Amyloid Fibril Formation of Arctic Amyloid-β 1-42 Peptide is Efficiently Inhibited by the BRICHOS Domain.

Amyloid-β peptide (Aβ) aggregation is one of the hallmarks of Alzheimer's disease (AD). Mutations in Aβ are associated with early onset familial AD, and the Arctic mutant E22G (Aβarc) is an extremely aggregation-prone variant. Here, we show that BRICHOS, a natural anti-amyloid chaperone domain, from Bri2 efficiently inhibits aggregation of Aβarc by mainly interfering with secondary nucleation. This is qualitatively different from the microscopic inhibition mechanism for the wild-type Aβ, against which Bri2 BRICHOS has a major effect on both secondary nucleation and fibril end elongation. The monomeric Aβ42arc peptide aggregates into amyloid fibrils significantly faster than wild-type Aβ (Aβ42wt), as monitored by thioflavin T (ThT) binding, but the final ThT intensity was strikingly lower for Aβ42arc compared to Aβ42wt fibrils. The Aβ42arc peptide formed large aggregates, single-filament fibrils, and multiple-filament fibrils without obvious twists, while Aβ42wt fibrils displayed a polymorphic pattern with typical twisted fibril architecture. Recombinant human Bri2 BRICHOS binds to the Aβ42arc fibril surface and interferes with the macroscopic fibril arrangement by promoting single-filament fibril formation. This study provides mechanistic insights on how BRICHOS efficiently affects the aggressive Aβ42arc aggregation, resulting in both delayed fibril formation kinetics and altered fibril structure.

Introduction

Proteins and peptides can self-assemble into fibrillar, cross β-sheet structures (commonly referred to as amyloid) that are relevant for about 40 human diseases including the neurodegenerative Alzheimer’s disease (AD).[1,2] AD is the most prevalent form of dementia, and so far, only the monoclonal antibody aducanumab has been approved for disease-modifying treatment by the US Federal Drug Administration, yet the reported effects are relatively minor.[3] Several observations support that amyloid-β peptide (Aβ) aggregation initiates AD development, whereof Aβ 1–42 peptide (Aβ42) is the most aggregation prone and toxic variant.[4] Familial, early onset AD is linked to mutations in the γ-secretase components presenilins 1/2 and the amyloid precursor protein, that is subjected to sequential cleavages by the β- and γ-secretases eventually generating the Aβ peptide.[5,6] Among the familial mutations, the Arctic mutant E22G (Aβ42arc) is not only the most aggregation-prone variant,[7] but it is also associated with aggressive early onset AD and rapid plaque deposition in the brain,[8] while the pathogenic mechanisms are still largely unclear.

The wild-type Aβ42 (Aβ42wt) fibrillates into nanoscale amyloid fibrils following nucleation-dependent microscopic events:[9] Aβ42 monomers associate and form a nucleus (primary nucleation), from which a fibril can start to elongate (elongation). Aβ42 monomers also can attach to the fibril surface and subsequently form a new nucleus (secondary nucleation) that further elongates to a fibril. The monomer-dependent fibril surface catalyzed secondary nucleation pathway is the main source of toxic Aβ42 species.[10] The Aβ42arc peptide follows a similar fibrillization mechanism as Aβ42wt, but the surface-catalyzed secondary nucleation process needs to be treated as a multistep process as the secondary nucleation is saturated.[7] Aβ42arc forms amyloid fibrils with a much faster rate compared to Aβ42wt; however, in vitro mature fibrils from both variants, from hundreds of nanometers to a few micrometers long and 5 to 10 nm thick, share similar morphology with a twisted structure, and can form large fibril bundles.[7] Recently, cryo-electron microscopy (cryo-EM) structure of Aβ amyloid fibrils from AD brain tissue showed fibrils that are polymorphic with three abundant morphologies.[11] Interestingly, different types of fibril arrangements have been observed from the brain of individuals with sporadic and familial AD, respectively.[12] In vitro, for generating homogeneous Aβ42 fibrils, several generations of seeding are normally applied,[13,14] and the Aβ42 fibrils were shown to be composed of two molecules per fibril layer, where residues 1–14 are only partially ordered and residues 15–42 form a cross-β-sheet entity with hydrophobic side chains maximally buried.[14] Without seeding, highly homogeneous Aβ42 fibrils were formed, which are unbranched, micrometer-long, and most of the fibrils showed a rather uniform diameter of about 7 nm.[13−15]

Molecular chaperones can prevent proteins from aggregating and exerting cytotoxic effects,[16] and several chaperones have been shown to interfere with amyloid formation but with different microscopic mechanisms.[17] One example is the BRICHOS domain that has been established as a molecular chaperone domain active against amyloid fibril formation and toxicity of peptides associated with severe human diseases.[18−20] We have shown that the recombinant human (rh) BRICHOS domain from familial dementia-associated Bri2 protein is efficient in inhibiting both Aβ42wt amyloid fibril formation and neurotoxicity.[19,21−23] How the BRICHOS domain interferes with familial Aβ mutants with more aggressive amyloid-forming propensity, like the arctic Aβ42 mutant (Aβ42arc), remains to be elucidated.

