Recent advances in mass spectrometry methods for the investigation of oligomeric intermediates of amyloid protein aggregation

1 Introduction

Protein misfolding and aggregation into amyloid structures is a competing pathway to the normal folding process of virtually any protein. This process leads to the accumulation of fibrils where monomers are stacked into β-rich structures (Sawaya et al., 2021) with variable degrees of structural disorder (Utami et al., 2025; Zhang et al., 2026). Amyloid structures can be physiologically relevant (Foust et al., 2026), but also correlate with degenerative pathologies, like Alzheimer’s disease (AD) and Parkinson’s disease (PD) (Siebner and Uversky, 2026). The aggregation process involves oligomeric structures, which can form by primary or secondary nucleation steps and represent either on-pathway or off-pathway intermediates (Wei et al., 2025). The structure of oligomeric intermediates remains elusive, due to their dynamic and heterogeneous nature. Nonetheless, their characterization could shed light on the aggregation mechanism and reveal potential targets of therapeutic strategies (Goncerz et al., 2026; Wei et al., 2025). Besides being recognized as toxic species (Gasca-Martínez et al., 2026; Ivanova et al., 2026a), they could carry structural determinants of “strains” involving the same protein in different pathologies (Lo Cascio et al., 2025; Saleem et al., 2026) and can also act as seeds in the so-called “prion-like” spreading, in which normally folded molecules of the aggregating protein are recruited in the pathogenic mechanism (Zampar et al., 2024). Thus, oligomeric states can reveal different aggregation pathways responsible of the diverse pathologies involving a given amyloid protein (Mahul-Mellier et al., 2026).

Mass spectrometry (MS) has become a central tool of structural biology (Britt and Robinson, 2025; Kaltashov et al., 2024). Its specific strength is dissecting heterogeneous samples into sub-populations, thanks to the ion-sorting mechanism inherent to MS analysis, thus delivering measurables for each detected ion (Reid et al., 2023). A whole class of methods is available for protein structural investigation and designated as structural MS (Wagner and Gross, 2024). One approach is the so-called native MS, which analyses intact proteins and supramolecular complexes preserving the non-covalent interactions responsible for folding and binding (Lawrence et al., 2025). The charge-state distribution (CSD) obtained by electrospray ionization (ESI) reports on protein compactness, while mass reports on binding and modifications. The shape of gas-phase ions can be explored by ion mobility (IM), which measures the rotationally averaged collisional cross section (CCS) (Panczyk et al., 2026). Another class of methods is based on conformation-sensitive, covalent modification of proteins in solution (footprinting), followed by bottom-up or top-down MS analysis of the reaction products (Castel et al., 2023). Different modification strategies have been used, including cross-linking (XL) (Balakrishnan et al., 2026a), hydrogen-deuterium exchange (HDX) (Nyvall et al., 2026), fast-photochemical oxidation (FPOP) (Wecksler et al., 2026), covalent protein painting (CPP) (Bamberger et al., 2026) and limited proteolysis (LIP) (Figure 1) (Elsässer et al., 2026).

Such an arsenal of methods help dissect dynamic structural ensembles, either by equilibrium or time-resolved measurements (Lento and Wilson, 2022; Wagner and Gross, 2024), overcoming ensemble averaging. Furthermore, these methods can be combined with conventional MS protocols to characterize post-translational modifications and mutations on each dissected structural species. Structural MS, however, does not achieve atomic resolution and suffers of possible structural rearrangements and limitations on buffer composition and protein concentration. Nonetheless, the applicability of these methods is rapidly evolving. They are tipically combined with thioflavin-T fluorescence for aggregation kinetics, small-angle x-ray scattering (SAXS) or nuclear magnetic resonance (NMR) for size distributions, NMR and computational simulations for modelling at atomic resolution (Happonen and Varjosalo, 2025). Native-MS methods also complement imaging by atomic-force microscopy (AFM), transmission and cryogenic electron microscopy (cryo-EM) and tomography (cryo-ET) (Hale and Cooper, 2025; Liang et al., 2025). This work highlights the specific contributions of structural MS dissecting heterogeneous and dynamic systems into structurally annotated sub-populations and yielding constraints for computational modelling (Cehlar et al., 2024; Ivanova et al., 2026b). The focus is kept on the oligomeric intermediates of amyloid proteins in the last 5 years (Wagner and Gross, 2024).

