PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Arch Biochem Biophys. Author manuscript; available in PMC 2017 May 1.
Published in final edited form as:
PMCID: PMC4841699
NIHMSID: NIHMS772745

Aβ40 has a subtle effect on Aβ42 protofibril formation, but to a lesser degree than Aβ42 concentration, in Aβ42/Aβ40 mixtures

Abstract

Recent findings suggest that the senile plaques in Alzheimer’s disease may contain soluble amyloid-β peptide (Aβ) fibril precursors along with insoluble fibrils.. These soluble Aβ species, including oligomers and protofibrils, have been well-studied in vitro and are formed via non-covalent self-assembly of Aβ monomers. While both 40- and 42-residue forms of Aβ are observed in the human body, the majority of the Aβ aggregation work has been conducted on Aβ42 or Aβ40 separately, with relatively few investigations of mixtures. In order to study the effect of different combinations of Aβ40 and Aβ42 on protofibril formation, mixtures of either dry solid peptide, or purified Aβ40 and Aβ42 monomer solutions were mixed together and protofibril/monomer distributions were quantified. Increases in the Aβ42/Aβ40 ratio increased protofibril formation but the presence of Aβ40 in the mixed Aβ solutions had a significant negative impact on protofibril formation compared to equivalent solutions of pure Aβ42. Protofibril size was less affected, but β-sheet structure increased with protofibrils formed from higher Aβ42/Aβ40 ratio solutions. Direct measurement of Aβ42/Aβ40 ratios by C-terminal-selective ELISA found very little Aβ40 incorporated into protofibrils. The cumulative data emphasizes the critical importance of Aβ42, yet establishes Aβ40 as a regulator of Aβ42 aggregation.

Keywords: Amyloid-beta protein, aggregation, protofibrils, light scattering

1. Introduction

Alzheimer’s disease (AD) is the most prevalent type of dementia in older adults and is caused by progressive neurodegeneration in the hippocampus and the amygdala in affected patients [1]. As of 2013, AD was the sixth leading cause of death in the United States. It has been estimated that 5.3 million Americans are currently living with AD and in the coming years the number will increase significantly due to the aging baby boom generation [2, 3]. Furthermore, it is estimated that the number of people currently living with dementia worldwide (46.8 million) is expected to double in 20 years [4],

Two pathological hallmarks, neurofibrillary tangles (NFTs) and senile (amyloid) plaques, still dominate the biochemistry of AD [5]. NFTs are intraneuronal deposits of insoluble filamentous aggregates of tau, a microtubule-associated protein. Senile plaques are extracellular deposits in the brain parenchyma of insoluble amyloid-β protein (Aβ) fibrils. Clinicopathologic studies have shown that the clinical progression of AD correlates closely with the density and distribution of abnormally-folded tau NFTs in the brain; while the initial cause of the disease appears to be linked to Aβ aggregation and accumulation [6].

Aβ is a collection of secreted peptide fragments produced by β- and γ-secretase-catalyzed proteolytic cleavage of the amyloid precursor protein (APP). The cleavage yields Aβ fragments 37–43 amino acids in length (reviewed in [710]). The two most commonly-studied fragments are Aβ40 and Aβ42. Both peptides are found circulating normally in the blood and cerebrospinal fluid with Aβ40 at much higher levels than Aβ42 [11, 12]. Even though the two isoforms differ by only two amino acid residues, their biophysical properties are very different. It is widely recognized that Aβ42 is more fibrillogenic than Aβ40 [13] and this property has many ramifications on the disease process. In early-onset familial AD, mutations in the APP gene surrounding the γ-secretase cleavage site alter Aβ metabolism by increasing the total Aβ amount and the Aβ42/Aβ40 ratio [1416]. While the circulating levels of Aβ are dominated by Aβ40, senile plaques consist overwhelmingly of Aβ42 [17]. Aβ deposits in the leptomeningeal vessels also have a greater percentage of Aβ42 [18]. Furthermore, studies in mouse models, human presenilin-expressing mouse fibroblast lines, and human plasma indicate that the total Aβ concentration may not be as important as the Aβ42/Aβ40 ratio for plaque formation [14, 1921]. These observations, along with other findings, support the critical, and detrimental, role of Aβ42 in AD pathogenesis [22]. In fact, the Aβ42 level in cerebrospinal fluid (CSF) is one component of an important biomarker test for AD patients that also includes CSF tau and amyloid imaging [23]. The reduction in circulating plasma and CSF Aβ42 is believed to be due to aggregation and deposition of Aβ42 in the brain prior to onset of AD symptoms [24, 25].