Here, we report a protocol for the recombinant preparation of Aβ42arc with high quality and yield and show the inhibition effect of rh Bri2 BRICHOS on Aβ42arc fibrillization kinetics and its modulation effect on the fibril morphology. The results further elucidate the aggregation properties of Aβ42arc and supply a basic understanding for the effects of BRICHOS on Aβ42arc fibril formation.

Results

Recombinant Preparations of Aβ42arc, Aβ42wt, and Tev Proteinase

First, we set out to establish an efficient and robust protocol for recombinant production of Aβ42arc. The N-terminal globular domain (NT) of major ampullate spider silk protein (MaSp) was genetically modified, referred to as NT*Masp, and implemented as a solubility tag for producing different problematic proteins and peptides.[24−30] In the recent protocol, we applied NT* derived from flagelliform spider silk protein (FlSp), NT*FlSp, which is more soluble than NT*Masp, to generate recombinant Aβ42wt.[31] Here, we follow a modified protocol without using urea, which might induce potential modifications to the final product.[32] Recombinant NT*FlSp-Aβ42wt and NT*FlSp-Aβ42arc were expressed in Escherichia coli, and the Ni-NTA column purified fusion proteins were subsequently cleaved by Tobacco etch virus (Tev) protease to release the tag-free Aβ42wt and Aβ42arc peptides without any extra amino acid residues (Figures a,b and S1a). The Aβ42wt and Aβ42arc monomers were isolated via size exclusion chromatography (SEC), which showed good quality in terms of purity (Figures c,d and S1b). To obtain pure Aβ42arc monomers for kinetic analysis, the SEC-isolated [superdex30 column (26/600)] monomers were lyophilized, solubilized with guanidium chloride, and isolated again by SEC using a superdex30 column (10/300), which showed very well-separated monomer and oligomer peaks (Figure d), indicating that a single SEC isolation is not enough to obtain pure monomeric Aβ42arc. Although Aβ42arc is highly prone to form amyloid aggregates and significant losses are observed during Ni-NTA column purification, the final yield of the monomeric Aβ42arc was up to ∼5 mg per liter LB medium. Tev proteinase used in this study was expressed in E. coli fused to the NT*FlSp tag, and the soluble fusion protein was purified by Ni-NTA chromatography (Figure S2a). The final yield of NT*FlSp-Tev reached 145 mg per liter LB medium and showed high purity (Figure S2b). The NT*FlSp-Tev fusion protein presented very good cleavage efficiency against NT*FlSp-Aβ42wt. The cleavage reaction was performed in the cold room at an enzyme to a substrate ratio of 1:100 (w/w) where the half-time for cleavage was estimated to be ∼3.2–4.1 h (Figure S2c–f). No visible protein aggregation was seen, and no aberrant degradation appeared as judged by SDS-PAGE (Figure S2c), indicating that fusion to NT*FlSp tag can enhance the stability of Tev and does not impair Tev activity.

Figure 1

Preparation of recombinant human Aβ42arc and Aβ42wt peptides using the NT*FlSp tag. (a) Schematic presentation of NT*FlSp-Aβ42arc and NT*FlSp-Aβ42wt. The Tev cleavage site is located immediately before Aβ42, which generates recombinant Aβ42 peptides without extra amino acid residues. The structure model of NT*FlSp is derived from the NMR structure of NT at pH 7.2 (PDB 2LPJ). (b) Amino acid sequence of human Aβ42arc and Aβ42wt. The arrow points to the mutated amino acid residue (E22G). (c) Chromatogram of recombinant Aβ42wt on a Superdex30 26/600 column. The shadowed area indicates the fraction collected for monomeric Aβ42wt species. (d) Chromatogram of recombinant Aβ42arc on an analytical Superdex30 10/300 column. The shadow area indicates the fraction collected for monomeric Aβ42wt species. The inset shows the SDS-PAGE analysis of final monomeric Aβ42arc and Aβ42wt.

