (blue) and TAMRA-Aβ_1–42 with (a) first eq 2 mL and (b) second eq 5 mL of 10 mM HEPES pH 8.0 (orange). Note: that after each addition, there is a 10 min incubation period before data is collected. Comparison of the average anisotropies of TAMRA-Aβ_1–42 in 10 mM HEPES, (B) pH 8.0, or (C) pH 6.5, before and after the addition of metal and EDTA. The final concentrations of TAMRA-Aβ_1–42, metal ions, and EDTA are 55 nM, 65 μM, and 2.5 mM. Bars represent mean ± SD (n = 3). Statistical significance was evaluated using one-way ANOVA with Tukey post hoc corrections.

The observed anisotropy responses are physically consistent with the short excited-state lifetime of the TAMRA fluorophore (≈2–3 ns), which limits steady-state anisotropy sensitivity to species undergoing rapid rotational diffusion. Consequently, anisotropy primarily reports on early stage oligomers and local rigidity changes rather than large, slowly rotating aggregates whose rotational correlation times exceed the fluorescence lifetime. Within this framework, statistical analysis confirms a strong metal-dependent hierarchy in anisotropy response. Cu²⁺ produced highly significant differences among TAMRA-Aβ_1–42, Cu²⁺, and Cu²⁺ + EDTA at both pH values (pH 6.5: F(2,6) = 24.56, p = 0.0013, η² = 0.89; pH 8.0: F(2,6) = 315.34, p ≈ 10^–6, η² = 0.99), with post hoc comparisons confirming complete reversal of Cu-induced anisotropy upon EDTA addition (p ≤ 0.024). Fe³⁺ exhibited more minor but significant group effects (pH 6.5: F(2,6) = 16.86, p = 0.0034; pH 8.0: F(2,6) = 7.89, p = 0.0209), consistent with partial anisotropy reversal at pH 6.5 and minimal response at pH 8.0. In contrast, Zn²⁺ produced anisotropy values near baseline at both pH conditions, with no statistically significant group differences observed. While abrupt anisotropy increases can, in some systems, signal liquid–liquid phase separation, the absence of micron-scale droplets and the low peptide concentrations used here indicate that the observed responses arise from nanoscale aggregation. Together, these results establish a clear hierarchy in early stage aggregation sensitivity (Cu²⁺ > Fe³⁺ > Zn²⁺) and demonstrate that EDTA restores anisotropy to baseline across all metal conditions (Figure A–C). ,

Ni-bme-dach Selectively Reverses Cu-Induced Aggregation Without Broad Metal Stripping

Ni-bme-dach, was next evaluated as a targeted chelator to determine whether it could selectively reverse Cu-induced Aβ_1–42 aggregation without the broad, nondiscriminating metal removal characteristic of EDTA. To assess this selectivity, 20 μL of 0.02 M Ni-bme-dach was added to TAMRA-Aβ_1–42 samples containing 110 nM peptide and 0.33 mM metal ions (metal: peptide = 1:300, metal: chelator = 1:11). Although the metal/peptide ratios differ between EDTA and Ni-bme-dach conditions, both sets were intentionally designed to maintain excess metal and chelator to ensure complete competitive binding under their respective regimes. Fluorescence anisotropy was measured before and after the addition of Ni-bme-dach at pH 8.0 and 6.5 to monitor metal coordination and chelation. For Cu²⁺, anisotropy measurements showed that metal binding induced moderate aggregation (Δr _avg = 0.02 ± 0.01 at pH 8.0; 0.03 ± 0.01 at pH 6.5) (Figure A,B). Strikingly, the addition of Ni-bme-dach completely reversed the increases in anisotropy at both pH values, restoring monomer-like rotational mobility and demonstrating complete disassembly of Cu–Aβ aggregates. This complete reversal confirms that Ni-bme-dach effectively disrupts Cu–Aβ interactions. The extent of anisotropy recovery parallels the Cu-specific reversibility observed with EDTA, yet, as described below, the underlying mechanism differs significantly. In addition, a slightly smaller recovery at pH 6.5 is consistent with decreased stability of the Cu-histidine coordination under mildly acidic conditions. Statistical analysis of the Δr values confirmed the Cu-specific nature of Ni-bme-dach reversal. At both pH 6.5 and 8.0, one-way ANOVA revealed significant differences among TAMRA-Aβ_1–42, Cu²⁺-treated, and Cu²⁺ + Ni-bme-dach samples (pH 6.5: F(2,6) = 18.40, p = 0.0028; pH 8.0: F(2,6) = 11.95, p = 0.0081). Post hoc comparisons demonstrated that Cu²⁺ produced a significantly larger Δr than both TAMRA-Aβ_1–42 (p = 0.018 at pH 6.5) and Cu²⁺ + Ni-bme-dach (p = 0.013 at pH 6.5; p = 0.039 at pH 8.0), whereas the Ni-bme-dach-treated samples were statistically indistinguishable from baseline at both pH values (p ≥ 0.85). These results corroborate the complete functional reversal of Cu-induced aggregation by Ni-bme-dach and reinforce that the chelator restores peptide mobility to monomer-like levels.

