Mechanism-selective inhibition of α-synuclein aggregation by the chaperone-like BRICHOS domain Authors: Laurène Adam, Willem Hendrik Molenkamp, Jan Stanislaw Nowak, Azad Farzadfard, Koen Gaarthuis, Rakesh Kumar, Janni Nielsen, Daniel E Otzen, Jan Johansson, Axel Abelein Published: 2026 ABSTRACT AbstractCurrent therapeutic approaches for Parkinson’s disease and other synucleinopathies alleviate symptoms but fail to effectively prevent disease progression. As a result, there is an increasing focus on alternative disease-modifying strategies where molecular chaperones are emerging candidates. Recently, the chaperone-like Bri2 BRICHOS domain has been shown to be a promising therapeutic candidate, inhibiting amyloid formation and associated toxicity of multiple amyloidogenic proteins including human α-synuclein (αSyn). To advance the development of Bri2 BRICHOS as a therapeutic, in vivo tests in mice are necessary, which commonly rely on the injections of preformed fibrils of mouse αSyn. Here, we investigate the inhibitory mechanism of Bri2 BRICHOS on mouse αSyn aggregation and fibril interaction. In contrast to previous results on human αSyn, we found that Bri2 BRICHOS exhibits a very modest inhibitory effect on mouse αSyn aggregation, which is only observed under gentle shaking or quiescent conditions. While Bri2 BRICHOS binds with similar affinities to the respective fibrils, we observed that differences in the underlying nucleation mechanisms of mouse versus human αSyn fibril formation explain the impaired suppression of mouse αSyn fibrillation under strong shaking conditions. The more fragile nature of mouse αSyn fibrils causes stronger contributions of fibril fragmentation processes compared to surface-catalyzed secondary nucleation—the dominant nucleation mechanism for human αSyn. In conclusion, these findings provide molecular insights into the mechanism-of-action of Bri2 BRICHOS-mediated inhibition of αSyn aggregation as a selective chaperone-based inhibitor of surface-catalyzed secondary nucleation pathways, which facilitates informed choices of in vivo model systems for future treatment studies.Keywords: amyloid, BRICHOS, molecular chaperone, Neurodegenerative disorders, Parkinson’s disease, α-synuclein No curative treatment is currently available for patients with Parkinson’s disease (PD) and other synucleinopathies. Although symptomatic therapies exist, no disease-modifying drug has yet proven capable of halting disease progression (1). Increasing hope for new therapeutic strategies lies in naturally occurring molecular chaperones, such as the BRICHOS protein domain, which has shown great promise due to its potent anti-amyloid activity (2). The BRICHOS domain, found in 10 human protein families, has been demonstrated to inhibit aggregation and its associated toxicity of diverse amyloidogenic proteins (2). In particular the BRICHOS domains from the Bri2 proprotein and from the pro-surfactant protein C (proSP-C) have shown pronounced effects (3, 4, 5, 6, 7, 8, 9). Recombinant Bri2 BRICHOS efficiently suppresses aggregation of amyloid-β (Aβ42) (4, 5), islet amyloid polypeptide (IAPP) (6), Tau (7) and human α-synuclein (h-αSyn) (8), and likewise, proSP-C BRICHOS has been reported to inhibit Aβ42 (3, 5) and h-αSyn (9). Notably, while recombinant human Bri2 BRICHOS appears in different assembly states with different amyloid-suppression activities (4), a single-point mutant (R221E) of Bri2 BRICHOS was designed that stabilized the monomeric form and showed improved abilities to prevent neurotoxic effects (10). In contrast to proSP-C BRICHOS, which naturally appears as a trimer, monomers of WT and R221E Bri2 BRICHOS have the ability to cross the blood-brain barrier (11, 12). Most intriguingly, Bri2 BRICHOS R221E (hereafter referred to as BRICHOS) has shown promising therapeutic outcome in Alzheimer’s disease APP knock-in mouse models by reducing plaque load and neuroinflammation, as well as alleviating memory impairment symptoms (12, 13). Building on the promising results in the case of treatment studies for AD (12, 13), and BRICHOS’ ability to prevent in vitro h-αSyn aggregation and ex vivo toxicity (8), translating studies to test the efficiency of BRICHOS in PD mouse models are considered essential, where an adequate choice of the in vivo model is required. Among the widely used PD in vivo models, pre-formed fibrils (PFFs)-based mouse models enable cell-to-cell propagation of αSyn pathology, ultimately leading to dopaminergic neuronal loss (14, 15). Notably, recombinant mouse αSyn (m-αSyn) PFFs exhibit a greater propensity to seed endogenous m-αSyn within the brain compared to recombinant h-αSyn PFFs (15, 16, 17, 18), highlighting the importance of investigating BRICHOS efficacy against m-αSyn aggregation. Remarkably, mice do not develop PD-like pathology, despite the presence of Thr53 in m-αSyn, and that the A53T mutation in h-αSyn is linked to early-onset familial PD (19). This discrepancy suggests that other amino acid sequence differences between mouse and human αSyn confer protective effects (20). m-αSyn differs from h-αSyn by seven residue differences located in all parts of the protein (Fig. 1A), making it an interesting system for comparative studies to discover determinants of αSyn pathology. Among these, A53T and S87N mutations or residue differences were reported to accelerate m-αSyn aggregation (21, 22, 23). Recently, we revealed the inhibitory mechanism of BRICHOS on h-αSyn aggregation, where BRICHOS binds to the fibril surface and predominantly suppresses surface-catalyzed secondary nucleation and fibril-end elongation events. This mechanism implies a reduction of small nucleation units, or oligomers, which can be related to attenuated fibril-related toxicity as measured by the impact on γ-oscillations of mouse hippocampal brain slices for h-αSyn (8). Mechanistically, shared inhibitory effects of BRICHOS were reported for Aβ (4, 10) and Tau (7) aggregation, where BRICHOS binds to the amyloid fibril surfaces (24) and suppresses surface-catalyzed secondary nucleation, which is the key pathway in Aβ42 (25), Tau (26), and αSyn (27, 28) aggregation. Notably, only a small proportion of BRICHOS is bound to the fibrils (7, 8, 29), where BRICHOS apparently recognizes aggregation hotspots on the fibril surface, as determined for Aβ42 fibrils (29). Effectof BRICHOS on αSyn variants’ aggregation under different conditions.A, schematic representation of the protein sequence of h-αSyn, h-αSyn A53T, h-αSyn A53T S87N, and m-αSyn. For m-αSyn, the seven residue differences to h-αSyn are identified by the residue found in the human sequence followed by the sequence location and the residue found in the mouse sequence. B, aggregation kinetics of 30 μM m-αSyn and h-αSyn in 20 mM NaP buffer, pH 7.4 under strong shaking conditions (9 min ON/1 min OFF, 300 rpm) with or without 50% BRICHOS (molar equivalent), showing no inhibitory effect of BRICHOS on m-αSyn in contrast to h-αSyn. C, aggregation half-time (thalf) at 50% molar equivalent of BRICHOS of 30 μM h-αSyn, 70 μM h-αSyn A53T (8), 30 μM h-αSyn A53T S87N, and m-αSyn under the same conditions as above. Error bars represent the standard deviations between six technical replicates. D, aggregation kinetics of 70 μM h-αSyn A53T S87N in 20 mM NaP buffer, pH 7.4 under strong shaking conditions with or without 50% BRICHOS. E, aggregation kinetics of 70 μM m-αSyn in PBS under gentle shaking conditions (1 min ON/9 min OFF) with increasing concentrations of BRICHOS (10–70%, color gradient from brown to green). Global fits using a secondary nucleation-dependent model (solid lines) where either kn, k+, or k2 is the free fitting parameter. The relative mean squared errors (MSE) are given for comparison between the fits. F, aggregation kinetics of 70 μM m-αSyn in PBS under gentle shaking conditions with 5% seeds and increasing concentrations of BRICHOS (2–20%, color gradient from brown to yellow). Linear fits of the first 48 min are shown in solid lines. G, dependences of the combined rate constants k+k2, for h-αSyn under strong shaking conditions (8) and m-αSyn, under gentle shaking conditions where knk+ is set as a global constant, and k+, extracted from the highly seeded kinetics, are plotted against the BRICHOS:αSyn ratio, showing that secondary nucleation is dominantly affected by BRICHOS for both h-αSyn and m-αSyn. H, aggregation kinetics of 30 μM m-αSyn under quiescent conditions in PBS with 1% seeds and with 50% BRICHOS (yellow), showing an inhibitory effect of BRICHOS. For all aggregation kinetics data, normalized and averaged traces (dots) of six technical replicates are shown. Analyses of independent biological replicates are presented in Figure S1, showing a qualitatively similar effect. Here, we test whether BRICHOS modulates m-αSyn aggregation. We find that BRICHOS slows down m-αSyn fibril formation, but its effect is less pronounced compared to h-αSyn and only present under gentle shaking or quiescent conditions. Under strong shaking conditions, surface-catalyzed secondary nucleation is the dominant mechanism for h-αSyn aggregation (28), while m-αSyn undergoes excessive fibril fragmentation, a process that BRICHOS cannot influence, which explains the mechanism-selective inhibitory effect of BRICHOS. The current findings may facilitate the choice of an adequate PD mouse model system for BRICHOS-based in vivo treatment. Results BRICHOS inhibits m-αSyn aggregation under gentle shaking and quiescent conditions To evaluate the inhibitory effect of BRICHOS on m-αSyn, Thioflavin T (ThT)-based aggregation kinetics were first performed using the same conditions as described previously for h-αSyn (8). The experiments were conducted using 30 μM m-αSyn or h-αSyn in 20 mM NaP, pH 7.4 under strong shaking conditions at 300 rpm for 9 min ON and 1 min OFF (Fig. 1B). While the aggregation of h-αSyn was effectively suppressed, as expected from previous results (8), surprisingly, the aggregation of m-αSyn was not significantly modulated by BRICHOS. These experiments were repeated using different concentrations of m-αSyn (30 and 70 μM) and different batches of proteins, together with h-αSyn as a control (Fig. S1). Consistently, we found that BRICHOS does not inhibit the aggregation of m-αSyn under the conditions which are effective for the human variant (8). Several mutations or residue differences between human and mouse αSyn, such as A53T and S87N, have been reported to be important for accelerating αSyn aggregation (21, 23). To test the effect of these residue differences between m-αSyn and h-αSyn aggregation behavior, we assessed BRICHOS efficacy on the h-αSyn A53T S87N double mutant compared to h-αSyn as well as previously analyzed h-αSyn A53T (8) (Fig. 1C). The analysis of the aggregation half-time (thalf) of h-αSyn, h-αSyn A53T (8), h-αSyn A53T S87N, and m-αSyn revealed that BRICHOS has a smaller effect on h-αSyn A53T compared to h-αSyn (8) and the h-αSyn A53T S87N double mutant (Fig. 1, C and D). As noted above, the inhibitory effect of BRICHOS was completely abolished for m-αSyn (Fig. 1, B and C). These findings suggest that A53T and S87N mutations or residue differences found in m-αSyn diminish the effect of BRICHOS but are not the only contributors to reduce BRICHOS’ efficiency. A recent structural study of in vitro–formed m-αSyn fibrils revealed a distinct β-sheet architecture that appears to be less stable than h-αSyn fibrils, thereby making them more prone to fibril fragmentation (30). Sonication of these fibrils and detergent sensitivity experiments demonstrated that m-αSyn fibrils were more fragile than their human counterparts (30). We hypothesized that BRICHOS fails to act as a surface-catalyzed secondary nucleation inhibitor because m-αSyn is more prone to undergo fibril fragmentation than the human variant. In other words, under strong shaking conditions, m-αSyn fibril fragmentation may become the dominant nucleation mechanism that drives its aggregation, a process that BRICHOS presumably