Much of the detailed knowledge about Aβ aggregation kinetics and mechanisms has been acquired through in vitro studies using synthetic peptides. These studies have been invaluable for understanding the process by which Aβ undergoes non-covalent self-assembly in solution and identifying the conditions that modulate this process. There is general agreement that unstructured monomeric Aβ42 and Aβ40, at sufficient concentration, will nucleate and begin to form soluble oligomers [26, 27] and protofibrils with significant β-sheet content [2830]. The initial nucleation step, conformational pathway, and subsequent oligomerization can be influenced by pH, ionic strength, temperature, and agitation [28, 3133]. Many of the soluble Aβ species ultimately progress to insoluble fibrils, which are indistinguishable from those isolated from brain tissue [34]. Although fibril-bearing senile plaques are the most visible Aβ pathological feature, the failure of insoluble Aβ load to correlate strongly with memory loss [35] significantly increased the interest in soluble Aβ aggregates for their role in neurodegeneration [36, 37].

As discussed above, numerous in vivo investigations have established that the Aβ42/Aβ40 ratio is an important factor in AD pathology. Despite this information, most studies utilize only one form of Aβ. Although fewer in number, there have been several reports on Aβ aggregation kinetics and mechanisms in Aβ42/Aβ40 mixtures with most suggesting that Aβ40 inhibits Aβ42 aggregation. For example, the inclusion of monomeric Aβ40 suppressed both Aβ42 monomer and protofibril conversion to fibrils and failed to promote Aβ42 fibril elongation [38]. Other reports have demonstrated shorter lag times and/or faster aggregation rates as the Aβ42/Aβ40 solution ratio increased [39, 40]. It has been postulated that Aβ40 inhibits Aβ42 aggregation by preferentially binding to Aβ42 aggregates thereby preventing further polymerization [39]. A contrasting theory maintains that Aβ40 potentially delays Aβ42 aggregation through “non-productive” interactions [41]. The findings that a minor increase in the solution Aβ42/Aβ40 ratio increases neurotoxicity [40, 42] emphasizes the importance of understanding the aggregation dynamics of Aβ42/Aβ40 mixtures and the impact on Aβ structure and biological activity.

The current study examined the effect of varying Aβ42/Aβ40 ratios on protofibril formation. Protofibrils are a well-characterized member of an ever-broadening class of soluble Aβ aggregates. Their formation in vivo may play a role in AD pathogenesis making them an important potential therapeutic target [43]. The investigation revealed marked changes in both the extent of protofibril formation and the character of the protofibrils as the Aβ42/Aβ40 ratio changed. The findings demonstrated profound effects by Aβ42 on the initial nucleation phase, which is closely coupled to rapid protofibril formation.

2. Materials and methods

2.1. Aβ preparation

Aβ42 and Aβ40 peptides were obtained from W.M. Keck Biotechnology Resource Laboratory (Yale School of Medicine, New Haven, CT) in lyophilized form, individually dissolved in 100% hexafluoroisopropanol (HFIP) (Sigma-Aldrich, St. Louis) at 1 mM, separated into aliquots in sterile microcentrifuge tubes, and evaporated uncovered at room temperature overnight in a fume hood. The following day, the aliquots were vacuum-centrifuged to remove any residual HFIP and stored in dessicant at −20°C. Preparation of A β42/Aβ40 dry peptide mixtures was accomplished by resuspending Aβ42 and Aβ40 aliquots separately in 100% HFIP and mixing the two solutions in the desired amounts to form either pure Aβ42 or Aβ40 or molar Aβ42/Aβ40 ratios of 4:1, 1:1, and 1:4. The volumes used to prepare the planned dry peptide ratios were calculated to yield a total Aβ concentration of 200 μM upon reconstitution. Briefly, Aβ42 and Aβ40 (1 mg) were resuspended in 100% HFIP to a concentration of 1 mM. Ratios of 4:1, 1:1, and 1:4 Aβ42/Aβ40 were prepared by mixing selected volumes of each Aβ/HFIP into three separate tubes. The volumes used were 160 μL Aβ42 (0.72 mg) and 40 μL Aβ40 (0.17 mg) for 4:1; 100 μL Aβ42 (0.45 mg) and 100 μL Aβ40 (0.43 mg) for 1:1; and 40 μL of Aβ42 (0.18 mg) and 160 μL Aβ40 (0.69 mg) for 1:4. The Aβ42/Aβ40 HFIP mixtures were evaporated overnight in a fume hood uncovered. Vacuum centrifugation was used to remove any remaining HFIP and samples were stored in a desiccant at −20°C.