3.

Comparison of the average anisotropies of TAMRA-Aβ_1–42 in (A) 10 mM HEPES pH 8.0 or (B) 10 mM HEPES pH 6.5, and after the subsequent addition of metal ions and Ni-bme-dach. The final concentrations of TAMRA-Aβ_1–42, metal ions, and Ni-bme-dach were 110 nM, 33 μM, and 0.35 mM, respectively. Bars represent mean ± SD (n = 3). Statistical significance was evaluated using one-way ANOVA with Tukey post hoc corrections.

In contrast to the robust Cu-specific reversibility, Ni-bme-dach produced substantially weaker effects on Fe- and Zn-induced aggregation across both pH values. At pH 8.0, Fe³⁺ caused only minimal aggregation (Δr _avg = −0.01 ± 0.01) (Figure B), consistent with the extremely low solubility of Fe³⁺ under alkaline conditions where Fe­(OH)₃ precipitates (K _sp = 1.0 × 10^–38). Correspondingly, the addition of Ni-bme-dach produced no measurable change in anisotropy (Δr _avg ≈ 0.00 ± 0.01). Although one-way ANOVA indicated a significant overall effect among TAMRA-Aβ_1–42, Fe³⁺, and Fe³⁺ + Ni-bme-dach (F(2,6) = 7.89, p = 0.0209, η² = 0.72), post hoc comparisons showed only a borderline difference between Fe³⁺ and TAMRA-Aβ_1–42 (p = 0.050) and no significant difference between Fe³⁺ and Fe³⁺ + Ni-bme-dach (p = 1.0). Thus, at pH 8.0, the minimal magnitude of Δr changes reflects limited soluble Fe³⁺ available to interact with TAMRA-Aβ_1–42, and Ni-bme-dach does not significantly perturb these already minimal effects. At pH 6.5, where Fe³⁺ remains soluble, metal binding produced anisotropy increases comparable to those observed for Cu²⁺ (Δr _avg = 0.03 ± 0.01). Ni-bme-dach partially reversed this increase, reducing Δr by approximately 58% to Δr _avg≈ 0.013 ± 0.01. One-way ANOVA confirmed a significant overall effect among conditions (F(2,6) = 16.86, p = 0.0034, η² = 0.85). Consistent with the observed Δr values, Fe³⁺ significantly differed from TAMRA-Aβ_1–42 (p = 0.015), while the Fe³⁺ vs Fe³⁺ + Ni-bme-dach comparison did not remain significant after correction (p = 0.064). Fe³⁺ + Ni-bme-dach was also indistinguishable from TAMRA-Aβ_1–42 (p = 0.24), consistent with partial, but not complete, reversal of Fe-induced aggregation.