cannot inhibit (8). Fragmentation is often attributed to agitation during the fibrillation reaction (31). To test this hypothesis, we designed additional aggregation kinetics experiments to reduce fibril fragmentation by lowering the shaking time from 9 min ON and 1 min OFF to 1 min ON and 9 min OFF, which we referred to as gentle shaking, and ultimately to quiescent conditions, that is no shaking. Under gentle shaking conditions, we observed that BRICHOS indeed had a concentration-dependent inhibitory effect on m-αSyn (Fig. 1E). This effect was apparently saturated at 25% molar equivalent of BRICHOS (Fig. 1E). Previously, saturation of BRICHOS’ inhibitory effect was also observed against h-αSyn (8), which is caused by saturation of the binding of BRICHOS on the fibril surface. As an additional negative control, we evaluated the effect of bovine serum albumin, which indeed did not inhibit m-αSyn aggregation (Fig. S2). To elucidate how BRICHOS inhibits m-αSyn, we focused on the mechanism of fibrillization, which can be described as a series of nucleation-dependent processes, including primary nucleation (kn) and elongation (k+) along secondary pathways such as monomer-dependent fibril surface–catalyzed secondary nucleation (k2) and monomer-independent fibril fragmentation (k−). First, a secondary nucleation dominated model was used to analyze the aggregation traces using the online platform AmyloFit (32). To determine the effect on an individual nucleation rate constant (kn, k+, and k2), we constrained two of the three fitting parameters, while letting the third one free (Fig. 1E). This analysis revealed that the best fits were obtained when k2 or k+ was the free fitting parameter, indicating that BRICHOS inhibits fibril-end elongation and/or surface-catalyzed secondary nucleation of m-αSyn under gentle shaking conditions. To evaluate whether fibril-end elongation was inhibited by BRICHOS, 5% seeds (sonicated preformed fibrils) were added to the aggregation reaction. While the presence of seeds bypasses primary nucleation, a high seed concentration also allows to directly monitor fibril-end elongation due to the availability of a large number of fibrils ends (31). The fibril-end elongation rate (k+) is thereby directly proportional to the initial slope of the hyperbolic aggregation trace. Remarkably, using 2 to 20% molar equivalent of BRICHOS, we could observe a concentration-dependent inhibition of the aggregation of m-αSyn (Fig. 1F). BRICHOS is therefore able to delay the fibril-end elongation of m-αSyn. Since the fitting parameters k+ and k2 are coupled, the fits of the individual fitting parameters can also be reproduced using the combined rate constants k+k2 (Fig. S2). The dependence on the BRICHOS ratio of the globally fitted rate constants k+k2 can be plotted together with k+ obtained from the initial slopes (Fig. 1G), showing that the observed reduction in k+k2 cannot solely be attributed to the inhibition of fibril-end elongation by BRICHOS. Instead, a major contribution stems from the suppression of surface-catalyzed secondary nucleation by BRICHOS. Notably, a comparison of the effect of BRICHOS on the combined rate constants k+k2 between m-αSyn and h-αSyn (8) (Fig. 1G) reveals that the effect of BRICHOS is around 10 times more pronounced for the human variant of αSyn than for the mouse variant. To confirm these findings, we applied quiescent conditions, under which BRICHOS should also exhibit an inhibitory effect on m-αSyn. Without shaking, seeds are required to induce aggregation of monomeric m-αSyn in the chosen time window. With 1% seeds, we found that 50% molar equivalent of BRICHOS was able to delay m-αSyn aggregation (Fig. 1H), supporting that BRICHOS also exhibits an inhibitory effect under seeded, quiescent conditions where fragmentation is minimized. In summary, gentle shaking or quiescent conditions are necessary for BRICHOS to execute its anti-amyloid property against m-αSyn, where BRICHOS remains noticeably less effective compared to its efficiency against h-αSyn. BRICHOS inhibits predominantly surface-catalyzed secondary nucleation in addition to fibril-end elongation for both m-αSyn and h-αSyn, however, its effectiveness, in particular the reduction of k2, is lower for the mouse variant than for the human variant of αSyn. Strong shaking induces fibril fragmentation in m-αSyn To prove the hypothesis that BRICHOS is unable to inhibit m-αSyn aggregation under strong shaking conditions due to excessive fibril fragmentation, we compared m-αSyn fibril lengths formed under two distinct conditions: strong shaking and quiescent, both in the presence of 1% seeds and h-αSyn fibrils in strong shaking conditions without seeds. Fibrils were obtained at the aggregation plateau, washed, and subsequently imaged using negative-stain transmission electron microscopy (TEM). Analysis of the measurable fibril lengths (<800 nm) revealed that m-αSyn fibrils (n = 198) formed under strong shaking conditions were significantly shorter than those formed under quiescent conditions (n = 198) (Fig. 2, A, B and D). Notably, only fibrils that could confidently be measured, that is with visible fibrils ends, were counted. Therefore, fibrils with lengths above 800 nm, which were present in all samples, could not be considered as their fibril ends were not clearly visible. Despite this limitation, the results show that there are significantly more short fibrils present for m-αSyn under strong shaking conditions than for m-αSyn under quiescent conditions. Similarly, in the same strong shaking conditions, there is a higher number of shorter m-αSyn fibrils than for h-αSyn fibrils (n = 179) (Fig. 2, C and D). Of note, while branching of fibrils might be present, which has previously been related to secondary nucleation (33), the observed fibrils are predominately unbranched, which suggests that nucleation units easily detach from the fibrils. Thus, these findings support that fibril fragmentation is a more dominant nucleation process at