Aβ42arc and Aβ42wt Aggregation and Kinetics

To compare the aggregation kinetics of Aβ42arc and Aβ42wt, we used thioflavin T (ThT)[33] to monitor the fibrillization kinetics as a function of time at a range of different initial monomer concentrations. Both Aβ42arc and Aβ42wt showed typical sigmoidal aggregation kinetics (Figures a and S3a), and the fibrillization half-time, τ1/2, increased with decreasing monomer concentrations, while the maximum rate of aggregation, rmax, decreased (Figure b,c), indicating a dose-dependent aggregation behavior for both Aβ42 variants. As indicated by rmax and τ1/2, Aβ42arc exhibited significantly faster aggregation than Aβ42wt (Figure b,c), in line with a previous report using an Aβ42arc variant with an additional methionine at position zero, that is, Met-Aβ42arc7. The dependence of the τ1/2 on the initial monomer concentration, m0, is captured by τ1/2 ∼ m0γ, where γ is the scaling exponent related to the reaction order (i.e., to the monomer dependence of the dominant processes) for each of the kinetics models and can be used to indicate the dominant mechanism of aggregation.[34] The aggregation half-time and the initial monomer concentration were plotted on a double logarithmic scale, and Aβ42arc showed a γ value of −0.8 ± 0.1, while for Aβ42wt, it was −1.4 ± 0.1 (Figure b), similar to the γ values determined in previous studies.[7,21,23,31] This indicates a multistep secondary nucleation and a secondary nucleation dominated pathway for the fibrillization of Aβ42arc and Aβ42wt, respectively. Aβ42 fibrillization kinetics can be described by a set of microscopic rate constants, that is, for primary (k) and secondary nucleation (monomer-dependent, k2) as well as elongation (k+),[34] and the combined rate constants for primary and for secondary pathways, respectively.[35−37] Global fitting with combined rate constants and showed that Aβ42wt aggregation traces could be sufficiently described by secondary nucleation dominated models (Figure S3a,b), whereas Aβ42arc traces were fitted with an additional Michaelis constant of 0.96 μM (Figures a and S3c), indicating that saturation of secondary nucleation applies to Aβ42arc fibrillization. The global combined rate constants and of Aβ42arc aggregation traces were 2.3 and 6.0 times higher, respectively, than that for Aβ42wt, indicating that the Arctic mutation accelerates Aβ42 peptide aggregation through predominantly secondary pathways. To further investigate the relationship between the initial monomer concentration and the final fluorescence intensity, the final intensities were plotted as a function of the initial monomer concentrations, which exhibited a linear relationship for both Aβ42arc and Aβ42wt (Figure d). Notably, there was a striking difference regarding the final ThT fluorescence intensity between Aβ42arc and Aβ42wt fibrils, where Aβ42arc showed much lower final intensity than Aβ42wt (Figure d), which probably indicates different fibril morphologies.

Figure 2

Kinetic analysis of Aβ42arc fibril formation. (a) Global fits (solid lines) of aggregation traces (dots) at different Aβ42arc peptide concentrations from 1.0 μM (dark red) to 4.0 μM (gray) with a multistep secondary nucleation dominated (unseeded) model. Best fitting parameters: = 41.0 ± 1.4 M–1 s–1, = 1.8 × 106 ± 0.1 × 106 M–3/2 s–1, and = 0.96 ± 0.06 μM. Fitting residuals are shown in Figure S3c. (b) Both Aβ42arc and Aβ42wt exhibit linear dependence of the aggregation half-time, τ1/2, on the initial peptide monomer concentration; however, the γ-exponent values are different with −1.4 ± 0.1 for Aβ42wt peptide and −0.8 ± 0.1 for the Aβ42arc peptide, indicating a secondary nucleation dominated and a multistep secondary nucleation pathway, respectively. (c) Linear dependence of the aggregation maximum rate (rmax) of Aβ42arc and Aβ42wt on the initial peptide monomer concentration, while the rmax saturates at high Aβ42wt concentrations. (d) Linear dependence of final ThT fluorescence intensity of Aβ42wt and Aβ42arc on different starting monomer concentrations from 1.0 to 9.0 μM. The left Y-axis is for Aβ42wt, and the right Y-axis is for Aβ42arc.

Aβ42arc and Aβ42wt Fibril Morphologies

The remarkable difference of the final intensity between Aβ42arc and Aβ42wt fibrils prompted us to image both types of fibrils by transmission electron microscopy (TEM) (Figure ). Under negative-staining TEM, Aβ42wt fibrils were straight and unbranched and displayed clear twisted architecture with two or more intertwined filaments (Figure a–c). There were at least three different crossover distances (twist–twist distances) (Figure a–c), representing polymorphic structures, that have been shown previously.[11,38] The twist body (position I, as shown in Figure h) of Aβ42wt fibrils showed an averaged diameter of 14.4 ± 2.1 nm, while the twist point (position II in Figure h) was around 6.6 ± 1.3 nm, indicating that most of the twisted fibrils were made up with two filaments. Compared to the wild-type fibrils, the Aβ42arc fibrils were curlier (Figure d,e). Interestingly, less obvious twists were observed for the Aβ42arc fibrils and more single filament-like fibrils were visible, but still thick fibrils consisting of multiple intertwined filaments were present (Figure d,e). We classified these fibrils as single-like (S) and multiple (M) fibrils by their appearance. The average diameter for the single-like fibrils was 9.6 ± 2.9 nm, and for the multiple fibrils, it was 18.8 ± 2.8 nm (Figure h), indicating that the multiple fibrils of Aβ42arc are also largely composed by two or more single-like filaments. However, the diameters of the single-like and multiple Aβ42arc fibrils were significantly different from the diameters of the twist point (position II, as shown in Figure h) and the twist body (position I, as shown in Figure h) of Aβ42wt fibrils. Furthermore, the Aβ42arc peptide formed small aggregates with different sizes (15–300 nm along the long axis) (Figure f,g) that were not observed for the Aβ42wt peptide (Figure a–c). This might be one reason for the observed lower ThT intensity of Aβ42arc than Aβ42wt fibrils (Figure d).

