Opposing Effects of Cucurbit[7]uril and 1,2,3,4,6-penta-O-galloyl-β-D-glucopyranose on Amyloid β_25–35 Assembly

Abstract

Alzheimer’s disease (AD) is a neurodegenerative disease characterized by extracellular deposits of amyloid β-protein (Aβ) in the brain. The conversion of soluble monomers to amyloid Aβ fibrils is a complicated process and involves several transient oligomeric species, which are widely believed to be highly toxic and play a crucial role in the etiology of AD. The development of inhibitors to prevent formation of small and mid-sized oligomers is a promising strategy for AD treatment. In this work, we employ ion mobility spectrometry (IMS), transmission electron microscopy (TEM) and molecular dynamics (MD) simulations to elucidate the structural modulation promoted by two potential inhibitors of Aβ oligomerization, cucurbit[7]uril (CB[7]) and 1,2,3,4,6-penta-O-galloyl-β-D-glucopyranose (PGG), on early oligomer and fibril formation of the Aβ_25–35 fragment. One and two CB[7] molecules bind to Aβ_25–35 monomers and dimers, respectively, and suppress aggregation by remodeling early oligomer structures and inhibiting the formation of higher-order oligomers. On the other hand, non-selective binding was observed between PGG and Aβ_25–35. The interactions between PGG and Aβ_25–35, surprisingly, enhanced the formation of Aβ aggregates by promoting extended Aβ_25–35 conformations in both homo- and hetero-oligomers. When both ligands were present, the inhibitory effect of CB[7] overrode the stimulatory effect of PGG on Aβ_25–35 aggregation, suppressing the formation of large amyloid oligomers and eliminating the structural conversion from isotropic to β-rich topologies induced by PGG. Our results provide mechanistic insights into CB[7] and PGG action on Aβ oligomerization. They also demonstrate the power of the IMS technique to investigate mechanisms of multiple small-molecule agents on the amyloid formation process.

INTRODUCTION

Alzheimer’s disease (AD) is a neurodegenerative disease pathologically characterized by extracellular deposits of amyloid β-protein (Aβ) in the brain, resulting in neuronal cell loss and subsequent dementia.^1–6 However, the toxicity of Aβ is not mediated by these large aggregates of Aβ. Rather, neurotoxicity correlates with the rapid self-assembly of soluble, oligomeric forms of Aβ. Interestingly, there are two major forms of Aβ: Aβ_1–42 constitutes only ~9 % of the total Aβ in the brain and is the more toxic alloform, whereas the more abundant species Aβ_1–40 (~90 %) is less toxic and has a slower aggregation rate.^7–11 The conversion of soluble Aβ monomers to amyloid Aβ fibrils is a complicated, dynamic process involving several transient oligomeric species,^12–17 several of which appear to be highly toxic and play a crucial role in the development of AD.^18,19

Although there are presently no effective treatments for AD, several therapeutic strategies are under development for prevention and treatment.^20–25 Approaches based on suppressing Aβ fibril formation appear to be insufficient and are likely to be unsuccessful given that soluble oligomers, rather than fibrils, are the probable toxic agents.^14,26,27 In this context, the search for potential inhibitors to shorten the lifetime of small and mid-sized oligomers has attracted great attention.^25,28–32 Promising results have been reported employing polyphenol derivatives as inhibitors of amyloid formation.^{30,31,33–35} In particular, 1,2,3,4,6-penta-O-galloyl-β-D-glucopyranose (PGG) (see Scheme 1), a large polyphenol from the traditional medicinal herb Paeonia suffruticosa, was reported to be a potent inhibitor for the aggregation of both Aβ_1–40 and Aβ_1–42.³³ However, it is unclear whether there are specific interaction motifs between polyphenols and amyloid peptides, as these compounds tend to form complexes with a wide range of proteins and peptides including Aβ, islet amyloid polypeptide (IAPP), α-synuclein, and tau.^30,34,36,37 It has been suggested that the inhibitory mechanism of polyphenols may be size and shape-dependent. For example, a previous study has shown that epigallocatechin gallate (EGCG; a polyphenol extracted from green tea) is an effective inhibitor of Aβ_25–35 aggregation because the peptide can tightly wrap around three faces of the inhibitor, whose shape resembles a propeller.³¹ Another study has shown that the effects of inositols, a class of small polyphenols, on amyloid aggregation are stereochemistry dependent.³⁸ However, there is not enough evidence to correlate inhibitory effects with the shapes of the polyphenols and the lengths of the peptide targets.