strong shaking compared to gentle shaking for m-αSyn, compatible with m-αSyn fibrils’ apparently more fragile nature. Strong shaking conditions promote fibril fragmentation of m-αSyn fibrils.A–C, negative-stain TEM images of m-αSyn fibrils made in the presence of 1% seeds under strong shaking conditions (9 min ON/1 min OFF) (A) and quiescent conditions (1 min ON/9 min OFF) in PBS (B) as well as h-αSyn fibrils under strong shaking conditions in PBS (C). D, fibril lengths of measurable fibrils (<800 nm) of m-αSyn (n = 198) and h-αSyn (n = 179) in strong shaking and m-αSyn in quiescent conditions, ∗∗∗∗p < 00,001, ns = non-significant with p > 0.05, using one-way ANOVA. E, schematic representation of the sucrose gradient fractions from 1 to 9. Sucrose gradient (15–60%) centrifugation was performed from n = 3 biological replicates using 70 μM of m-αSyn fibrils made under strong shaking and quiescent conditions in the presence of 1% seed and one replicate for m-αSyn monomers as a control. Sucrose gradient fractions were analyzed by SDS PAGE followed by immunoblotting and the amount of m-αSyn in each fraction was analyzed. F, quantification of m-αSyn in fractions corresponding to short fibrils (sum of fractions 4–6 of the sucrose gradient) and long fibrils (fraction 9 of the sucrose gradient) under strong shaking conditions compared to quiescent conditions. Data are presented as mean of biological replicates ± SD (n = 3). Statistical analysis was performed using an unpaired t test (strong shaking vs quiescent). The proportion of short fibril differed significantly between conditions (p = 0.0137), while long fibrils showed a trend towards significance (p = 0.0505). To further validate this observation, we performed a sucrose gradient centrifugation experiment to separate αSyn species based on their size and sedimentation properties. According to previous published protocols, this method allows the separation of long and short fibrils as well as monomeric αSyn (34, 35). We applied this approach to m-αSyn fibrils generated under strong shaking and quiescent conditions. Following centrifugation, nine fractions for each gradient were analyzed by SDS PAGE followed by immunoblotting (Figs. 2, E and F, and S3). m-αSyn monomers were used as a control, where most of the monomers were found in the first three fractions (Fig. 2E). To avoid potential contamination of m-αSyn monomers in the samples, we defined the small fibril proportion as fraction 4 to 6. The data revealed that a significantly higher proportion of shorter fibrils (fractions 4–6) is present under strong shaking conditions compared to quiescent conditions (Fig. 2, E and F). In agreement, a greater proportion of m-αSyn long fibrils was measured in quiescent conditions compared to strong shaking conditions, with a trend (p = 0.0505) towards a significant difference (Fig. 2F). Although the difference did not reach statistical significance (p < 0.05), limited by the small (n = 3) number of replicates, the results further support that the nucleation mechanism caused by fibril fragmentation is more dominant under strong shaking compared to gentle shaking conditions. Notably, insights into the underlying nucleation process can also be deducted from aggregation kinetics analyses where a mechanistic dependence of the aggregation half time on the protein concentration, characterized by the scaling factor γ (describing the slope in a double-logarithmic plot), reflects the monomer-dependence of the mechanism (36). In theory, a surface-catalyzed secondary nucleation–dominated mechanism has a γ-value of −1.5, while a fragmentation-based process exhibits a slope of −0.5 (25, 36). However, in more complex scenarios where secondary nucleation is saturated and independent of the monomer concentration, the γ-value is shifted toward −0.5 (37). In our experiments, the scaling factor γ is close to −0.5 (37) for both human and mouse αSyn (Fig. S4), making these two processes indistinguishable from this analysis. In agreement, a previous study also reported a proportionality factor of −0.5 for h-αSyn, but, importantly, the authors could assign the underlying mechanism to surface-catalyzed nucleation using complementary microfluidics and TEM images analysis (28). Taken together, based on the present data and previous reports, we conclude that h-αSyn is dominated by surface-catalyzed secondary nucleation, while this mechanism is shifted toward fragmentation for m-αSyn under high shaking conditions. The change of the predominant underlying nucleation mechanism hence impairs the ability of BRICHOS to function as an amyloid inhibitor. BRICHOS binds to m-αSyn and h-αSyn fibrils with binding affinities of the same magnitude To assess BRICHOS binding to m-αSyn fibrils, we performed flow-induced dispersion analysis (FIDA) and surface plasmon resonance (SPR) experiments. FIDA measures the diffusion behavior of fluorescently labeled proteins, here BRICHOS-Alexa488, within a microfluidic channel, allowing the calculation of their hydrodynamic radius (Rh), where binding with an unlabeled partner leads to a change in hydrodynamic radius. FIDA experiments revealed that the Rh of BRICHOS-Alexa488 remained unchanged across varying concentrations of m-αSyn monomers, suggesting no interaction between BRICHOS and monomeric m-αSyn (Fig. S5), as previously found for h-αSyn monomers (8). However, in the presence of m-αSyn fibrils, BRICHOS-Alexa488 formed higher molecular weight complexes, indicating binding to the fibrils. The binding affinity was estimated based on the shift in fluorescence, which revealed a KD value of 42.8 ± 3.5 μM (Fig. 3A). This is a notably lower affinity than the measured KD value of 2.4 ± 0.2 μM, which in agreement with the previously reported KD value of 2.3 ± 0.2 μM, for BRICHOS binding to h-αSyn fibrils (8). Of note, the shape of the curves did not allow calculating the hydrodynamic radius, which would yield a more precise determination of the KD value. The current measurements may therefore