Zn²⁺ induced minimal aggregation at both pH 8.0 and 6.5 (Δr _avg = −0.01 ± 0.01; Figure A,B), and addition of Ni-bme-dach produced no measurable change in anisotropy (Δr _avg ≈ −0.01 to 0.00 ± 0.01), indicating negligible Zn²⁺ extraction under these conditions. Consistent with these weak responses, anisotropy is inherently less sensitive to the modest structural perturbations associated with Zn–Aβ interactions. In contrast, Fe³⁺ exhibited partial, pH-dependent reversibility, whereas only Cu²⁺ showed a statistically validated, near-complete restoration of monomer-like anisotropy upon Ni-bme-dach treatment. Lastly, Ni-bme-dach exhibited fluorescence intensity comparable to that of the metal-free peptide, without exceeding it, indicating that it effectively extracts Cu²⁺ via direct metal extraction and complexation. This selective behavior suggests it does not disrupt local TAMRA–Tyr10 interactions or cause significant conformational changes. Additionally, the lack of anisotropy changes with Zn²⁺ indicates that this optical platform is not ideal for assessing Zn²⁺ chelators, as any structural changes are subtle. Furthermore, the solubility of metal species, as in the case of Fe³⁺ at pH 8.0, is crucial for binding to and aggregating TAMRA-Aβ_1–42.

Molecular Evidence for Targeted Cu²⁺ Chelation by Ni-bme-dach

UV–vis spectroscopy was used to probe metal coordination and chelation underlying the anisotropy responses directly. The spectrum of TAMRA-Aβ_1–42 displays a weak absorption at λ_max = 552 nm, characteristic of the TAMRA fluorophore. CuSO₄ exhibits a shoulder at λ_max = 246 nm, which red-shifts to 250 nm upon addition to TAMRA-Aβ_1–42, indicating Cu-peptide coordination (Figure A). Ni-bme-dach shows a prominent d–d transition at λ_max ≈ 456 nm, arising from N, S-chelation of Ni²⁺. Upon addition of Ni-bme-dach to Cu²⁺-TAMRA-Aβ_1–42, a new shoulder emerges at λ_max ≈ 341 nm alongside the Ni-centered d–d band, consistent with Cu extraction from the peptide and formation of a Cu–Ni-bme-dach complex (Figure A, inset). At pH 6.5, similar weak features are observed at λ_max ≈ 339 and 450 nm. These spectral signatures closely match those of the reported [(Ni-bme-dach)₃Cu₂Br]­Br complex, which exhibits weak d–d transitions near 390 and 440 nm (Figure S6A), supporting formation of a Cu–NiN₂S₂ “paddlewheel”-type structure.

4.

UV–vis spectroscopy comparing changes in the absorbance of TAMRA-Aβ_1–42, (A) Cu^II, (B) Zn^II, or (C,D) Fe^III and Ni-bme-dach in 10 mM phosphate buffer pH 8.0. as well as (D) Fe^III and Ni-bme-dach in 10 mM HEPES buffer pH 6.5. Final concentrations of the TAMRA-Aβ_1–42, metals, and Ni-bme-dach were 110 nM, 33 μM, and 0.35 mM, respectively.

In contrast, Zn²⁺, a d¹⁰ metal, does not produce distinct metal–ligand charge-transfer features upon binding to TAMRA-Aβ_1–42 (Figure B). Addition of Ni-bme-dach to Zn²⁺–TAMRA-Aβ_1–42 at either pH 8.0 or 6.5 leads to increased absorbance of Ni-bme-dach-associated MLCT bands at λ_max = 227 and 274 nm, while the Ni-centered d–d transition at λ_max ≈ 452 nm remains unchanged. These spectra closely resemble those of Ni-bme-dach₃(ZnCl)₂, which exhibits absorptions at λ_max = 226, 270, and 460 nm, consistent with weak Zn²⁺ coordination by Ni-bme-dach (Figure S6B). The minimal UV–vis perturbations observed are consistent with the weak, nonaggregating nature of Zn–Aβ interactions.