Scheme 1.

Structural formulas of 1,2,3,4,6-penta-O-galloyl-β-D-glucopyranose (PGG), cucurbit[7]uril (CB[7]) and Aβ_25–35.

Recently, a new supramolecular strategy approach has been employed to target certain types of residues explicitly.^32,39–41 For example, cucurbit[7]uril (CB[7]) (see Scheme 1), a synthetic macrocyclic receptor molecule,^42–44 has demonstrated inhibitory effects by locking aromatic residues inside its hollow core.⁴⁰ Each cucurbituril possesses a rigid hydrophobic cavity lying between the two oxygen-atom-rich rims, making it a compatible structure for host-guest chemistry.⁴⁵ The inhibitory effect of CB[7] on Aβ aggregation was suggested to be associated primarily with its hydrophobic cavity capable of capturing aromatic residues.⁴⁶ Polar and charged residues can in principle also interact with the hydrophilic surfaces of the molecule, but there is currently lack of experimental support for this effect.

The mechanistic effect of ligands like PGG and CB[7], individually or in combination, on transient early-stage soluble Aβ oligomers is unknown. Here, we investigate the effect of these ligands on oligomer and fibril formation of Aβ_25–35 (see Scheme 1), a cytotoxic fragment of Aβ that has been studied extensively by experimental and computational techniques.^31,47–54 The transient and polymorphic nature of aggregating systems poses significant challenges for traditional techniques aimed at detecting and characterizing early oligomer formation. Ion mobility spectrometry coupled mass spectrometry (IMS-MS) overcomes many of these challenges, and has been applied successfully to studies of structural changes in the self-assembly of Aβ and similar systems,^14,55,56 as well as to the evaluation of small-molecules that interfere with the Aβ assembly process.^{29,31,32,57–60} We report herein the effects of CB[7] and PGG on oligomer formation and aggregate morphologies of Aβ_25–35 using IMS-MS, transmission electron microscopy (TEM), and molecular dynamics (MD) simulations.

RESULTS and DISCUSSION

Aβ_25–35 aggregation and the transition of oligomeric states from isotropic to β-sheet

The nano-ESI-Q mass spectrum of Aβ_25–35 is given in Figure 1A. Numerous peaks are reported with respect to their n/z ratios, where n is the oligomer number and z the charge, including n/z = 1/2 (m/z = 530) and n/z = 1/1 (m/z = 1060) and numerous less intense peaks for higher order oligomers. Since these are relatively low resolution spectra a given n/z can contain several oligomeric states. For example n/z = 1/1 can also be n/z = 2/2. The various contributions of different oligomeric states to each peak can be obtained from the arrival time distributions (ATDs) given in Figure S2 in Supporting Information.

Figure 1.

ESI-Q mass spectra of (A) pure Aβ_25–35, (B) Aβ_25–35 incubated with CB[7] at 1:1, (C) Aβ_25–35 incubated with PGG at 1:1, and (D) Aβ_25–35 incubated with PGG and CB[7] at 1:1:1 molar ratio. The Aβ_25–35 concentration was 100 μM in acidified water (0.1% (v/v) of formic acid) in all cases. The peaks of the Aβ_25–35 homo-oligomers are annotated as ratios of the oligomer size n and the charge z, whereas the hetero-oligomer peaks of Aβ_25–35 with ligands are annotated by (n+p)/z, (n+k)/z and (n+p+k)/z where p and k are the numbers of PGG and CB[7] molecules, respectively.

The experimental collisional cross sections (CCS) (reported in Figures 2A and Supporting Information, Figure S2 and Table S1) are obtained from measuring ATDs at different pressure to voltage (P/V) ratios from a plot of arrival time (t_A) versus P/V (Equations S3 and S4). To analyze the transient states of the soluble oligomers, the experimental data were compared to the ideal isotropic growth model (Figure 2), in which the isotropic growth approximates the cross section of an oligomer whose structure is equally distributed in all spatial dimensions (σ_n = σ₁ × n^2/3).^31,61 For Aβ_25–35 a positive deviation from the ideal isotropic growth model is observed at the dimer and increasingly significant deviations are observed for trimers and larger oligomeric species (Figure 2A). Recent studies^31,61,62 have convincingly shown that significant and persistent positive deviations of oligomers above the isotropic growth curve are reliable indicators of β-sheet formation. This occurs because β-sheet growth occurs in a linear manner and hence the CCSs of β-sheet oligomers increase linearly with n, the oligomer number, not as n^2/3 which describes isotropic growth. More specifically, previous work has shown that the formation of β-sheet-like trimers and tetramers is a crucial step in the process of Aβ_25–35 aggregation.^31,47,63 In summary, the IMS-MS data indicate that extended β-sheet conformations are populated more abundantly as the Aβ_25–35 oligomer number increases, a clear signature of eventual amyloid fibril formation. These results are consistent with TEM images (Figure 3A) that shows the formation of amyloid fibrils and were obtained from the same Aβ_25–35 sample used in the IMS-MS experiments. Our data here are also consistent with previous studies on the same system.³¹