be considered as estimates, which are in the same order of magnitude. To further confirm the interaction, we performed SPR with m-αSyn fibrils immobilized on a sensor chip (Fig. 3B). The profile of the SPR curves fitted best to a heterogenous ligand-binding model. This analysis resulted in a weak and strong binding dissociation constant of 65 ± 14 μM and 25 ± 15 nM, respectively, in a similar range, in particular for the strong binding, as the reported affinity of BRICHOS binding to h-αSyn fibrils of 350 ± 60 μM and 22 ± 2.0 nM (8), regardless of slightly different experimental conditions used to produce m-αSyn and h-αSyn fibrils. Of note, here two technical replicates were used, where the errors were derived from the range of the individual values of the replicates, while the errors reported for h-αSyn correspond to fitting errors from the model (8), which are given in Table S1 for m-αSyn. A limitation of SPR is the potential appearance of artifacts due to the immobilization of large heterogenous aggregates on the chip. In contrast, FIDA measures equilibrium binding directly in solution, however, the large and inhomogeneous size distribution of the fibrils produces curve shapes, which only allowed estimates of the KD value. Overall, these results show that BRICHOS binds with affinities in the same range to both m-αSyn and h-αSyn. Hence, the absence of inhibitory effect of BRICHOS on m-αSyn aggregation in high shaking conditions cannot only be explained by profound differences in binding affinities. BRICHOS binds to mouse αSyn fibrils.A, pre-mix FIDA titration of h-αSyn (8) or m-αSyn fibrils to fluorescently labeled BRICHOS-Alexa488 represented by the fraction of BRICHOS bound to the fibrils (fbound) calculated from max and minimal RFU values, showing a lower affinity of BRICHOS to m-αSyn (42.8 ± 3.5 μM) than h-αSyn (2.4 ± 0.2 μM). Error bars represent standard deviation from three technical replicates. B, SPR measurements of one replicate showing BRICHOS binding to m-αSyn fibrils in PBS. A two-phase profile was fitted (replicate 2 is shown in Fig. S5), resulting in a weak and strong dissociation constant of 65 ± 14 μM and 25 ± 15 nM, respectively. The stated KD refers to mean ± half-range of two technical replicates. Discussion This study demonstrates that the BRICHOS domain from the Bri2 protein can inhibit the aggregation of m-αSyn in a mechanism-selective manner. Our findings reveal, for the first time, experimental conditions under which BRICHOS does not inhibit amyloid formation, highlighting the importance of studying aggregation systems where secondary nucleation is not the dominant pathway. While under strong shaking conditions, BRICHOS fails to inhibit m-αSyn aggregation, BRICHOS efficiently slows down the fibrillation process under gentle shaking and quiescent conditions (Fig. 1). Similarly as previously reported for h-αSyn (under strong shaking conditions), BRICHOS predominately affects surface-catalyzed secondary nucleation in addition to fibril-end elongation (Fig. 1). This illustrates how subtle changes in experimental conditions can shift the dominant aggregation mechanism. The binding affinities of BRICHOS to m-αSyn and h-αSyn fibrils, as determined by FIDA and SPR data (Fig. 3), are in a similar range, indicating that the sequence difference between the two variants does not likely influence BRICHOS–fibril interaction. This observation is consistent with the previous observations that the BRICHOS domain recognizes shared structural features of amyloid fibrils, that is β-sheet structure, rather than relying on a specific amino acid sequence (24, 38). In particular, the apparent generic ability to bind to and target very different amyloidogenic proteins, e.g. different Aβ isoforms and disease-related mutants (3, 4, 5), IAPP (6, 39), h-αSyn (8, 9), and Tau (7), appears to suggest a common inhibitory mechanism based on binding to amyloid fibrils and sensing catalytic sites on the fibril surface. For Aβ42, the binding site has been identified by solid-state NMR and constitutes of the three C-terminal β strands (29). Remarkably, only a small proportion of BRICHOS is bound at the fibril surface, which is a common feature for Aβ42, Tau, and h-αSyn fibrils (7, 8, 29). For Aβ42-specific aggregation, hotspots were suggested to be blocked by BRICHOS (29). For human and mouse αSyn as well as for Tau, high-resolution data are currently lacking, and it hence remains speculative whether similar binding mechanisms as for Aβ42 are present. The scenario of specific secondary nucleation sites is supported by a recent study providing evidence for structural defects on the Aβ fibrils that enable secondary nucleation (40). An alternative mode was suggested by a different study where proSP-C BRICHOS was reported to bind to the fuzzy C-terminal tail of αSyn fibrils, which was based on competition experiments with monomeric αSyn for the same binding site, and sweep along the fibril axis, thereby preventing monomeric αSyn undergoing secondary nucleation (9). The experimental data from both studies would also fit to a binding mechanism, where BRICHOS binds to the fibril core thereby inducing conformational changes in the C-terminal fuzzy tail, which is more consistent with the generic ability of BRICHOS to bind different amyloid fibrils. Interestingly, the structures of fibrils from S20G IAPP mutant (41), αSyn (42), and truncated Tau (43) (residues 297–391, under conditions resembling the AD fold) were shown to change and maturate over time. This arises the intriguing question whether the BRICHOS binding efficiency may vary during fibril maturation. We conclude that the difference in BRICHOS’ inhibitory effect on m-αSyn vs h-αSyn is presumably not caused by conceptionally different binding mechanisms. Instead, the origin of the different BRICHOS efficiency might be found in different nucleation processes and fibril stabilities. Indeed, analysis of fibril length distributions by TEM and sucrose gradient experiments suggest that m-αSyn