Figure 3

TEM of Aβ42arc and Aβ42wt fibrils. (a–c) Representative negative staining TEM images of Aβ42wt fibrils. Three representative morphologies are shown in (a–c), respectively. (d,e) Representative negative staining TEM images of Aβ42arc fibrils. (f,g) Negative staining TEM images of Aβ42arc aggregates. The single back dot in (d) is likely a staining artifact. (h) Characterizations of Aβ42wt fibrils, i.e., the diameter at twist body (I) and the diameter at the twist point (crossover point, II). The left panel is a schematic cartoon for the Aβ42wt fibril. The fibrils were divided into two kinds of fibrils generally, i.e., the multiple and single-like fibrils. The diameters of both types of fibrils were measured and compared to the diameters at twist body (I) and at the twist point (II) of the Aβ42wt fibrils. The data are present as mean ± SEM (****p < 0.0001). The sizes of the scale bars are 100 nm.

BRICHOS Inhibition of Aβ42arc Aggregation

The Rh Bri2 BRICHOS domain has been shown to inhibit amyloid fibril formation of several peptides efficiently, including Aβ42wt peptide,[19,21,23,39] but it is not evident whether BRICHOS has the ability to suppress also Aβ42arc aggregation since its aggregation mechanism is considerably different from Aβ42wt. To evaluate the inhibition effects of rh Bri2 BRICHOS on the fibrillization process of Aβ42arc, monomeric rh Bri2 BRICHOS species were isolated by SEC and added to Aβ42arc. In line with previous studies,[21,23,39] rh Bri2 BRICHOS showed efficient inhibition of Aβ42wt fibrillar aggregation, as indicated by linearly increased τ1/2 and mono-exponentially declined rmax with increased BRICHOS concentrations (Figure S3d,e). Although Aβ42arc showed substantially faster aggregation than Aβ42wt (Figure b,c), rh Bri2 BRICHOS monomers showed dose-dependent inhibition effects on τ1/2 and rmax (Figure a,b). The aggregation traces for both Aβ42wt and Aβ42arc were further analyzed by global fits with combined parameters and to dissect the underlying mechanisms. Using individual fits of a secondary nucleation dominated model, increasing relative rh Bri2 BRICHOS monomer concentration did not change drastically the (for the primary pathway) but decreased the (for the secondary pathway) (Figure S3f), indicating that rh Bri2 BRICHOS monomer mainly interferes with the secondary pathway rather than the primary pathway of Aβ42wt fibril formation, as proposed previously.[21] A similar mechanism but with an additional secondary nucleation saturation effect (a multistep dominated secondary nucleation model) was applied for Aβ42arc in the presence of rh Bri2 BRICHOS monomers. Also for Aβ42arc, a noticeable decrease in compared to was observed (Figure a,c). Furthermore, keeping as the sole fitting parameter could not account for the kinetic behavior, while the traces were sufficiently described when was the only free fitting parameter (Figure S3g,h). These results indicate that rh Bri2 BRICHOS possesses the capacity to suppress Aβ42arc assembly into fibrils, by mainly interfering with the secondary pathway.

Figure 4

Aβ42arc fibril formation and toxic oligomer generation are inhibited by rh Bri2 BRICHOS. (a) Global fits (solid lines) of aggregation traces (dots) of 3.0 μM Aβ42arc with different concentrations of rh Bri2 BRICHOS monomer from 10 to 100% with a multistep secondary nucleation dominated (unseeded) model. Combined parameters and were kept free, and was set to 0.96 μM. (b) Aggregation half-time τ1/2 and the maximal growth rate rmax determined from the fitting of Aβ42arc aggregation traces with different concentrations of rh Bri2 BRICHOS monomers, as shown in (a), and linear and exponential decay fits were applied, respectively. (c) Dependencies of the relative combined rate constants obtained reveal a strong effect of rh Bri2 BRICHOS monomers on secondary (k+k2) but not primary (kk+) pathways. (d) Seeded aggregation traces of Aβ42arc in the presence and absence of rh Bri2 BRICHOS monomer. Seeded aggregation traces of 3 μM Aβ42arc with 0.6 μM preformed Aβ42arc fibrils in the presence of different concentrations of Bri2 BRICHOS monomers. (e) Estimation of the elongation rates (k+) from the highly pre-seeded aggregation kinetics in (d). The elongation rates (k+) of the Aβ42wt are from ref (21). (f) Immuno-EM of Aβ42arc fibrils with rh Bri2 BRICHOS monomer. The samples were treated with a Bri2 BRICHOS antibody and a gold-labeled secondary antibody and characterized by TEM. The size of the scale bar is 100 nm. (g) Simulated nucleation generation rates of Aβ42arc in the absence and presence of different concentrations of rh Bri2 RBICHOS monomers with the parameters from (c,e). (h) With the individual fitting parameters derived from (c) and the elongation rates (k+) from (e), the relative number of Aβ42arc nucleation unit generated in the presence of rh Bri2 BRICHOS monomers at different concentrations was estimated.