UV–vis studies involving Fe³⁺ are more complex due to its redox activity and pH-dependent solubility. At pH 8.0, FeCl₃ displays an absorption at λ_max = 278 nm that undergoes a ∼15 nm blue shift to ∼263 nm upon addition to TAMRA-Aβ_1–42, indicating metal coordination (Figure C). This feature diminishes upon the addition of Ni-bme-dach, consistent with reversible Fe binding. At pH 6.5, Ni-bme-dach induces only weak spectral features near λ_max ≈ 338 and 450 nm (Figure D), indicating modest Fe-ligand interactions. Although the structure of the resulting Fe–Ni-bme-dach after peptide chelation is unclear, literature precedents indicate that Fe can coordinate to the Ni-bme-dach in monodentate or bidentate fashion. − Together, these pH-dependent spectral changes suggest that Fe redox behavior and solubility limit both its interaction with Aβ_1–42 and its chelation by Ni-bme-dach.

Collectively, the UV–vis data provide molecular-level support for the selective extraction of Cu²⁺ from Aβ_1–42 by Ni-bme-dach, resulting in the formation of a well-defined [(Ni-bme-dach)₃Cu₂Br]­Br. In contrast, Zn²⁺ and Fe³⁺ exhibit weaker and more reversible interactions, consistent with their minimal or partial responses in anisotropy measurements. For comparison, UV–vis spectra obtained in the presence of EDTA show complete removal of both Cu²⁺ and Fe³⁺ at pH 8.0 and 6.5, yielding characteristic Cu-EDTA and Fe-EDTA signatures (Figure S7) and highlighting the nonselective nature of EDTA relative to Ni-bme-dach.

Morphological and Nanoscale Validation via TEM and AFM

Because fluorescence anisotropy demonstrated that Ni-bme-dach selectively reverses Cu-induced Aβ aggregation, and UV–vis identified the coordination states and metal-extraction mechanisms, nanoscale imaging was used to determine how these molecular events alter aggregate morphology. Fluorescence microscopy was not used because early stage assemblies at nanomolar peptide concentrations are too small for diffraction-limited imaging. AFM and TEM offer better nanoscale characterization to support Ni-bme-dach’s role as a selective Cu-extracting ligand alongside solution-phase anisotropy measurements. All samples were prepared under conditions identical to those used for anisotropy measurements, thereby enabling a direct correlation between hydrodynamic radius and nanoscale morphology. Briefly, baseline TEM images of TAMRA-Aβ_1–42 stained with uranyl acetate revealed an extensive network of small aggregates, likely oligomers (Figure (i)). These structures are nanoscale oligomers and aggregate networks, not mature fibrillar plaques. Due to low peptide concentrations and short incubation times, micrometer-scale plaques are not expected, and terminology has been updated for clarity. Upon Cu²⁺ addition, a vast network of nanoscale aggregates formed, appearing interconnected and significantly denser, often adopting globular and clustered structures (Figures (ii) and S8). The shift in the size of the extensive network of nanoscale aggregates is consistent with the increase in anisotropy observed in solution (Figure ). Interestingly, TEM images after Ni-bme-dach treatment still showed significant peptide aggregation, which appeared less dense but more compact than TAMRA-Aβ_1–42 alone (Figures (iii) and S8). The network morphology made it challenging to measure the dimensions of the nanoscale aggregates, likely due to drying effects and the presence of metal-chelate complexes that were not removed from the samples before deposition.

5.

TEM images of (i) 1.10 μM TAMRA-Aβ_1–42 in 10 mM phosphate buffer pH 8.0, with additions of (ii) Cu^II, (iii) Cu^II and Ni-bme-dach, (iv) Zn^II, (v) Zn^II and Ni-bme-dach, (vi) Fe^III, and (vii) Fe^III and Ni-bme-dach. Final concentrations of metals and Ni-bme-dach were 0.33 mM and 0.35 mM, respectively. The scale bar for all images is 100 nm.