fibrils are more fragile and, thus, more prone to break into smaller fibrils (Fig. 2). In support, a recent study reported significant difference in the molecular structure of m-αSyn fibrils compared to h-αSyn fibrils, which made them more sensitive to fragmentate in sonication and detergent sensitivity experiments (30). The authors identified that different β-fold stacking arrangements in m-αSyn fibrils cause lower tensile strength and resilience (30). The role of the m-αSyn residue differences to h-αSyn in the reduced stability of m-αSyn fibrils is still unclear. Both A53T and S87N accelerate aggregation possibly by disrupting long-range interactions between N- and C-terminal regions, weakening the normally protective role of the C-terminus (21, 44). Furthermore, the A53T mutation was reported to accelerate the aggregation kinetics and the combination of both A53T and S87N mutations drastically reduces the lag time, pointing out that these residue differences, in particular A53T, as decisive differences between h-αSyn and m-αSyn (23). While BRICHOS can still inhibit the A53T and A53T S87N mutants, the effectiveness drastically decreases in a similar manner for both mutants, supporting that the A53T residue replacement is a key difference between m-αSyn and h-αSyn. These residue differences, found in the fibril core, likely reduce fibril stability and promote the formation of different fold and β-strand packing, resulting in more pronounced fibril fragmentation and reduced BRICHOS efficiency. However, also additional contributions from the C-terminal sequence differences in m-αSyn likely influence how the protein aggregates since the inhibitory effect is even more reduced for m-αSyn compared to these single or double point mutants of h-αSyn. Taken together, our findings support a model (Fig. 4), in which strong shaking drives different nucleation mechanisms for mouse and human αSyn. h-αSyn aggregation remains surface-catalyzed secondary nucleation dominated, while m-αSyn aggregation becomes fragmentation dominated. Notably, both mechanisms can exhibit indistinguishable dependencies of the scaling factor γ, with a value of around −0.5, when surface-catalyzed secondary nucleation is saturated and this effect that was previously demonstrated for h-αSyn (28). Under reduced agitation, m-αSyn aggregation shifts towards monomer-dependent secondary nucleation, where BRICHOS is effective. Due to the low population of bound BRICHOS, a stabilization of the fibril by BRICHOS is mechanistically not likely. Hence, BRICHOS potently suppresses secondary nucleation but cannot counteract fragmentation, explaining the selective inhibitory effect. For other, yet unexplored, amyloids, such as TDP-43 or superoxide dismutase 1 (SOD1), the origin of the underlying nucleation mechanism may hence determine the efficiency of BRICHOS in these systems. Mechanism-of-action of BRICHOS inhibition of αSyn aggregation under different conditions. Schematic representation of the aggregation pathways for h-αSyn (blue) and m-αSyn (red) under varying shaking conditions. For h-αSyn, surface-catalyzed secondary nucleation (k2) is the dominant mechanism and BRICHOS (green) effectively inhibits aggregation by binding to the fibril surface (8). In contrast, for m-αSyn, under strong shaking conditions, fragmentation (k−) is the dominant mechanism, which cannot be inhibited by BRICHOS. However, under gentle shaking or quiescent conditions, fragmentation is reduced, and surface-catalyzed secondary nucleation becomes the dominant pathway, allowing BRICHOS to inhibit aggregation of m-αSyn. Other anti-amyloid chaperone-like proteins and molecular chaperones have been shown to exhibit inhibitory effects against amyloid formation, yet, their mechanism-of-action is typically different to BRICHOS’ effect, targeting other nucleation reactions such as primary nucleation or fibril-end elongation (45, 46). A recent study using the molecular chaperones DNAJB6b (also called HSP40) demonstrated that it co-aggregates with αSyn, thereby suppressing aggregation (47). It is also important to note that in vitro conditions do not replicate physiological conditions and how the nucleation mechanism translates to the in vivo situation is not known. However, a report showed that the in vitro aggregation mechanism of recombinantly produced ALS-associated protein SOD1 can be transferred to more complex in vivo conditions using transgenic mice (48). Additionally, in contrast to the in vitro mechanical forces induced by shaking in our experiments, degradation or fragmentation of fibrils in vivo are likely influenced by other factors such as proteolytic enzymes, heat shock proteins (HSP) and other molecular chaperones, which has been comprehensively reviewed (49, 50, 51). Remarkably, HSP chaperones were found to disaggregate α-synuclein fibrils (52), thereby potentially increasing the number of small fibrillar units, a scenario similar to fragmentation. In vitro studies showed that fragmentation increased the ability of fibrils to penetrate cell membranes (53) and promoted fibrillation and propagation of the pathology in preformed fibril (PFF)-based mice models (54). However, it is still unknown whether fragmentation is a natural process that contributes to propagation of the amyloid fibrils and their seeding efficiency in vivo. Nevertheless, BRICHOS has an overall stronger effect on h-αSyn than m-αSyn in vitro, suggesting that these differences should be considered for future in vivo studies. Although direct translation remains to be established due to the complex in vivo situation, the current findings can help to design BRICHOS-based in vivo treatment studies, which may preferentially choose animal models that are not based on endogenous m-αSyn or initiated by injections of PFFs from m-αSyn. In conclusion, our findings reveal that the efficacy of Bri2 BRICHOS as an amyloid inhibitor depends on the contribution of different underlying nucleation mechanisms, where BRICHOS targets secondary