To figure out which of the microscopic events are affected by rh Bri2 BRICHOS against Aβ42arc fibril formation, we carried out aggregation kinetics with a high seed concentration. Aggregation traces typically display a concave aggregation behavior under such conditions (Figure d), where the relative elongation rate k+ could be determined by the initial slope.[40] These experiments, interestingly, revealed that the rh Bri2 BRICHOS monomers only slightly affect the elongation rate k+ of Aβ42arc (Figure e), which is qualitatively different from the effects on the Aβ42wt peptide fibril formation where the elongation rate is deceased significantly in a concentration-dependent manner by rh Bri2 BRICHOS.[21] Together with the fitting results using the combined rate constants, these finding suggest that secondary nucleation (k2) of Aβ42arc peptide is primarily blocked by rh Bri2 BRICHOS, and only a small effect is visible on the elongation rate k+.

The immuno-EM observations confirmed that rh Bri2 BRICHOS can bind to the surface of Aβ42arc fibrils (Figure f). Interference with discrete microscopic rates during Aβ42 fibrillization affects differently the generation of nucleation units, which may be the building blocks of toxic oligomers: it is decreased when secondary nucleation (k2) is inhibited, but it is increased when elongation (k+) is blocked.[41] It has been shown that rh Bri2 BRICHOS monomers can reduce nucleation unit generation by 70% during Aβ42wt fibril formation,[23] while the rh proSP-C BRICHOS, mainly blocking the secondary nucleation of Aβ42wt fibrillization, exhibits an efficiency of 80%.[41] To illustrate the generation of nucleation units during Aβ42arc fibrillization in the presence or absence of rh Bri2 BRICHOS monomers (Figure g,h), the time evolution of the fibril-forming rate was evaluated. The nucleation rate, from the individual fits (Figure c) and elongation k+ from the seeding experiment (Figure d,e), was integrated to calculate the number of nucleation units. We found that the generation of nucleation units during Aβ42arc fibrillization is reduced in a dose-dependent manner, and up to 80% in the presence of monomeric rh Bri2 BRICHOS at an equal ratio (in the presence of monomeric rh Bri2 BRICHOS at an equal ratio (Figure h). The results indicate that rh Bri2 BRICHOS monomers inhibiting the secondary nucleation event of Aβ42arc can largely reduce the new nucleation unit generation and thereby potentially toxic oligomers.

BRICHOS Affects Aβ42arc Fibril Arrangement

Rh Bri2 BRICHOS is able to suppress fibrillar aggregation and reduce the neurotoxicity of Aβ42wt by binding to the fibril surface.[21,23] In the current study, the immuno-EM observations showed that rh Bri2 BRICHOS can bind to the surface of the Aβ42arc fibrils (Figure f). The fibrils from Aβ42arc with and without BRICHOS were further analyzed by TEM (Figure a–d). Coincubation of monomeric Aβ42arc and BRICHOS [(Aβ42arc + BRICHOS)] resulted in the fact that more single-like (S) fibrils were observed (Figure e), and the multiple fibrils (M) presented significantly smaller diameters compared to that of the M fibrils of Aβ42arc alone (Figure f). This indicates that a smaller number of fibrils are bundled together in the presence of BRICHOS. Furthermore, the single-like Aβ42arc fibrils (S) with BRICHOS were narrower compared to the Aβ42arc alone fibrils (Figure f). To investigate the effects of BRICHOS on preformed fibrils, rh Bri2 BRICHOS monomer was added to preformed Aβ42arc fibrils [(Aβ42arc)fibril + BRICHOS]. Under TEM (Figure g,h), Aβ42arc fibrils with rh Bri2 BRICHOS monomers displayed large number of short fibrils and oligomer-like assemblies (Figure g,h), and the fibrils were covered with material that could represent BRICHOS (Figure g,h). To further confirm whether BRICHOS can bind to preformed Aβ42arc fibrils, immuno-EM was performed with an anti-BRICHOS antibody, which confirmed the presence of BRICHOS on the surface (Figure i).