2 Amyloid β

A hallmark of AD is the aggregation of amyloid β (Aβ) peptides, derived from the cleavage of the amyloid precursor protein (APP) by β- and γ-secretases (Zetterberg and Bendlin, 2026). Aβ40 and Aβ42 are the most studied variants, but Aβ43 is also involved in the pathology and all these fragments accumulate in the AD brain (Llorente-Saguer et al., 2026; Tambini et al., 2023). Aβ42 and Aβ43 have a stronger propensity to aggregate and higher cytotoxicity than Aβ40 (Habeck et al., 2024). Native IM-MS has been used to compare homo- and hetero-oligomers (Habeck et al., 2024). This study shows that Aβ40 dimers and trimers have bimodal CCS distributions, revealing a considerable amount of extended oligomers coexisting with a more compact component. Aβ42 and Aβ43, instead, lack the extended populations of dimers and trimers. Interestingly, the hybrid dimers and trimers of Aβ40 with either Aβ42 or Aβ43 have similar CCS profiles as pure Aβ42 and Aβ43, where the extended population is depleted. Thus, Aβ42 and Aβ43 are capable of remodeling the oligomeric structure toward higher compaction. Considering their stimulating effect on Aβ40 aggregation kinetics, it is suggested that the more compact oligomers have higher aggregation propensity, offering a mechanistic interpretation of the reciprocal influence of the Aβ variants (Habeck et al., 2024). Native IM-MS has also detected the hexameric cylindrin state of the Aβ25-35 peptide, which represents the putative toxic form capable of inserting into membranes (Laos et al., 2021).

Further improvements are offered by cyclic IM (cIM) technology, which offers enhanced sensitivity and resolution through multiple ion cycling, amplifying differences in arrival times (Naylor and Nagy, 2025). An accurate methodological work on Aβ42 oligomers (Vlk et al., 2026) sets the background for applications of cIM to amyloid oligomers, analyzing the effects of instrumental parameters. It is shown that upon optimization of the experimental settings, Aβ42 oligomers up to the dodecamer can be detected, minimizing dissociation and denaturation. Gas-phase unfolding by collision-induced unfolding (CIU) highlights higher stability of the trimer relative to the dimer and distinct unfolding pathways (Vlk et al., 2026).

cIM has also been applied to the investigation of synthetic seeds consisting of covalent homo-trimers obtained by crosslinking the Aβ fragments corresponding to residues 17–36 or 16–36 (Guaglianone et al., 2022; Kreutzer et al., 2023). This portion spans the central and C-terminal regions, which are the most involved in aggregation (Chen et al., 2026). Each trimer is constrained in a β-hairpin conformation, as observed in the outer layers of mature fibrils (Ghosh et al., 2021), by an additional covalent bond, stapling together the N- and C-termini. These triangular seeds can associate into higher-order structures, offering a potentially useful tool to mimic early assemblies of Aβ42 (Guaglianone et al., 2022; Kreutzer et al., 2023).

Alkylation of carboxylic residues has been exploited by glycine ethyl ester (GEE) footprinting to investigate the early stages of Aβ42 oligomerization (Kuang et al., 2025). The results are consistent with involvement of the central and C-terminal regions in the aggregation process, while the N-terminus remains mostly exposed. Kinetic modelling identifies four populations that sequentially act as seeds of the following aggregation step: small oligomers assemble first (M1) and convert into secondary seeds (M2) by an irreversible structural rearrangement. M2 seeds grow into protofibrils (M3) and eventually mature fibrils (M4) (Kuang et al., 2025).

3 α-Synuclein

The aggregation of the presynaptic protein α-synuclein (α-syn) plays a central role in the etiology of PD and other synucleopathies (Arino et al., 2026). The wild-type protein is involved in sporadic forms of the disease, while mutants are implicated with familial PD (Ranjan and Kumar, 2026). Three distinct protein regions have been recognized: an N-terminal region capable of forming α-helices, interacting with membranes and hosting most of the pathogenic mutations, a central region that drives aggregation (non-amyloid component, NAC, region) and an acidic C-terminal region mostly involved in interactions with other proteins (Singh et al., 2025). Native-MS spectra of wt α-syn without preincubation generally reveal predominance of the monomeric form and very small amounts of oligomers (Wagner and Gross, 2024). It has been shown, however, that an N-terminally truncated form accumulating upon expression in Escherichia coli, shows predominant signals of pentamers, hexamers, and heptamers, suggesting a faster aggregation kinetics (Báez Bolívar et al., 2026; McGlinchey et al., 2021). These studies add to a growing body of evidence hinting to a role of C- and N-terminal α-syn truncation in PD pathogenesis (Mahul-Mellier et al., 2025; McGlinchey et al., 2025). This study uses Tris- and NaCl-containing solutions, normally not compatible with ESI, thanks to sub-micron emitters, which expand the range of explorable conditions (Báez Bolívar and McLuckey, 2025; Österlund et al., 2025; Susa et al., 2018).