nucleation pathways in contrast to fibril fragmentation. Moreover, these results are of importance for designing future studies of therapeutic strategies against synucleinopathies, for example in PD mouse models, to increase clinical translational relevance. Experimental procedure Purification of proteins BL21(DE3) Escherichia coli were transformed with the bacterial expression plasmid pET21a-αSynuclein (Addgene) for h-αSyn and pD434-SR (Addgene) for m-αSyn. Mouse and human αSyn were purified according to a published protocol (8). In short, cells were grown to optical density at 600 nm of0.8 at 37 °C, induced with 0.5 mM IPTG, incubated overnight at 20 °C, and harvested. Pellets were sonicated, boiled for 5 min, centrifuged, and the supernatant was purified on a HiTrap QFF anion-exchange column. Reverse phase chromatography was performed (0.1% trifluoroacedic acid, eluting with 70% acetonitrile), and elution fractions were lyophilized. Approximately 8 mg of proteins was resuspended in 7 M Gdn–HCl and further purified on a Superose 6 Increase or Superdex 75 Increase column to obtain monomeric αSyn using either 20 mM NaP, pH 7.4 or PBS, pH 7.4. The samples were stored at −20 °C or directly used for kinetics experiments. Bri2 BRICHOS monomers were purified according to a published protocol (8) In brief, recombinant R221E Bri2 BRICHOS was expressed in SHuffle T7 E. coli transformed with a plasmid encoding residues 113 to 231 fused to Gly-His6-NT∗, with a thrombin-cleavable site. Cells were grown at 30 °C to optical density at 600 nm of0.8, induced with 0.5 mM IPTG, incubated overnight at 20 °C, harvested, and frozen in 20 mM Tris pH 8. Pellets were thawed with protease inhibitor, sonicated, clarified at 24,000g, and the supernatant was purified on a Ni-NTA column (wash: 15 mM imidazole; elution: 300 mM imidazole). Pooled fractions were dialyzed with thrombin to remove the NT∗ tag, then passed over Ni-NTA. The flowthrough was concentrated, heat-treated at 60 °C for 10 min, and finally purified by size-exclusion chromatography to isolate monomeric BRICHOS using 20 mM NaP, pH 7.4. The protein was concentrated using a 10 kDa cut-off filter and buffer exchanged, when necessary, to PBS pH 7.4. The last step of SEC was performed using 20 mM NaP, pH 7.4. The protein was concentrated using a 10 kDa cut-off filter and buffer exchanged, when necessary, to PBS. ThT aggregation assays The aggregation kinetics were performed at 37 °C using a 384-well plate, using 2 glas beads of 1 mm per well, where each sample contained 20 μM ThT, 30 or 70 μM of h-αSyn or m-αSyn, and different concentration of BRICHOS ranging from 2% to 70% BRICHOS or 2% to 40% BSA (in molar equivalent of αSyn monomers). For the scaling factor γ for m-αSyn and h-αSyn, ranges of concentration were used (from 5 to 140 μM). The buffer used was either 20 mM NaP, pH 7.4 or PBS. Shaking conditions varied depending on the experiment. Strong shaking conditions were defined as 9 min ON, 1 min OFF of orbital shaking at 300 rpm while gentle shaking conditions were 1 min ON, 9 min OFF and quiescent conditions correspond to no shaking. For seed preparation, m-αSyn monomers were incubated in Eppendorf tubes and shaken during 7 days at 37 °C at 600 rpm or prepared in the plate reader similarly as for ThT kinetics. The fibrils were then washed 2 times using the appropriate buffer and sonicated using a probe sonicator (SONICS Vibra Cell), with 2-s ON and 2-s OFF intervals at 20% amplitude. Analysis of the ThT aggregation kinetics Aggregation profiles were normalized, combined by taking the median across six different replicates and truncated at the plateau phase of the aggregation process. Kinetic analysis was performed using the AmyloFit online platform (32) applying a model dominated by secondary nucleation. This model incorporates rate constants for primary nucleation, secondary nucleation, and fibril elongation. To assess the impact of individual nucleation steps, two nucleation rate constants were treated as global fit parameters (which were kept constant across all BRICHOS concentrations), while the remaining constant (kn, k2 or k+) was fitted individually. Normalized mean square error values were calculated for each fitting. The dependence of k+ and k2 on the BRICHOS concentration was determined using an unseeded secondary nucleation model using the combined fitting parameter k+k2, where the other combined constants knk+ was fixed globally. For highly seeded reactions, aggregation kinetics were analyzed using GraphPad. Elongation rates were obtained by linearly fitting the first 48 or 60 min of the aggregation curves, using normalized, averaged, and truncated data. For the scaling factor γ analysis, the aggregation half-time (thalf) was plotted against the concentration of αSyn used. The scaling factor γ was then extracted through the fitting of a straight line in double logarithmic axes, including standard fitting errors. Negative-stain TEM analysis m-αSyn monomers were aggregated using a plate reader under the same conditions as the ThT aggregation assay with 70 μM m-αSyn, 1% seed, in PBS with two different shaking conditions: strong shaking (9 min ON, 1 min OFF of 300 rpm orbital shaking) and quiescent. h-αSyn monomers were aggregated in the same conditions without seeds using strong shaking conditions. Aggregation was monitored until a plateau was reached, approximately 70 h for the strong shaking conditions and 90 h for the quiescent conditions to reach the plateau. The resulting fibrils were pelleted by centrifugation at 16,200g for 30 min, followed by two washes with equal volume of supernatant with fresh PBS buffer and an additional wash with 6 mM HCl to reduce fibril clumping. The fibrils were then pelleted again and resuspended in the resulting supernatant. Grid preparation for subsequent analysis was carried out following a published protocol using 70 μM for h-αSyn and 10 μM for m-αSyn for grid preparation (8). Sucrose gradient m-αSyn fibrils were prepared in the same