Figure 5

TEM of Aβ42arc fibrils in the presence of rh Bri2 BRICHOS. (a–d) Representative negative staining TEM images of (3.0 μM Aβ42arc + 3.0 μM rh Bri2 BRICHOS) co-incubated fibrils. The sizes of the scale bars are 100 nm. (e) Ratio of single-like fibrils in each micrograph, in total for each type of sample, eight micrographs were analyzed. The data are presented as mean ± SEM. *p < 0.05. The sizes of the scale bar are 100 nm. (f) Characterizations of Aβ42arc fibrils in the presence of rh Bri2 BRICHOS. The diameters of the thick and thin fibrils were measured and compared to the diameters without BRICHOS. The data are presented as mean ± SEM (***p < 0.001 and ****p < 0.0001). (g,h) Representative negative staining TEM images of Aβ42arc fibrils incubated with rh Bri2 BRICHOS [(Aβ42arc)fibril + BRICHOS]. The sizes of the scale bars are 100 nm. (i) Immuno-EM of preformed Aβ42arc fibrils incubated with the rh Bri2 BRICHOS monomer [(Aβ42arc)fibril + BRICHOS]. The samples were treated with a Bri2 BRICHOS primary antibody and a gold-labeled secondary antibody and characterized by TEM. The size of the scale bar is 100 nm.

Discussion

In this study, we provide facile protocols for the recombinant preparation of Tev proteinase and Aβ42arc. The protocols can likely be adapted for production of other Aβ mutants and proteinases. The Arctic mutation E22G significantly accelerated the amyloid fibril formation of Aβ42 and gave a different fibril arrangement pattern compared to wild-type fibrils. Rh Bri2 BRICHOS was able to inhibit Aβ42arc fibril formation and oligomer generation as well as affect the fibril arrangement.

While amyloid fibrils formed from various proteins and peptides contain a common cross-β sheet architecture,[42] amyloid fibrils assembled from the same protein and peptide can end up with different morphologies, including varying filament number and arrangements as well as different polypeptide conformations.[38] Altered Aβ42/Aβ40 ratio and deposition of Aβ42 is thought to be a main pathogenic factor in AD. Both Aβ42wt and Aβ40wt can form twisted fibrils, but they show different morphologies, including crossover distance and diameter.[43,44] In the current study, Aβ42wt fibrils with at least three kinds of morphologies and multiple (more than two) intertwined filaments with twists were observed (Figure a–c), whereas the Aβ42arc fibrils were morphologically different (Figure d–g). Notably, a similar fibril morphology as now observed for Aβ42wt with highly twisted structure was observed for Met-Aβ42arc.[7] These results suggest that even small residue differences and/or different preparations might result in significantly different Aβ fibrils. It has been shown that aggregation proceeds more rapidly for Aβ40arc than Aβ40wt, and Aβ40arc fibrils present at least five polymorphs, including both coiled and non-coiled structures. Furthermore, at the end of the lag phase of fibrillization of Aβ40arc, ∼ 3 nm size aggregates with a homogeneous morphology were identified.[45] Here, the arctic mutation also accelerated the overall aggregation of Aβ42, and multiple types of intertwined curly fibrils and more single-like fibrils were found (Figure d,e), supporting the observation that different types of fibril arrangements present in the brain of individuals with sporadic and familial AD, respectively.[12] Different from Aβ40arc, heterogeneous Aβ42arc aggregates formed at the end of the fibrillization reaction, not visible for Aβ42wt during fibril formation (Figure f,g), which might be one reason for the significantly lower final ThT density of Aβ42arc (Figure d). In line with that, Aβ40 showed much higher final ThT intensity compared to Aβ42, which was suggested to be caused by the exposure of β-sheet in Aβ fibrils and hence to differences in fibril morphology.[46] Cytotoxicity can be induced by both Aβ40 and Aβ42, but it has been shown that Aβ42 is more cytotoxic and more directly related to AD pathology.[47] However, together with the data in this study, it is not clear whether or not there is a correlation between the fibril morphology and toxicity.

Molecule chaperones have been shown to interfere with amyloid formation but with different underlying mechanisms[17] for example, DNAJB6 inhibits Aβ42wt fibril formation by interacting with the growing aggregates (oligomer formation during primary nucleation),[48] while proSP-C BRICHOS specifically inhibits secondary nucleation.[41] Recently, Bri2 BRICHOS has been shown to affect both Aβ42wt secondary nucleation and elongation;[21] however, this situation is changed for Aβ42arc, where mainly secondary nucleation but not the elongation was affected (Figure c–e). The molecular chaperone αB-crystallin colocalizes with Aβ amyloid fibrils in the extracellular plaques, binds to Aβ42wt fibrils and fibril ends with micromolar affinity, and inhibits Aβ42 fibril elongation.[49] Additionally, αB-crystallin delays the aggregation of Aβ40wt, favors more disordered aggregates, and hence interferes with ordered amyloid fibril formation.[50] The molecular chaperone BRICHOS binds to Aβ42wt fibrils with nanomolar affinity,[41,51] and here we show that rh Bri2 BRICHOS also affects Aβ42arc fibril formation, binds to the fibril surface, and affects the fibril structure (Figures f and 5). Modulation by molecular chaperones might be one explanation underlying why in vivo fibrils show different morphology and protease stability compared to in vitro fibrils.[52]