α-Syn oligomers have also been investigated by IM-MS in response to protein oxidation by reactive oxygen species (ROS), which has been implicated in pathogenesis (de Araújo et al., 2026). Controlled oxidation was achieved by low-temperature plasma (LTP) treatment, exploiting ultrafast protein modification (microsecond to millisecond timescale) (Wang et al., 2025a). Bottom-up LC-MS/MS analysis following the modification shows that oxidation upon 2–8-min incubation modifies preferentially the C-terminal region of monomeric α-syn (specifically M116, M127 and Y133). The results are consistent with this being the most disordered part of the protein at neutral pH and points to LTP treatment as an effective conformational probe, in addition to the more frequently employed FPOP (Cheng and Gross, 2025). Native IM-MS of untreated α-syn shows well-detectable oligomers including dimers, trimers and tetramers. Increasing treatment times (2–8 min) lead to progressive loss of α-syn oligomers, particularly trimers and tetramers, indicating a lower propensity to oligomerize of the modified protein (Wang et al., 2025a). Circular-dichroism analysis suggests that C-terminal α-syn oxidation hinders β-sheet formation. LTP-oxidized α-syn also shows lower propensity to form fibrils and reduced cytotoxicity while oxidation of preformed fibrils does not mitigate toxicity. This result suggests a complex relation between α-syn oxidation and pathogenic mechanisms and will have to be reconciled with the available evidence concerning the role of oxidative stress in neurodegeneration (Wang et al., 2025a).

HDX is a widely used method in structural MS (Nyvall et al., 2026). It is based on the exchange of amide hydrogen with deuterium from surrounding D₂O (or vice versa). Hydrogen bonding or solvent shielding protects from exchange, resulting in lower isotope incorporation. Since the exchange affects mass, methods and software have been developed for the analysis of reaction products by MS, besides NMR. Every residue is informative, regardless of its chemical structure. HDX-MS has been described the effects of epigallocatechin-3-gallate (EGCG) on large α-syn oligomers (∼23 subunits and multiples of that) (Grønnemose et al., 2022). The results show selective protection of the N-terminus by EGCG, suggesting stabilization of a closed conformation. At the same time, EGCG induces oxidation of methionine residues in the same region, which appears highly disordered upon oxidation. Since a disordered and not modified N-terminus seems to mediate interaction with membranes and toxicity, these effects could be at the basis of EGCG protective action (Grønnemose et al., 2022).

HDX has also been used to investigate resveratrol interference with α-syn oligomerization (Illes-Toth et al., 2024). Pulsed labelling along the aggregation process shows the inhibitory role of resveratrol, particularly on peptides spanning the NAC region. Aggregation still proceeds in the presence of resveratrol, but the exchange patterns suggest remodeling towards a more open oligomeric structure (Illes-Toth et al., 2024).

Quite a unique role in structural MS is played by XL-MS. Reagents containing two reactive groups at defined distance are used to crosslink specific protein functional groups (Cupp-Sutton and Wu, 2026). Product analysis by bottom-up or top-down MS yields an upper limit to the distance between the crosslinked sites. Thus, crosslinkers act as chemical rulers to investigate protein structure, dynamics and interactions (Piersimoni et al., 2022). In this sense, XL-MS is analogous to FRET or NOE-NMR experiments, providing distance constraints that can be used for computational modelling. XL-MS has been applied to investigate α-syn oligomers bound to membrane mimics, revealing two structures (linear and circular) developing from α-syn dimers, depending on the relative orientation of the two subunits (Schwarz et al., 2023). Even low concentrations of the phase-2 drug UCB0599 lead to a rearrangement of the membrane-bound oligomers, changing crosslink pattern at the dimer interface. These results provide structural insights into the specific targeting of oligomeric α-syn by the drug (Schwarz et al., 2023). Analogously, XL-MS has been used to map the interaction between α-syn large oligomers (∼30 subunits) and PSMa3, a peptide known to abrogate oligomer toxicity and to inhibit conversion to fibrils (Santos et al., 2024). Four crosslinks were detected, all located in α-syn N-terminal region (residues 1–61). HDX measurements provide consistent protection patterns and indicate that PSMa3 binding site remains accessible in the oligomeric state (Santos et al., 2024). This region could represent an interesting structural target for therapeutic strategies (Santos et al., 2024).