TEM images after Zn²⁺ addition show moderate compaction of TAMRA-Aβ_1–42 nanoscale aggregates, resulting in a thinner clustered network that appears less dense in the presence of Ni-bme-dach, indicating that the ligand does have mild disruption of Zn-Aβ interactions or peptide-to-peptide interactions (Figure (iv) and (v)). This is consistent with the relatively minor anisotropy changes, confirming that Zn²⁺ induces mild peptide–peptide interactions. In contrast, TEM images of Cu²⁺ and Zn²⁺-TAMRA-Aβ_1–42 nanoscale aggregates after the treatment with broad-spectrum chelator EDTA showed a significant reduction in aggregate size and network interactions (Figure S9). Lastly, Fe³⁺ also induced an extensive and dense large network of nanoscale aggregates similar to Cu²⁺ (Figure (iv)). However, in the presence of Ni-bme-dach, aggregate size and density are drastically reduced, yielding smaller spherical structures with little evidence of network-like features (Figure vii), indicating that Ni-bme-dach strongly disrupts Fe³⁺-induced aggregation.

Fluorescence anisotropy indicates that Ni-bme-dach reverses metal-induced TAMRA-Aβ_1–42 aggregation, whereas TEM reveals residual nanoscale aggregates due to differences in the techniques’ respective sensitivity. Anisotropy is sensitive to the rotational dynamics of TAMRA-Aβ_1–42 in solution, where larger, metal-cross-linked structures decrease molecular mobility and thus lower anisotropy, even if smaller aggregates remain. In contrast, TEM captures all electron-dense species after drying, including baseline self-assembly and aggregated forms not reflected in the anisotropy signal. Consequently, Ni-bme-dach reduces metal-driven aggregation, as seen in anisotropy, but does not completely prevent Aβ-Aβ association in the absence of solvent, which accounts for the observed nanoscale aggregates in TEM. Across all chelator-lacking conditions, a clear hierarchy emerges in how metals shape Aβ aggregation. Cu²⁺ and Fe³⁺ induce the most pronounced structural perturbations, generating dense, compact, and irregular nanoscale aggregates, whereas Zn²⁺ produces only modest morphological changes. These baseline behaviors define the relative ability of each metal to perturb TAMRA-Aβ_1–42 structure and serve as a critical reference for interpreting ligand-mediated effects across the data set.

AFM measurements reinforced the TEM observations, quantified the hierarchy of chelation responsiveness, and provided information about the relative sizes of the species formed. TAMRA-Aβ_1–42 alone exhibits small spherical structures (2.1 ± 0.6 nm height; 7.5 ± 1.3 nm width) (Figures (i) and S10). In the presence of Cu²⁺ the nanoscale aggregates are large (9.6 ± 1.6 nm height; 154 ± 74.9 nm width) (Figure (ii)) and is significantly decreased to 4.4 ± 1.2 nm and 51.9 ± 21.6 nm in the presence of Ni-bme-dach (Figures (iii) and (S10)), representing a near-complete collapse that closely mirrors the full anisotropy recovery and UV–vis evidence for Cu²⁺extraction. Zn²⁺ caused only modest increases in aggregate size (7.3 ± 2.0 nm height), and Ni-bme-dach had no detectable structural impact on these assemblies (Figures (iv) and (v) and S12), consistent with Zn²⁺ evidence within for its low propensity to induce significant TAMRA-Aβ_1–42. Fe³⁺ aggregates displayed heterogeneous dimensions (8.4 ± 6.6 nm height; 90 ± 80 nm width) and remained large following Ni-bme-dach treatment (10.5 ± 10.1 nm height; 166 ± 138 nm width) (Figures (vi–vii) and S13), consistent with 58% recovery observed in the anisotropy studies and the aggregates observed by TEM. Overall, AFM confirms the selectivity hierarchy: Cu (fully reversible) > Fe (partially reversible) ≫ Zn (no effect), a pattern that aligns precisely with Ni-bme-dach thermodynamic preferences and mechanistic mode of action.

6.

AFM image scans with 2 × 2 μm x–y and 5 nm total z-range of (i) 0.55 μM Aβ_1–42 in 10 mM phosphate buffer pH 8.0, with additions of (ii) Cu^II, (iii) Cu^II and Ni-bme-dach, (iv) Zn^II, (v) Zn^II and Ni-bme-dach, (vi) Fe^III, and (vii) Fe^III and Ni-bme-dach. Final concentrations of metals and Ni-bme-dach were 0.33 mM and 3.6 mM, respectively.

7.