Methods

Construct and Recombinant Protein Preparation

The recombinant protein NT*MaSp-Bri2 BRICHOS was expressed in SHuffle T7 E. coli cells, purified by a Ni-NTA column, separated by a Superdex200 column (Cytiva), and cleaved by thrombin, and eventually the tag-free Bri2 BRICHOS monomers were isolated by a Superdex75 column (Cytiva), as described in previous study.[21] The 42 amino acid residues (1–42) of Aβ were fused to the NT*FlSp tag and expressed in BL21(DE3) E. coli.[31] In brief, the NT*FlSp-Aβ42wt was purified with a Ni-NTA column with following the protocol, as described previously[31] but without using denaturant (i.e., urea) to avoid potential urea-induced modification. The fusion NT*FlSp-Aβ42wt proteins were cleaved by NT*FlSp-Tev and lyophilized. The lyophilized powder was solubilized in 20 mM Tris pH 8.0 with 7 M guanidium chloride, and the Aβ42wt monomers were isolated by a Superdex30 26/600 (Cytiva) in 20 mM NaPi pH 8.0 with 0.2 mM EDTA and aliquoted in low-binding Eppendorf tubes (Axygene). The Aβ42wt concentration was calculated through an extinction coefficient of 1424 M–1 cm–1 for (A280–A300). For generating arctic mutant (E22G) of Aβ42, the primers 5′-ctggtgttcttcgctggagacgtgggttctaac-3′ and 5′-gttagaacccacgtctccagcgaagaacaccag-3′ were synthesized. With the NT*FlSp-Aβ42wt plasmid as the polymerase chain reaction (PCR) template, NT*FlSp-Aβ42arc was obtained with the QuikChange II XL site-directed mutagenesis kit (Agilent, US). The preparation of Aβ42arc monomers was performed with following the same protocol as described above, but the final Aβ42arc monomers were refined with an analytical superdex30 10/300 column (Cytiva). Regarding the Tev construct, gene coding for Tev proteinase was cloned into the modified pET vector with NT*FlSp solubility tag, encoding the fusion protein NT*FlSp-Tev. NT*FlSp-Tev plasmid was transformed into BL21(DE3) E. coli competent cells, which were cultured at 37 °C in LB medium with 70 μg/mL kanamycin until an OD600nm ∼ 0.8. The temperature was turned down to 20 °C, and 0.5 mM (final concentration) isopropyl β-d-1-thiogalactopyranoside was added for overnight expression. The cells were collected by 7000 g centrifugation at 22 °C for 20 min and resuspended in 50 mM NaPi pH 8.0 with 200 mM NaCl and 10% glycerol. After 5 min on ice sonication (65% power, 2 s on, 2 s off), the cell lysate was centrifuged for 30 min at 4 °C with a speed of 24 000 g, and NT*FlSp-Tev present in the supernatant was purified with a Ni-NTA column. The final target proteins were eluted by 50 mM NaPi pH 8.0 containing 200 mM NaCl, 10% glycerol, and 250 mM imidazole and immediately buffer-exchanged to 25 mM NaPi pH 7.5 with 100 mM NaCl and 10% glycerol with a HiPrep 26/10 desalting column (Cytiva). The cleavage efficiency was evaluated by cleaving NT*FlSp-Aβ42wt at a ratio of 1:100 (proteinase/substrate, w/w) at 4 °C via analyzing band intensities at different time points on SDS-PAGE. For all the constructs above, the final DNA sequences were confirmed by sequencing (GATC Bioteq, Germany).

ThT Assay

For monitoring amyloid fibril formation and the kinetics, 20 μL of solution (20 mM NaPi pH 8.0 with 0.2 mM EDTA) containing monomeric Aβ42wt (1.0, 1.3, 1.6, 2.0, 3.0, 4.0, 5.0, 7.0, and 9.0 μM) and Aβ42arc (1.0, 1.3, 1.5, 2.0, 2.5, 3.0, 3.5, and 4.0 μM) at different concentrations in the presence of 10 μM ThT were added to each well of half-area 384-well black polystyrene microplates with clear bottom and nonbinding surface (Corning Glass 3766, USA) and incubated at 37 °C under quiescent conditions. The ThT fluorescence was continuously recorded using a 440 nm excitation filter and a 480 nm emission filter (FLUOStar Galaxy from BMG Labtech, Germany). For investigating the inhibition effects of rh Bri2 BRICHOS monomers on Aβ42arc fibril formation, 20 μL of solution (20 mM NaPi pH 8.0 with 0.2 mM EDTA) containing Aβ42arc monomers, 10 μM ThT, and different concentrations of rh Bri2 BRICHOS monomers at molar ratios 0, 10, 50, 70, and 100% relative to the Aβ42arc monomer concentration were added to each well of half-area 384-well black polystyrene microplates with clear bottom and nonbinding surface (Corning Glass 3766, USA) and incubated under quiescent conditions at 37 °C. The fluorescence was recorded as described above. To prepare fibrils for EM observation of both Aβ42wt and Aβ42arc fibrils, 20 μL of solution (20 mM NaPi pH 8.0 with 0.2 mM EDTA) containing 3.0 μM Aβ42wt or 3.0 μM Aβ42arc monomers with and without 100% BRICHOS was added to each well (four replicates) of half-area 384-well black polystyrene microplates with clear bottom and nonbinding surface (Corning Glass 3766, USA) and incubated at 37 °C under quiescent conditions overnight, among them one well for each was added with 10 μM ThT to monitor the aggregation. Furthermore, 100% (molar ratio) of rh Bri2 RRICHOS monomers were added to each well after the formation of fibrils and incubated again at 37 °C under quiescent conditions overnight. For investigating Aβ42 fibrillization kinetics with seeds, 20 μL of solution containing 10 μM ThT, 3 μM Aβ42 monomer, different concentrations of monomeric rh Bri2 BRICHOS, and 0.6 μM seeds (calculated from the concentration of initial Aβ42 monomers) were added in cold room to each well of half-area 96-well plates and incubated at 37 °C under quiescent conditions. The fluorescence measurement settings were carried out as described above. Linear fits were applied to the concave aggregation traces (the first 24 min) to determine the initial slopes. For all the experiments, aggregation traces were normalized and averaged using four replicates.