2.2. Size exclusion chromatography

Dry Aβ42/Aβ40 peptide mixtures were dissolved in 50 mM NaOH to yield a 2 mM solution followed by dilution to 200 μM Aβ in prefiltered artificial cerebrospinal fluid (aCSF, 15 mM NaHCO3, 1 mM Na2HPO4, 130 mM NaCl, 3 mM KCl, pH 7.8) and incubation for 30 min at 25 °C to allow protofibril formation. Protofibrils and monomers were separated as previously described [44]. The Aβ solutions were centrifuged at 18,000g for 10 min and the supernatant was fractionated using size exclusion chromatography (SEC) on a Tricorn Superdex 75 10/300GL column (fractionation range 3–70 kD, GE Healthcare) attached to an AKTA FPLC system (GE Healthcare). Bovine serum albumin (Sigma) was routinely run each day to prevent non-specific binding of Aβ to the column matrix. Aβ was eluted at 0.5 mL min−1 in aCSF and 0.5 mL fractions were collected and immediately placed on ice. Aβ concentrations were determined in-line by UV absorbance using an extinction coefficient for Aβ of 1450 cm−1 M−1 at 280 nm. For aggregation of Aβ42/Aβ40 monomer mixtures directly isolated by SEC, solutions were diluted with aCSF to 40 μM total Aβ, supplemented with 0.05% sodium azide (NaN3) to prevent contamination, and quiescently incubated at 25 °C for 25 days with num erous measurements during the incubation. For aggregation of Aβ42/Aβ40 monomer mixtures prepared from separately SEC-purified Aβ42 and Aβ40 monomers, solutions were mixed together at different ratios to a total Aβ concentration of 40 μM in aCSF buffer and incubated at 37 °C without disturbance for 24 h. Following an 18,000g spin, supernatants were further separated on SEC in-line with light scattering.

2.3. Dynamic light scattering

Hydrodynamic radius (RH) measurements were made as previously described [44]. Measurements were taken at room temperature with a DynaPro Titan instrument (Wyatt Technology, Santa Barbara, CA). Samples (30 μl) were placed into a quartz cuvette and light scattering intensity was collected at a 90º angle using a 5-second acquisition time. Particle diffusion coefficients were calculated from auto-correlated light intensity data and converted to RH with the Stokes-Einstein equation using Dynamics software (version 6.12.0.3). Histograms of percent intensity vs. RH were generated by data regularization and intensity-weighted mean RH values were derived from the regularized histograms.

2.4. Thioflavin T fluorescence assay

Aβ samples were assessed by ThT fluorescence as previously described [45]. Briefly Aβ samples were diluted to 5 μM (unless otherwise noted) in aCSF pH 7.8 containing 5 μM ThT. Fluorescence emission scans (460–520 nm) were acquired on a Cary Eclipse fluorescence spectrophotometer using an excitation wavelength of 450 nm and integrated from 470–500 nm to obtain ThT relative fluorescence values. Excitation and emission slit widths and power settings were the same for all fluorescence measurements. Buffer controls did not show any significant ThT fluorescence in the absence of Aβ. All ThT fluorescence numbers are reported in relative fluorescence units (RFU). Extended aggregation assays of SEC-purified Aβ42/Aβ40 monomer mixtures were monitored by ThT fluorescence similarly.

2.5. Multi-angle light scattering (MALS)

MALS data was obtained in-line with SEC using a DAWN DSP instrument (Wyatt Technology, Santa Barbara, CA). Samples were analyzed under continuous flow through a K12 quartz cell and were illuminated with a helium-neon laser. Light scattering data were obtained every 0.167 mL at fixed-angle detectors 4–16 and plotted using the Debye formalism of the Rayleigh-Debye-Gans approximation [46, 47] for large anisotropic particles (eq 1).