Figure 2.

Oligomer growth curve data of the Aβ_25–35 homo-oligomers of (A) pure Aβ_25–35, (B) Aβ_25–35 incubated with CB[7] at 1:1, (C) Aβ_25–35 incubated with PGG at 1:1, and (D) Aβ_25–35 incubated with PGG and CB[7] at 1:1:1 molar ratio.

Figure 3.

Representative TEM images of (A) pure Aβ_25–35, (B, C) Aβ_25–35 incubated with CB[7] at 1:1 and 1:10, (D, E) Aβ_25–35 incubated with PGG at 1:1 and 1:10, and (F) Aβ_25–35 incubated with PGG and CB[7] at 1:1:1 molar ratio. The scale bars for all images are given in the two panels of section F.

Inhibitory effect of cucurbit[7]uril (CB[7]) on Aβ_25–35 amyloid aggregation

As shown in Figures 1B and 2B, the presence of an equimolar concentration of CB[7] remodels and dramatically reduces the steady-state oligomer distribution of Aβ_25–35. In the presence of CB[7], the only identified Aβ_25–35 homo-oligomers are monomers at n/z = 1/2 (m/z = 530) and n/z = 1/1 (m/z = 1060), dimer at n/z = 2/2 (m/z = 1060), trimer at n/z = 3/4 (m/z = 795) and hexamers at n/z = 6/7 (m/z = 908) and n/z = 6/11 (m/z = 578). The low charge state mass spectral peaks at high m/z are absent, indicating that CB[7] not only modulated oligomer structures (see discussion below), but also prevented the formation of large oligomers.

Two new mass spectral peaks are recognized in Figure 1B as Aβ_25–35:CB[7] hetero-oligomers, in which the most abundant peak at m/z = 741 is a triply charged heterodimer of an Aβ_25–35 monomer and one CB[7] molecule ((n+k)/z = (1+1)/3), where k is the number of CB[7] molecules, and the peak at m/z = 1111 is a doubly charged heterodimer (see Figure S3 and Table S2). The CCSs obtained for Aβ_25–35:CB[7] hetero-oligomers (Figure 4A) suggest that the dimer is globular but the monomer has two conformations, one globular and one extended conformer.

Figure 4.

Oligomer growth curve data of hetero-oligomers of (A) Aβ_25–35 incubated with CB[7] at 1:1, (B) Aβ_25–35 incubated with PGG at 1:1, and (C) Aβ_25–35 incubated with PGG and CB[7] at 1:1:1 molar ratio. The cross sections of Aβ_25–35 oligomers reported here were extracted from the hetero-oligomer cross sections (see Supporting Information Tables S2, S3 and S4).

At 1:10 Aβ_25–35:CB[7] molar ratio, there are only three remaining homo-oligomers: Aβ_25–35 monomers and dimers at n/z = 1/2 (m/z = 530), n/z = 1/1 (m/z = 1060) and n/z = 2/2 (m/z = 1060) (Figure S1). The remaining peaks are Aβ_25–35:CB[7] hetero complexes previously described. The higher intensities of these Aβ_25–35:CB[7] complex peaks at 1:10 ratio suggests that CB[7] competitively binds to Aβ_25–35 species and suppresses Aβ_25–35 oligomer growth. Our data indicate that CB[7] is a very effective inhibitor that intervenes in the aggregation cascade by blocking self-assembly at the dimer stage. TEM imaging (Figures 3B, C) shows that CB[7] reduces the formation of macroscopic fibrillar aggregates compared to the TEM image of the pure Aβ_25–35, supporting the IMS-MS data.