Electron Microscopy Sample Preparation and Imaging

For immunogold staining of Aβ42 fibrils, the final incubation products (3.0 μM Aβ42arc) with BRICHOS added initially and after fibril preformed, respectively, were applied to form var-coated nickel grids and incubated for 2 min. Excess solution was removed with the filter paper (Whatman, grade 1). Blocking was performed by incubating the grids for 30 min in 1% BSA in TBS (Tris-buffered saline), followed by 3 × 10 min TBS washing. The grids were then incubated with primary antibody (goat anti-Bri2 BRICHOS antibody, 1:200 dilution) in cold room overnight, followed again by 3 × 10 min TBS washing. The grids were incubated with 10 nm gold particle-coupled secondary antibody (anti-goat IgG, 1:40 dilution, BBI Solutions, UK, EM.RAG10) at room temperature for 2 h and then washed with 1× TBS for 5 × 10 min. For staining, 2.5% uranyl acetate (2 μL) was added to each grid (for 20 s), and excess solution was carefully removed. The grids were air-dried and analyzed by TEM (Jeol JEM2100F at 200 kV). For imaging fibrils of Aβ42wt and Aβ42arc co-incubated with and without rh Bri2 BRICHOS monomers or with added BRICHOS to the preformed fibrils, the final incubation products were applied to carbon-coated copper grids (400 mesh, Analytical Standards) and incubated for 2 min. Excess solution was removed by blotting with the filter paper (Whatman, grade 1), and the grids were washed with two drops of Milli-Q water. For staining, 7 μL of 2% uranyl acetate was added to each grid for 45 s before final blotting and air-drying. The grids were analyzed by TEM (Jeol JEM2100F at 200 kV). All measurements were performed using ImageJ 1.53k. The single-like fibers with no visible twists or bundle structures were classified as S, whereas the multiple fibrils were classified as M. The measurements of twist body and twist body (crossover point) of Aβ42wt fibrils included 39 and 40 points, respectively. For Aβ42arc fibrils, 223 and 65 measurement points, respectively, were selected randomly for the diameter measurements. For Aβ42arc and rh Bri2 BRICHOS co-incubated fibrils, the diameter measurements of single-like and multiple fibrils were performed on 77 and 76 measurements, respectively.

Kinetic Analysis

For extracting the aggregation half-time τ1/2 and the maximal growth rate rmax, the aggregation traces of Aβ42wt and Aβ42arc with and without rh Bri2 BRICHOS monomers were fitted to a sigmoidal equationwhere A is the amplitude and F0 is the base value.[21,23] For global fit analysis, the aggregation trace of the total fibril mass concentration, M(t), is described by an integrated rate law, as described by Cohen et al.[41,53]where k, k+, and k2 are the microscopic rate constants for primary nucleation, elongation, and secondary nucleation, respectively, and n and n2 are the reaction orders of primary and secondary nucleation, respectively. The aggregations trace of Aβ42wt and Aβ42arc with and without rh Bri2 BRICHOS monomers were globally fitted using IgorPro and the AmyloFit 2.0 platform[34] (https://amylofit.com/amylofitmain/fitter/) with models for secondary nucleation dominated (unseeded) and multistep secondary nucleation dominated (unseeded) according to the γ values and previous reports,[7] respectively, where the k+k and k+k2 were constrained globally or free for aggregation traces with BRICHOS. The parameters n and n2 both were set to 2. The nucleation unit generation was calculated by integrating the nucleation rate r(t) over the reaction,[41] where r(t) = km(t) + k2M(t)m(t).

Statistical Analysis

All the statistically analyses were performed in Prism 9. Student’s t test (unpaired) was used for statistical analysis of two groups of data. The multiple groups were statistically compared with the ordinary one-way analysis of variance following by multiple comparisons with Tukey correction. Significance levels are *p < 0.05; **p < 0.01; ***p < 0.001; and ****p < 0.0001.

Data Availability

All data and materials related to this paper are available upon request.