UV–vis spectroscopy comparing changes in the absorbance of TAMRA-Aβ_1–42, Fe²⁺, and Ni-bme-dach in (A) 10 mM phosphate buffer pH 8.0 and (B) 10 mM HEPES pH 6.5, individually and in combination with each other. Final concentrations of the TAMRA-Aβ_1–42, metals, and Ni-bme-dach were 55 nM, 66 μM, and 2.5 mM, respectively.

Upon introduction of Ni-bme-dach, the ligand remodels these metal-associated aggregates in a distinctly metal-dependent manner. Cu-derived aggregates undergo the most substantial structural disassembly, consistent with selective chelation observed by fluorescence anisotropy and UV–vis spectroscopy. Fe-derived aggregates exhibit only partial reorganization, and Zn-associated structures remain essentially unchanged, reflecting their weaker interactions with both Aβ and Ni-bme-dach. Collectively, these ligand-treated morphologies demonstrate that Ni-bme-dach does not function as a broad-spectrum chelator and instead acts through selective disruption of Cu-driven assemblies, while minimally influencing metals that perturb Aβ less strongly. Note that the residual aggregates visible by TEM and AFM reflect drying-induced stabilization of weakly associated Aβ species, and do not contradict the complete restoration of monomer-like rotational mobility observed in solution by anisotropy.

Integrated Mechanistic Model of Metal-Specific Aggregation and Chelation

Across fluorescence anisotropy, UV–vis spectroscopy, TEM, and AFM, a unified mechanistic framework emerges in which Cu²⁺, Zn²⁺, and Fe³⁺ drive fundamentally distinct Aβ_1–42 aggregation pathways with sharply different susceptibilities to chelation. Among these metals, Cu²⁺ uniquely promotes large, dense, and highly dynamic aggregates, as evidenced by the largest anisotropy increases (Figure A–C), characteristic Cu-Aβ charge-transfer features in UV–vis spectra (Figure A), and extensive network-like assemblies observed by TEM and AFM (Figures (ii) and (ii)). These Cu-derived aggregates are fully reversible, collapsing to monomer-like states upon treatment with either EDTA or Ni-bme-dach, as reflected by complete anisotropy recovery (Figures A and A-B) and pronounced reductions in aggregate dimensions by AFM (Figures (iii) and S11). This identifies Cu²⁺ as the only metal whose aggregation pathway is fully amenable to chelation-driven disassembly.

Central to this reversibility is Ni-bme-dach’s selective Cu-extraction mechanism. UV–vis spectroscopy reveals direct capture of Cu²⁺ into a discrete Cu–Ni complex, manifested as new absorption features near ∼341 and 456 nm (Figures A; S6A). This specific coordination pathway restores anisotropy and reduces aggregate size without inducing fluorescence hyper-recovery, demonstrating that Ni-bme-dach removes Cu²⁺ without broadly stripping metals or perturbing local fluorophore–peptide interactions. In contrast, EDTA reverses aggregation by indiscriminately removing metals, producing fluorescence hyper-recovery, global anisotropy collapse, and extensive aggregate dissolution across all metals.

Zn²⁺ forms small, amorphous, weakly associated oligomers that exhibit minimal fluorescence quenching, negligible anisotropy changes, and modest particle growth. Because Zn²⁺ binds weakly to Aβ_1–42 and Ni-bme-dach exhibits low Zn²⁺ affinity, Zn-induced assemblies remain largely unaffected by chelation. Fe³⁺ induces a third, more complex aggregation regime governed by redox chemistry and pH-dependent solubility. Fe-derived aggregates are heterogeneous and intermediate in size (Figures (vi) and (vi)), show complete reversibility with EDTA, and only partial reversibility with Ni-bme-dach (Figures and ). UV–vis spectra reveal extensive Fe³⁺/Fe²⁺ interconversion: FeCl₃ undergoes a blue shift from 278 to ∼263 nm upon binding to TAMRA-Aβ_1–42 at pH 8.0 (Figure C,D), closely resembling FeCl₂ binding behavior under alkaline conditions (Figure A), while FeCl₂ features disappear upon peptide addition at pH 6.5 (Figures B and S16). These transitions indicate that Fe coordination is dictated by redox speciation and solubility, with Fe²⁺ oxidizing before binding at pH 8.0, Fe³⁺ reducing before binding at pH 6.5, and both species converting to insoluble Fe­(OH)₃ under more acidic conditions (Figure S16). This mechanism is consistent with reports that Aβ can reduce Fe³⁺ to redox-active Fe²⁺, providing a plausible biochemical pathway for Fe-mediated redox cycling and ROS generation in Alzheimer’s disease lesions. Together, these observations define a mechanistic hierarchy in which Cu²⁺ uniquely forms a fully reversible aggregation pathway, whereas Zn²⁺ and Fe³⁺ drive aggregation processes with restricted chelation sensitivity.