Computational modeling provides evidence supporting an inhibitory effect of CB[7] on the amyloid formation of Aβ_25–35. The simulations reveal that CB[7] ligands preferentially capture the amino-groups of lysines (Lys28), and to a lesser extent, the positively charged N-termini of Aβ_25–35 monomers (see Figure 5). The binding of CB[7] specifically to lysines is also observed with the 38-residue islet amyloid polypeptide whose aggregation is a pathological hallmark of Type 2 Diabetes.⁴¹ The interaction motif is similar to the molecular tweezer ligands that possess torus-shaped binding pockets.^32,64,65 The polar pockets serve as specific hosts for charged residues, and the strong binding between protein:ligand complexes can modulate the aggregation cascade. Each CB[7] molecule has only one binding site for Aβ_25–35 and the bulkiness of CB[7] helps to isolate the Aβ_25–35 monomers from each other and prevent large oligomer formation.

Figure 5.

Representative structures of Aβ_25–35:CB[7] complexes obtained from simulations. (1+1) complexes extracted from the simulation, and their theoretical cross sections. The experimental cross section is an average of the (1+1)/2 and (1+1)/3 complexes. The CB[7] molecules are represented in filled surface with oxygens in red, carbons in cyan, nitrogens in blue and hydrogens in white. The Aβ_25–35 peptides are shown in cartoon representation and colored in violet. Lysine (yellow) and N-terminus (green) are shown in licorice representation.

1,2,3,4,6-penta-O-galloyl-β-D-glucopyranose (PGG) enhances Aβ_25–35 amyloid aggregation

In the mass spectrum of Aβ_25–35 incubated with PGG at a 1:1 molar ratio (Figure 1C) significant homo- and hetero-oligomers are present. This is in contrast to the mixture of Aβ_25–35 with CB[7] where very few oligomers are observed. The CCS data (Figures 2C and 4B, and Supporting Information, Figure S4 and Table S3) indicate the formation of extended Aβ_25–35 homo-oligomers and Aβ_25–35:PGG hetero-oligomers, which correlates with a conformational conversion from isotropic to β-sheet structures. It has been previously shown that the conformational transition of Aβ_25–35 trimer and tetramer from native to β-sheet is a crucial step for Aβ_25–35 fibril formation.^31,47 Therefore we examined the ATDs of the mass spectral peaks corresponding to these oligomers in the presence of PGG. The ATD at 795 m/z for the pure Aβ_25–35 sample displays only a compact trimer at ~0.51 ms (Figure 6A). However, the ATD of the same m/z in the Aβ_25–35:PGG mixture shows two distinct features corresponding to a compact structure (at ~0.51 ms) and a new dominant feature for an extended trimer at ~0.58 ms (Figure 6B). For the Aβ_25–35 tetramer (n/z = 4/3) at m/z = 1413, the ATDs show two features for both Aβ_25–35 and Aβ_25–35:PGG (Figure 6C, D), however, there is an increase in the population of the extended species when the ligand is added. The enhanced formation of extended Aβ trimers and tetramers in the presence of PGG unambiguously demonstrates that the ligand facilitates the conformational transition of Aβ_25–35 oligomers from compact to extended. This result is very surprising and intriguing because PGG was originally thought to be an amyloid inhibitor.^33,35 The TEM images (Figure 3D, E) of the same samples used in IMS-MS experiments show a concentration-dependent increase in the abundance of fibrillar Aβ_25–35 aggregates in the presence of PGG. The fibrils in the mixed system show a twisted morphology, whereas the fibrils obtained from the pure Aβ_25–35 sample showed a combination of straight and twisted morphologies. Lastly, the fibrils become more abundant in the presence of excess PGG (1:10 ratio), further suggesting that this ligand is not an aggregation inhibitor for Aβ_25–35, but rather an aggregation accelerator.

Figure 6.

Representative arrival time distributions (ATDs) for m/z = 795 and m/z = 1413 of (A, C) pure Aβ_25–35 and (B, D) Aβ_25–35 incubated with PGG at 1:1 molar ratio.

The MD simulations provide insight into Aβ_25–35 behavior in the presence of PGG. Since PGG is a large polyphenol, Aβ_25–35 is unable to wrap around this molecule like it could with EGCG (Figure 7).³¹ Unlike the case of CB[7], there are no specific binding sites between the ligand and peptide. The PGG core formed by one or more ligand molecules provides a template surface with both hydrophobic and hydrophilic sites to which one or more amphipathic Aβ_25–35 peptides can weakly attach and undergo structural conversion. This observation from the MD simulations is consistent with the experimental detection of complexes made of multiple PGG molecules and Aβ_25–35 in which the Aβ_25–35 oligomers adopt extended conformations. Lastly, the theoretical cross sections of hetero-oligomers extracted from the simulation show good agreement with experimental data.