Rθ/Kc=MwP(θ)
(eq 1)

This equation relates the Rayleigh ratio (scattering intensity) to molecular weight (Mw) multiplied by a form factor, P(θ), where c is solute concentration (g/mL), θ is the scattering angle, K is an optical constant equal to 4 π2 (dn/dc)2 no2 λo−4 NA−1, n is the refractive index of the solution, no is the refractive index of the solvent, λo is the swavelength of incident light in a vacuum (632.8 nm), and NA is Avagadro’s number. P(θ), which takes into account the destructive interference of scattering by large particles, is the scattering intensity for a large particle divided by the scattering intensity without interference and can be expanded in a power series. The limiting form of the form factor equation as θ approaches zero is shown in eq 2 where Rg2 is the mean square radius, and λ is λo / no.

P(θ)=1-(16π2/3λ2)Rg2sin2(θ/2)
(eq 2)

Substitution of eq 2 into eq 1 and plotting of Rθ / Kc versus sin2(θ/2) yields Mw as the y-intercept and a limiting slope equal to Mw (16 π2 / 3λ2) Rg2. Debye plots were fit with a linear regression using dn/dc values for Aβ and BSA of 0.186 mL/g and 0.190 mL/g respectively.

2.6. Circular dichroism

Samples (minimum 0.2 ml) were placed into a Jasco J0556 quartz rectangular quartz cuvette with a 0.1 cm pathlength, or smaller volumes (15 μL) were measured via quartz-windowed microsampling disc (Jasco MSD-462) with 0.1 cm pathlength. Spectra were obtained with a Jasco J-1500 circular dichroism spectrometer by wavelength scan from 260 to 190 nm using scan speeds, data integration times, and response factors as provided in the Figure legend. Five successive wavelength scans were averaged for each Aβ sample. Buffer control spectra were averaged and subtracted from Aβ sample spectra, and each resulting point ([θ]obs, deg) was converted to mean residue ellipticity ([θ], deg cm2 dmol−1) with the equation [θ] = [θ]obs × (MRW/10lc), where MRW is the mean residue molecular weight of Aβ(1–42) (4514.1 g/mol divided by 42 residues), l is the optical path length (cm), and c is the concentration (g/cm3).

2.7. Aβ C-terminal-selective ELISA

Aβ antibodies were obtained as a generous gift from Mayo Clinic Jacksonville Department of Neuroscience. C-terminal selective antibodies Ab2.1.3 (Aβ42 specific) and Ab13.1.1 (Aβ40 specific) were employed as capture antibodies. Antibody Ab5, which recognizes both Aβ peptides, was used as the detection antibody. Ab5 was conjugated to HRP using a previously described protocol [48]. Briefly, 2 mg HRP (Sigma) was dissolved in 400 μL water and incubated at room temperature for 20 min with 100 μL of 0.1 M sodium periodate on an orbital shaker protected from light. The enzyme was then dialyzed at 4°C against 1 mM sodium acetate buffer (pH 4.4) overnight. The dialyzed HRP solution (190 μL) was supplemented with 10 μL of 200 mM sodium carbonate buffer (pH 9.5) and mixed with 200 μL of Ab5 (1.77 mg/mL) and 300 μL of 10 mM sodium carbonate buffer (pH 9.5). The mixture was incubated for 2 h at room temperature on an orbital shaker followed by the addition of 50 μL of freshly prepared sodium borohydride solution (4 mg/mL) and further mixing at 4°C for 2 h. The Ab5-HRP conjugat e was purified by SEC as described above. The absorbance of each fraction was assessed at 280 nm (protein) and 403 nm (HRP heme groups) for determination of Ab5 concentration.

Measurement of Aβ42 and Aβ40 in SEC-isolated Aβ42/Aβ40 mixtures by ELISA with all steps carried out at room temperature. Briefly, a 96-well plate was coated overnight with 100 μL of 5 μg/mL Ab2.1.3 or Ab13.1.1 capture antibodies. The wells were washed with PBS containing 0.05% Tween 20, blocked for 1 hr with PBS containing 1.0% BSA, 5.0% sucrose and 0.05% NaN3, and washed again. Successive treatments with washing in between were done with samples or standards (2 h), Ab5-HRP detection antibody (2 h), and equal volumes of HRP substrates TMB and hydrogen peroxide (0.5 h). The reaction was stopped by the addition of 1% H2SO4 solution and the optical density for each sample was determined at 450 nm with a reference reading at 630 nm. SEC-purified Aβ42 and Aβ40 monomers were used to construct separate standard curves from 1,000–10,000 pM and were done for each ELISA. Standards and samples were diluted to 1 μM in 7.2 M GuHCl for 30 min and then further diluted to fall within the standard curve. The denaturant step enhanced antibody recognition for both monomeric and oligomeric Aβ. Additional ELISA development included testing of Aβ40 standard curves with Ab2.1.3 (Aβ42 specific) and Aβ42 with Ab13.1.1 (Aβ40 specific). No cross-reactivity was observed.