The aggregation behavior of the NAC region has also been investigated using synthetic peptides corresponding to α-syn residues 68–78, in the NAC region (PD1 peptide), and residues 47–56, in the N-terminal region (PD2 peptide) (Depland et al., 2026; Depraz Depland et al., 2025). These studies have been performed on recently developed hybrid instruments that combine IM-MS with infrared multiple-photon dissociation (IRMPD) spectroscopy. Since MS does not provide direct information on protein secondary structure, the possibility to perform spectroscopic conformational analysis on the sub-populations resolved by IM-MS is extremely attractive. Despite very similar CCS values, the amide-A band shows that PD1 dimers form antiparallel β structure, while PD2 dimers contain parallel β-sheets. Dimers of the G51D pathogenic variant maintain a helical conformation similar to the monomer. The latter dimeric structures are also more compact and stable, based on resistance to ion heating by IRMPD (Depland et al., 2026; Depraz Depland et al., 2025).

4 Transthyretin

The recent technological development of direct charge detection MS (CDMS) overcomes the need for deconvolution procedures or isotopic resolution for charge assessment and mass calculation, enabling the analysis of heterogeneous and high-mass assemblies, such as higher-order protein oligomers (Kaltashov et al., 2024). IM-MS combined with CDMS has been used to study misfolding and aggregation of transthyretin (TTR) (Hampton et al., 2025). This protein is responsible for the physiological transport of thyroxine and retinol in plasma and cerebrospinal fluid. Its aggregation has been involved in senile systemic amyloidosis. TTR is natively tetrameric and aggregation is triggered by tetramer dissociation and misfolding of the monomers, which then re-associate into higher-order oligomers and eventually fibrils (Sun et al., 2025). Oligomerization has been explored as a function of pH, pre-incubation temperature and sequence variants (Hampton et al., 2025). IM-MS shows accumulation of extended monomers under acidic pH, enhanced by the pathogenic mutation L55 P. At the same time, CDMS shows that acidic pH, low temperature and the L55 P mutation induce heterogenous and higher-order oligomers (up to 19 subunits detected). Instead, the octamer dominates in wt TTR, consistent with association of native-like tetramers, under pH and temperature conditions less favorable to aggregation (Hampton et al., 2025). The temperature-dependence of aggregation suggests a role of hydration as supported by further analyses at variable-temperature ESI and effects of D₂O. Furthermore, HDX-MS could map “cold” water molecules, whose exchange with the bulk solvent is inhibited by the interactions with the protein structure, at the dimer/dimer interface and shows that they are responsible for stabilization of the native tetramer (Lantz et al., 2025, 2024).

5 TDP-43

TAR DNA-binding protein 43 (TDP-43) is the target of emerging therapeutic strategies for amyotrophic lateral sclerosis (ALS) (Estevez-Fraga et al., 2026). Its oligomerization has been investigated by IM-MS (Depraz Depland et al., 2023; Laos et al., 2021), also in interplay with an Aβ peptide, revealing reciprocal influence (Laos et al., 2021). Furthermore, formation of the toxic hexameric cylindrin state of TDP-43_307−319 could be monitored by IM-MS, testing the effects of drugs and identifying promising inhibitors (Liu X. et al., 2023). The effects of mutations and protein concentration has also been explored, identifying a dual kinetic aggregation pathway leading to either fibrils or cylindrin and offering an interpretation of toxicity phenotypes (Jin et al., 2025).

6 Other amyloid systems

Native IM-MS has also been employed to study oligomerization of peptides (either designed or representing distinct regions of amyloid proteins), evaluating the effects of mutations and conditions on size distribution and stability (Hirschbeck et al., 2025; Jin et al., 2025; Wang et al., 2025b). Furthermore, it has been employed to explore the effect of disulfide shuffling on insulin aggregation pathway (Qin et al., 2025). Higher-order oligomers of the therapeutic peptide liraglutide have been detected by DCMS, revealing a dual aggregation pathway (Kuo et al., 2025).

XL-MS has been applied to the investigation of the early stages of aggregation of the protein β2 microglobulin, responsible for the dialysis-related amyloidosis. An off-pathway covalent dimer has been produced by photo-induced crosslinking and shown to inhibit the aggregation of unmodified protein (Maya-Martinez et al., 2022).

HDX-MS has been applied to structural characterization of oligomeric intermediates in the aggregation pathway of insulin β-chain (Renawala and Topp, 2022), glucagon-analog peptides (Renawala et al., 2024), and tau (James et al., 2025). In the latter case, by relating the oligomer structure to fibril morphology, it was possible to show that the polymorphism due to aggregation conditions is imprinted in the early stages of aggregation (James et al., 2025).