Figure 7.

Representative structures of Aβ_25–35:PGG complexes obtained from simulations. (A, B) Hetero-dimers of one Aβ_25–35 and one PGG. Aβ_25–35 monomers adopt a wide range of structures from random coil to partial helix. (C, D, E) Hetero-tetramers of two Aβ_25–35 and two PGG molecules. The PGG molecules are represented in filled surface with oxygens in red, carbon in cyan and hydrogen in white. The Aβ_25–35 peptides are shown in cartoon representation where non-polar residues are shown in pink, charged residues in blue and polar residues in green.

Cucurbit[7]uril (CB[7]) overrides the effect of 1,2,3,4,6-penta-O-galloyl-β-D-glucopyranose (PGG) on Aβ_25–35 amyloid aggregation

Interestingly, when fresh Aβ_25–35 peptide is incubated together with both PGG and CB[7] at a 1:1:1 molar ratio, most Aβ_25–35 homo-oligomers become depleted (Figures 1D and 2D) and Aβ_25–35:CB[7] complexes dominate. No Aβ_25–35:PGG complexes are found in the mass spectrum. Instead, two new peaks appear at m/z = 791 and m/z = 845, which are accounted for by the “triple” complexes formed from Aβ_25–35, PGG and CB[7] molecules (i.e., (n+p+k)/z = (1+1+1)/4 and (2+1+1)/5), respectively).

The dominance of binary complexes between Aβ_25–35 and CB[7] in the 1:1:1 mixture indicates that Aβ_25–35 forms more stable complexes with CB[7] than with PGG. In addition, the formation of triple complexes between the peptide and the two ligands is intriguing. From the cross section data (Figure S5 and Table S4), the hetero-oligomer topologies are consistent with an isotropic conformation (Figure 4C). The data indicate that the strong inhibitory effect of CB[7] overrides the conformational conversion from isotropic to β-sheet structures and the multimeric Aβ_25–35 homo-oligomer formation promoted by the polyphenol. In a good agreement with the IMS-MS results, the TEM imaging in the presence of both CB[7] and PGG (Figure 3F) shows the formation of globular aggregates, and very few short and irregular fibers (data not shown).

In a further experiment, CB[7] was added to an Aβ_25–35 sample previously incubated with PGG (for 1 week at 1:1 ratio) to give final molar ratios of 1:1:5 and 1:1:10 of Aβ_25–35:PGG:CB[7], and the samples were incubated an additional two days. The TEM images (Figure S6) indicate that addition of CB[7] to the aged Aβ_25–35:PGG sample reduces the amount of fibrillar aggregates. Although some of the remaining fibers retain a typical amyloid fiber morphology, most exhibit some degree of irregularity, with regions that are thicker and more heavily stained. Indeed, the irregular fibers were similar to those few fibers that were observed when PGG and CB[7] were both present from the onset of aggregation (not shown). Collectively, these results demonstrate that CB[7] inhibits the assembly of fibrillar Aβ_25–35 aggregates and also triggers their disassembly in a concentration-dependent manner.

SUMMARY and CONCLUSIONS

In the current study, IMS-MS data reveal distinctive effects of two assembly modulators, CB[7] and PGG, on the amyloid aggregation system, as well as on the structures of Aβ_25–35 homo- and hetero-oligomers. The self-assembly of pure Aβ_25–35 (Figures 1A, 2A) proceeds from compact monomers into β-rich oligomers emerging at the dimer and trimer stages. In the presence of CB[7], one and two CB[7] molecules selectively bind to Aβ_25–35 monomer and dimer, respectively. The most probable binding site for CB[7] on Aβ_25–35 is the lysine residue (Lys28). This observation expands our knowledge of the action mechanism of CB[7], since its inhibitory effect on β-amyloid fibrillation was previously understood to involve the capture of only aromatic residues: The hydrophilic rims of CB[7] created by oxygen atoms are a potential site to capture polar and charged residues, while the hydrophobic cavity hosts aromatic residues. These two possible host sites reveal the versatility of CB[7] as a possible novel therapeutic agent for AD. The IMS-MS data reveal that CB[7] suppresses Aβ_25–35 aggregation by preventing homo-oligomer formation. Complex formation between CB[7] and Aβ_25–35 monomers and dimers depletes the population of Aβ_25–35 oligomers larger than dimer. Consequently, fewer fibrillar Aβ_25–35 aggregates are found by TEM imaging at a 1:1 Aβ_25–35:CB[7] molar ratio and aggregates are only very rarely observed at a 1:10 Aβ_25–35:CB[7] ratio (Figure 3B, C).