3. Results

3.1. Protofibril formation increases as the Aβ42/Aβ40 ratio increases in mixed solutions

We have previously detailed the preparation and characterization of the biophysical and proinflammatory properties of protofibrils rapidly formed by Aβ42 and to a slower extent by Aβ40 [44, 45, 49]. A variety of solution and microscopic techniques indicated that Aβ42 protofibrils were short, curvilinear species less than 100 nm in length, rich in β-sheet structure, yet relatively small and quite soluble (hydrodynamic radius, RH 20–25 nm and molecular weight, Mw, 200–2600 kDa). Those findings are consistent with previous reports [29, 30, 50] and expand the current analysis of these soluble precursors to fibrils. Aβ protofibrils may be an important therapeutic target in AD therefore a better understanding of their properties will enhance this effort. Furthermore, the observation that circulating Aβ in plasma and CSF contains both Aβ42 and Aβ40 underscores the importance of understanding the effect of varying ratios of the two isoforms on protofibril formation. Since protofibril formation occurs rapidly upon reconstitution, different mixtures of dry Aβ42 and Aβ40 peptide were prepared to ensure that the correct mixture would be present upon reconstitution. This procedure is described in the Methods and was done by mixing 100% HFIP solutions of Aβ42 and Aβ40 together at predetermined volumes to obtain the correct molar ratios and then removing the HFIP under vacuum centrifugation. Upon reconstitution in the appropriate volume of modified artificial cerebrospinal fluid (aCSF) [45], the desired final individual solution concentrations of Aβ42 and Aβ40 were achieved while maintaining the total Aβ concentration at 200 μM (depicted in Figure 1A). The extent of rapid protofibril formation in the Aβ42/Aβ40 mixtures after a 30 min incubation at 25 °C was assessed by SEC-separation and spectroscopic quantitation. A solution of Aβ42 alone yielded a typical 2-peak elution profile on Superdex 75 (fractionation range 3–70 kD) with roughly equivalent amounts of protofibrils and monomers (Figure 1A). Our previous work showed that protofibrils were larger than 70 kD [49], so protofibrils eluted in the void volume. The calculated distribution was 56% protofibrils and 44% monomers (Figure 1B). This finding is consistent with our previous reports on Aβ42 protofibril formation [44, 45]. However, a decrease in the Aβ42 concentration and an increase in the Aβ40 concentration resulted in a decrease in protofibril formation. Accordingly, the decrease in protofibril formation corresponded with an increase in residual monomers (Figure1). The quantitation of this effect is depicted in Figure 1B. These findings demonstrate the effect that the Aβ42/Aβ40 ratio and Aβ42 concentration have on Aβ nucleation and protofibril formation.

Figure 1
A change in the Aβ42/Aβ40 ratio directly impacts the extent of protofibril formation. Panel A. Lyophylized peptide mixtures containing varying ratios of Aβ42 and Aβ40 were solubilized with NaOH followed by dilution into ...

3.2. Protofibril size range slightly increases as the Aβ42/Aβ40 ratio increases in mixed solutions

Dynamic light scattering (DLS) analysis was conducted on the protofibril pools in Figure 1A obtained following SEC-separation of the selected Aβ42/Aβ40 mixtures. Diffusion coefficients were measured and converted to hydrodynamic radius (RH) in Dynamics software 6.12.0.3 using the Stokes-Einstein equation. Regularization of the light scattering data separated the light scattering of the protofibrils from the buffer contribution and any trace presence of large particles with RH values greater than 1000 nm. Regularized intensity-weighted histograms revealed a subtle change in protofibril size for the different Aβ42/Aβ40 ratios (Figure 2). Determination of the mean RH from the histograms of each sample yielded values of 21 nm for 200/0, 16 nm for 160/40, 16 nm for 100/100, 14 nm for 40/160, and 13 nm for 0/200 (data not shown). The decrease in the Aβ42/Aβ40 peptide ratio appeared to alter protofibril amount more dramatically than protofibril size.