Conclusion

Taken together, this work establishes a clear metal-specific hierarchy in Aβ_1–42 aggregation and its susceptibility to chelation. Cu²⁺ emerges as the most potent promoter of conformational reorganization, producing large, dynamic aggregates that, importantly, remain fully reversible upon chelation. Fe³⁺ and Zn²⁺ generate smaller or structurally heterogeneous aggregates shaped by redox cycling and weaker coordination, and these assemblies exhibit limited or negligible responsiveness to selective chelation. Within this landscape, EDTA and Ni-bme-dach display mechanistically distinct behaviors. EDTA disrupts all metal-Aβ states through nonselective stripping, yielding a hyperfluorescent, near-monomeric end point that masks key features of metal-specific aggregation. In contrast, Ni-bme-dach extracts Cu²⁺ through formation of spectroscopically distinct Cu–Ni complexes, fully reversing Cu-induced aggregation while minimally perturbing Fe- or Zn-associated species. This selectivity is reflected across fluorescence anisotropy, UV–vis coordination signatures, and nanoscale morphology, demonstrating that Ni-bme-dach does not act as a general metal remover but as a targeted Cu-extracting ligand while leaving Fe and Zn coordination largely intact. Consequently, this work identifies a viable design principle for next-generation, metal-selective AD chelation strategies.

A key strength of this study is the integration of real-time anisotropy kinetics with UV–visible spectroscopy and TEM/AFM imaging to build a unified mechanistic model of metal-driven aggregation and chelation-mediated disassembly. To our knowledge, this is one of the few direct comparative analyses of selective and nonselective chelation in metallopeptide aggregation. While the focus here is on early stage aggregation under controlled conditions, the mechanistic principles revealed, particularly the importance of thermodynamic selectivity and ligand field matching, provide a foundation for rational design of next-generation Cu-targeted chelators. These methods enable future work to compare TAMRA-Aβ_1–42 kinetics with unlabeled Aβ monitored by ThT or orthogonal probes, thereby further validating the correspondence between labeled and native aggregation pathways. More broadly, these results illustrate how coordination chemistry can be leveraged to selectively modulate metallopeptide structure while preserving essential metal–protein equilibria, a necessary criterion for therapeutic intervention.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c11345.

• The Supporting Information contains supplemental fluorescence spectra, UV–vis spectroscopy, TEM images, AFM image scans, AFM histogram analyses, and statistical analysis workflows supporting the conclusions of this study. Specifically, it includes representative fluorescence spectra of TAMRA-Aβ_1–42 under metal binding and chelation conditions at pH 6.5 and 8.0 (Figures S1–S5), UV–vis spectroscopy examining metal-peptide and chelator interactions (Figures S6, S7, S15, and S16), TEM images illustrating metal-induced aggregation and chelator-mediated reversal (Figures S8 and S9), AFM image scans and corresponding height distribution histograms quantifying aggregate morphology and reversibility (Figures S10–S14), and a detailed description of the statistical workflow used for fluorescence intensity and anisotropy analyses (PDF)

§.

A.N.S. and E.K.A contributed equally to this work. Fluorescence Studies were completed by A.S., E.A., D.F., E.L., J.G. TEM studies were completed by A.S., while E.A. completed the AFM studies. The conception of the study, experimental design, data analysis, and writing were completed by M.M, A.S., E.A, and DC.

The authors declare no competing financial interest.