In contrast to CB[7], PGG significantly enhances the formation of Aβ_25–35 aggregates (Figures 1C, 2C and 4B). Complex formation between PGG and Aβ_25–35 promotes extended Aβ_25–35 conformations in both homo- and hetero-oligomers (Figures 6 and 7). Consequently, long fibrillar Aβ_25–35 aggregates are found by TEM in the presence of PGG at a 1:1 Aβ_25–35:PGG ratio, and these become remarkably more abundant with an increase in PGG molar ratio (1:10 of Aβ_25–35:PGG) (Figure 3D, E). This is the first time a polyphenol has been shown to accelerate amyloid formation. These experimental results diverge from previous reports in the literature,³³ which describe a strong inhibitory effect of PGG on the fibril formation of Aβ_1–40 and Aβ_1–42 (of which Aβ_25–35 is a fragment). The chemical structure of PGG possesses both hydrophobic (aromatic rings) and hydrophilic (hydroxyl groups) sites for Aβ binding. In fact, both IMS-MS and MD data indicate that the binding of this ligand to the peptide is non-selective. The opposing effects of PGG on Aβ_25–35 and full-length Aβ_1–40/Aβ_1–42 may arise from the difference in the conformations required for aggregation. Short peptides such as Aβ_25–35 can aggregate when adopting extended conformations^31,47 whereas the monomers of Aβ_1–40/Aβ_1–42 in the fibril state exist in strand-loop-strand conformations.^66,67 Another possibility is that PGG suppresses fibril formation in Aβ_1–40/Aβ_1–42 by interacting with segments other than Aβ_25–35. Aβ_16–22 [KLVFFAE]⁶⁸ could be a probable target for PGG as it contains an important hydrophobic stretch and aromatic residues that can interact with PGG through π-stacking.

Finally, when both CB[7] and PGG are simultaneously present during aggregation, the dominance of binary complexes of Aβ_25–35 and CB[7] over Aβ_25–35 and PGG indicates that the inhibitory effect of CB[7] overrides the stimulatory effect of PGG (Figures 1D, 2D and 4C). CB[7] competitively binds to not only Aβ_25–35 homo-oligomers, but also hetero-Aβ_25–35:PGG complexes, thereby suppressing the formation of large Aβ_25–35 oligomers and eliminating the isotropic to β-rich structural conversion induced by PGG.

METHODS

A detailed description of materials and methods is given in Supporting Information (Section S1). The stock solutions were prepared as follows: Aβ_25–35 and CB[7] samples were prepared in acidified water (0.1% (v/v) formic acid), whereas PGG was dissolved in water containing 20% methanol (v/v). These stock solutions were used for all experiments (pure Aβ_25–35, Aβ_25–35 with CB[7], Aβ_25–35 with PGG and Aβ_25–35 with PGG and CB[7] samples). For IMS-MS analysis, the samples were loaded into gold coated nano-ESI capillaries and electrosprayed in a home-built ion mobility spectrometry-mass spectrometer.⁶⁹ The theoretical cross sections were calculated using the projected superposition approximation (PSA) method available at http://luschka.bic.ucsb.edu:8080/WebPSA/.^70,71 For the TEM analyses, aliquots of the same sample solutions used in IMS-MS experiments were adsorbed onto mesh formvar/carbon copper grids and imaged on a JEOL 123 microscope with an ORCA camera and AMT Image Capture Software v. 5.24.

ABREVIATIONS USED

AD

Alzheimer’s disease

Aβ

amyloid β-protein

Aβ_1–40

amyloid β-protein (1–40)

Aβ_1–42

amyloid β-protein (1–42)

Aβ_25–35

amyloid β-protein (25–35)

Aβ_16–22

amyloid β-protein (16–22)

ATD

arrival time distribution

CB[7]

cucurbit[7]uril

CCS

collision cross sections

EGCG

epigallocatechin gallate

IAPP

islet amyloid polypeptide

IMS

ion mobility spectrometry

IMS-MS

ion mobility spectrometry coupled mass spectrometry

Lys

lysines

MD

molecular dynamics

PGG

1,2,3,4,6-penta-O-galloyl-β-D-glucopyranose

t_A

arrival time

TEM

transmission electron microscopy