conditions as described for the TEM analysis using both strong shaking and quiescent conditions. The reactions were stopped at 70 h for both conditions. The resulting fibrils were washed twice with fresh PBS, pH 7.4 and 100 μl was applied to 3.5 ml of a 15 to 60% sucrose gradient and centrifuged at 86,419 g (35,000 rpm) for 2.5 h at 4 °C. For m-αSyn monomers, 100 μl of 70 μM were applied on the sucrose gradient. Following centrifugation, nine fractions of 388 μl each were manually collected from top to bottom. Each fraction was analyzed by SDS PAGE and immunoblotting (using the antibody ab155038 anti-alpha-synuclein). Glycerol was omitted from the SDS-sample buffer. We quantified relative amounts of αSyn in each fraction by immunoblotting using ImageJ software and data was represented as triplicates using three independent experiments except for m-αSyn monomers where one replicate was used as the control. Normality was visually assessed using QQ plots. As the data were approximately normally distributed, an unpaired t test was used for statistical analysis. Flow-induced dispersion analysis BRICHOS was labeled with Alexa Fluor 488 by reacting 43.6 μM protein in 100 mM sodium bicarbonate buffer (pH 8.3) with Alexa Fluor 488 NHS ester (Thermo Fisher Scientific) in DMSO at a 1:2 M ratio for 1 h at room temperature, followed by removal of free dye using a PD-10 column (GE Healthcare) equilibrated in PBS. FIDA experiments were performed on a FIDA One instrument (FIDA Biosystems) using fluorescence detection (λex = 480 nm) and a 75-μm inner diameter capillary, which was rinsed with 1 M NaOH and equilibrated in PBS. Measurements were conducted at 25 °C after priming the capillary with buffer at 3500 mbar. Monomeric or fibrillar α-synuclein (human or mouse) was titrated at varying concentrations against 50 nM Alexa488-labeled BRICHOS using the standard injection protocol allowing in-capillary mixing of analyte and indicator (55). Fibrils were fragmented prior to analysis by probe sonication (QSONICA Q500) for 90 s (20 s ON, 10 s OFF, 20% amplitude). Resulting taylograms were analyzed using FIDA software v2.3 (FIDA Biosystems) to obtain hydrodynamic radii (Rh). Binding to fibrillar species was quantified from shifts in BRICHOS-Alexa488 fluorescence, and the bound fraction (fbound) was calculated by normalizing RFU between minimum and maximum signals and fitted to fbound=[A]KD+[A], where [A] is the analyte concentration (analyzed by OriginPro, OriginLab). Surface plasmon resonance The interaction between BRICHOS and m-αSyn was performed using SPR on a BIAcore T200. m-αSyn fibrils were sonicated using a sonicator bath for 5 min at 30% amplitude. The immobilization was performed via amino coupling method as recommended by the manufacturer (Biacore) on the sensor chip CM5 at 108 μg/ml of m-αSyn fibril (monomer equivalent), at pH 4.0. The binding analysis was performed at 25 °C. The system was primed with PBS, pH 7.4. The different concentration of BRICHOS (0, 0.091, 0.195, 0.391, 0.781, 1.56, 3.125, 6.25, 12.5, 25, 50, and 100 μM) were injected for 5 min at a flow rate of 20 μl/min. Channels 1 was left empty and served as a reference to control for possible nonspecific binding of BRICHOS. Each sample was run in duplicates. The analysis consisted of first adjusting the sensor grams to zero and removing buffer spikes as well as subtracting the empty flow cell signal as well as the bulk effect of the injection of 0 μM of BRICHOS. Between two measurements, the chip surface was regenerated with 20 mM NaOH. Data on SPR sensorgrams were evaluated using GraphPad. The curves from the two technical replicates were then fitted with a two-phase association and dissociation model using GraphPad Prism with the association rate determined by RAsso(t) and the dissociation rate by RDiss(t). where kon and koff refer to the on- and off-rate constants, kobs1 and kobs2 are the observed association rate constants, c is the BRICHOS concentration, Bmax1 & Bmax2 and R0 are the amplitudes for association and dissociation, respectively, and a is a value between 0 and 1, describing the contribution of the two dissociation phases. The fast and slow apparent KD values were calculated as a ratio of the corresponding dissociation rate constants and association rate constants. The given errors correspond to the half-range between the two technical replicates. Data availability All data are available in the paper or as supporting information. Additional raw data can be obtained from the corresponding author upon request. Supporting information This article contains supporting information (8, 32). Conflict of interest The authors declare that they have no conflicts of interest with the contents of this article. Acknowledgments The authors acknowledge financial support from the Swedish Research Council (A. A., J. J.), the Swedish Alzheimer foundation (A. A.), the Swedish Brain Foundation (A. A., J. J.), StratNeuro (A. A.), CIMED (A. A.), FORMAS (A. A.), the Åke Wiberg foundation (A. A.), the Magnus Bergvall foundation (A. A.) and the Åhlen foundation (A. A.). D. E. O. is funded by the Lundbeck Foundation (grant R453-2024-359 to the Collaborative Project NanoPANS). Author contributions L. A. and A. A. writing–original draft; L. A. visualization; L. A., W. H. M., J. S. N., K. G., R. K., and J. N. investigation; L. A., W. H. M., J. S. N., A. F., and A. A. formal analysis; W. H. M., J. S. N., A. F., K. G., R. K., J. N., D. E. O., and J. J. writing–review and editing; D. E. O., J. J., and A. A. supervision; D. E. O., J. J., and A. A. funding acquisition; J. J. and A. A. conceptualization; A. A. project administration. Reviewed by members of the JBC Editorial Board. Edited by Ursula Jakob Supporting information Supplementary Material References Associated Data Supplementary Materials Supplementary Material Data Availability Statement All data are available in the paper or as supporting information. Additional raw data can be obtained from the corresponding author upon request.