Islet amyloid polypeptide (IAPP) is a polypeptide hormone that aggregates in the islets of Langerhans, contributing to β-cell dysfunction in type-2 diabetes and showing pathophysiological connections to insulin and Aβ (Bousch et al., 2026; Dai et al., 2024; Fu et al., 2026). Structural MS has long been applied to the study of IAPP oligomeric intermediates (Abedini et al., 2016; Eldrid et al., 2021; Lim et al., 2024; Ridgway et al., 2020; Xu et al., 2022; Young et al., 2015). Among the postulated toxicity mechanisms, interactions of IAPP with membranes play a key role (Bousch et al., 2026; Dai et al., 2024; Mithu et al., 2025). These interactions have been recently investigated by native MS of IAPP bound to nanodiscs (Sanders et al., 2023), membrane mimetics consisting of lipid bilayers surrounded by a scaffold protein belt (Townsend and Marty, 2023). Oligomerization of a host protein bound to nanodiscs can be assessed by mass defect analysis of the raw nMS spectra (Townsend and Marty, 2023). This approach has shown that human IAPP forms tetramers in a nanodisc environment, while the non-amyloidogenic rat variant binds sequentially and only up to a 2:1 stoichiometry. Still another behavior is shown by the non-amyloidogenic pig protein, lacking binding at all (Sanders et al., 2023). These experiments set the basis for further mechanistic studies and screening of inhibitors.

7 Future perspectives

Two powerful tools of structural MS have not been applied yet to the investigation of amyloid oligomers, limited proteolysis coupled to MS (LiP-MS) (Elsässer et al., 2026) and covalent protein painting (CPP) (Son et al., 2023). These approaches probe local solvent accessibility and conformational dynamics, respectively, by enzymatic proteolysis and isotope-coded chemical methylation. Downstream, quantitative, bottom-up MS generates fingerprints of the original structures, with site-specific information on the degree of solvent exposure. These methods have been applied to structural investigation of amyloid fibrils, e.g., α-syn (Serdiuk et al., 2026) and Aβ40/Aβ42 (Bamberger et al., 2026), characterizing alternative polymorphs, as well as to monomeric and fibrillar α-syn, identifying mitochondrial interactors (Serdiuk et al., 2025). Future extension of these approaches to oligomeric structures is highly promising. It has been shown that sensitivity to proteolysis discriminates stability and morphology of α-syn and tau oligomers (Lo Cascio et al., 2025).

Protein amyloid aggregation inside the cell is promoted by liquid-liquid phase separation (LLPS), a process with complex physiological and pathological implications (Lucas et al., 2026; Mukherjee et al., 2024; Tang et al., 2025). XL-MS has been successfully applied to α-syn characterization inside LLPS “droplets” showing accumulation of an elongated conformer (Ubbiali et al., 2022). It would be important to also catch the oligomeric species of amyloid proteins that form under LLPS conditions by this and other MS-based techniques (Sahin et al., 2023).

A major Frontier lies in the integration of MS with artificial intelligence (AI), for instance, facilitating automated feature extraction, spectra deconvolution, and quantitative analysis (Yu et al., 2023). AI-based tools, such as AlphaFold, can be combined with structural MS, e.g., with distance constraints derived from XL-MS, to model dynamic macromolecular assemblies and non-native or self-associating amyloid architectures (Liu, 2022). Furthermore, hybrid workflows combining structural MS with cryo-EM and cryo-ET offer exceptional synergy for integrative structural characterization (Greenblatt et al., 2024; Quintana-Gallardo et al., 2022). The combination of single particle cryo-EM and structural MS can capture transient or low-abundance species that resist conventional structure determination.

The different approaches of structural MS are increasingly implemented ex-vivo (El-Baba et al., 2026) and XL-MS (Balakrishnan et al., 2026b) also sees applicability in-vivo (Balakrishnan et al., 2026a). Methods for native ambient MS (NAMS) are rapidly growing, paving the way to tissue-derived and patient-derived oligomer analysis (Du and Cooper, 2025; Hughes et al., 2024; Liu R. et al., 2023; Venter, 2025; Villacob et al., 2024; Wong et al., 2023). Future developments in this direction will give important contributions to the understanding of amyloid biology and amyloid diseases. These methods also have the potential to reach the omic scale (Khanppnavar et al., 2024) and time-resolved implementation (Nkyaagye et al., 2026; Overby et al., 2026), opening exciting new perspectives in amyloid structural biology, not only by assisting drug discovery and elucidating aggregation mechanisms, but also helping discovery of novel biomarkers and understand the biological response to protein aggregation.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Chiranjeevi Pasala, Sloan Kettering Insitute, United States

Reviewed by: Shoyab Ansari, University of Texas Southwestern Medical Center, United States