Figure 2
RH histograms of protofibrils formed from varying Aβ42/Aβ40 mixtures. The protofibril pools shown in Figure 1A were analyzed immediately after SEC-isolation by DLS and regularized histograms of percent intensity versus RH are presented ...

3.3. Protofibril thioflavin T fluorescence increases as the Aβ42/Aβ40 ratio increases in mixed solutions

Thioflavin-T (ThT) is a benzothiazole dye, which shows increased fluorescence when bound to amyloid fibrils [51]. The β-sheet architecture of aggregated Aβ in particular creates a ThT binding site [52]. ThT binding and fluorescence has previously been used to characterize soluble Aβ protofibril species [30, 44]. Here we used ThT to analyze the SEC-isolated protofibril pools obtained from each of the different Aβ42/Aβ40 mixtures in Figure 1A. ThT fluorescence measurements were conducted at a final Aβ concentration of 5 μM for each of the protofibril samples to ensure that a change in fluorescence was not due to lesser or greater numbers of protofibrils in the sample. ThT analysis revealed that the extent of ThT binding and fluorescence decreased as the Aβ42/Aβ40 ratio decreased (Figure 3). The ThT data suggested that protofibrils formed from different Aβ42/Aβ40 mixtures differed in their conformation or the ThT binding and/or ThT molecular rotation was altered in protofibrils formed from solutions containing a lower Aβ42/Aβ40 ratio. The small peak of material eluting in the void volume (protofibril peak) for the 40 μM Aβ42/160 μM Aβ40 and the 200 μM Aβ40 solutions exhibited minimal β-sheet character based on ThT fluorescence levels very close to the background levels.

Figure 3
ThT fluorescence for protofibrils formed from Aβ42/Aβ40 mixtures decreases as the Aβ42/Aβ40 ratio decreases. The protofibril pools shown in Figure 1A were analyzed immediately after SEC-isolation by ThT binding and fluorescence ...

3.4. Mixed monomer fractions after SEC-isolation display same Aβ42/Aβ40 ratio-dependence on nucleation

A second set of dry Aβ42/Aβ40 peptide mixtures were prepared and separated by SEC as in Figure 1. The same trend of protofibril and residual monomer elution was observed for the same Aβ42/Aβ40 mixtures. The aggregation kinetics of the SEC-eluted monomers (containing mixtures of Aβ42 and Aβ40) was conducted at a lower concentration than the protofibril formation described earlier at 200 μM Aβ. In contrast to rapid protofibril formation, this process permitted observation of aggregation lag times. The SEC-purified Aβ42/Aβ40 monomer mixtures were each brought to 40 μM total Aβ in a solution of aCSF containing 0.05% NaN3 and incubated at room temperature. ThT fluorescence measurements were taken at numerous time points to assess the aggregation of each Aβ42/Aβ40 monomer mixture over time. The SEC-purified all-Aβ42 (40 μM) and the 4:1 Aβ42/Aβ40 (≈32 μM/8 μM) monomer exhibited a very rapid initial aggregation with significant ThT fluorescence by 21 hrs (Figure 4). The 1:1 Aβ42/Aβ40 monomer solution (≈20 μM/20 μM) showed an extended aggregation lag phase with little ThT fluorescence at 21 hrs, but significant levels at 47 hrs (Figure 4). All five solutions, once nucleated, reached roughly the same ThT fluorescence levels. The 1:4 Aβ42/Aβ40 (≈8 μM/32 μM) and all-Aβ40 (40 μM) monomer solutions aggregated much slower with lag times of 3 days, decreased polymerization rates, and lower overall maximum ThT fluorescence levels (Figure 4). Similar to rapid protofibril formation, this data shows that as Aβ42 increases and Aβ40 decreases within a monomeric Aβ solution, lag times decrease and the overall ThT binding/fluorescence increases.

Figure 4
The Aβ42/Aβ40 ratio modulates the nucleation phase prior to aggregation. SEC-purified monomer pools from a second series of Aβ42/Aβ40 reconstituted dry peptide mixtures were prepared as described in the Methods and Figure ...

3.5. Aβ40 has an impact on Aβ42 aggregate yields

The findings heretofore indicate that at high Aβ concentrations, particularly Aβ42, nucleation is rapid and protofibrils are formed but with a significant dependence on the ratio between the two peptides. However, working with stored dry Aβ42, and to a lesser extent Aβ40, peptide samples without further purification increases the potential for pre-existing structures that may influence the ensuing aggregation process. In order to preclude this possibility, studies were done in solution mixtures of Aβ42 and Aβ40 using separately SEC-purified monomers of each. For these studies, both Aβ42 and Aβ40 were reconstituted in 6 M GuHCl containing 10 mM NH4OH to suppress rapid protofibril formation and increase monomer yields. Purified monomer fractions were obtained from the included volume of a Superdex 75 column, immediately placed on ice, and promptly used for aggregation studies. A total Aβ (Aβ42 + Aβ40) concentration of 40 μM was selected so that 0.5 mL (90 μg Aβ) could be loaded for in-line SEC-MALS analysis and post-elution evaluation of fractions. This amount was the realistic minimum for sample detection and analysis [49]. Aβ42 and Aβ40 monomers were isolated in aCSF pH 7.8 containing 30 mM NaCl concentration, rather than 130 mM to encourage soluble aggregate formation and discourage insoluble fibril formation [49]. Using this paradigm previously with Aβ42, we have observed the formation of both protofibrils and protofilaments [49]. Furthermore, in order to better elucidate whether Aβ40 directly impacts Aβ42 aggregation, or is merely a bystander during Aβ42 nucleation, additional and separate control experiments were conducted with Aβ42 alone.

Aβ42 and Aβ40 mixtures prepared from SEC-purified monomers were incubated for 24 h at 37 °C and then monitored for ThT fluore scence. Increased Aβ42/Aβ40 ratio correlated with higher ThT fluorescence after 24 h (Figure 5A). A solution of purely Aβ40 did not display any aggregation at all while purely Aβ42 displayed the greatest aggregation. The presence of Aβ40 at a small ratio (4:1 Aβ42:Aβ40) slowed the overall aggregation based on ThT fluorescence (Figure 5A). A separate experiment was conducted using identical conditions and the same four Aβ42 concentrations (40, 32, 20, and 8 μM) but without Aβ40 in the solutions. The decrease in ThT fluorescence as Aβ42 concentration decreased (Figure 5B, 40 μM to 32 μM) was not nearly as dramatic as with the inclusion of Aβ40 (e.g. 32 μM Aβ42/8 μM Aβ40). The aforementioned Aβ42/Aβ40 solution mixtures were centrifuged at 18,000g for 10 min and the supernatants were re-purified on SEC. The pure Aβ40 solution did not yield any soluble protofibrils in the Superdex 75 void volume (Figure 6A), but increasing Aβ42/Aβ40 ratio significantly increased soluble protofibril levels (Figure 6A). A plot of the protofibril/monomer distribution (Figure 6B) highlighted the significant impact of Aβ42 on protofibril formation and the difference in aggregation propensity between Aβ42 and Aβ40. The ThT fluorescence in Figure 5A and elution chromatograms in Figure 6A indicated that more protofibrils were being formed as the Aβ42/Aβ40 ratio increased. Further inspection of the data also indicates that the presence of Aβ40 significantly decreased protofibril formation by Aβ42 (Figure 6A). This observation is more clearly evident when comparing the protofibril/monomer chromatogram distributions in Aβ42/Aβ40 solution mixtures (Figure 6A, B) with the control solutions containing Aβ42 alone (Figure 6C, D). The effect of Aβ42 concentration on protofibril formation was not as dramatic as with the inclusion of Aβ40. Certainly, the increased concentration of Aβ40 monomers, which are slower to aggregate, in the aggregation solutions will skew the protofibril/monomer distribution. However, careful inspection of the protofibril peaks formed by Aβ42 alone and those including Aβ40 demonstrates an apparent subtle inhibitory effect of Aβ40 on Aβ42 protofibril formation (Figure 6A and 6C).

Figure 5
Increasing Aβ42/Aβ40 ratio and Aβ42 concentration accelerates monomer aggregation. Solution mixtures of SEC-purified Aβ42 and Aβ40 monomers (total Aβ 40 μM) in aCSF buffer were prepared and incubated ...
Figure 6
Increasing Aβ42/Aβ40 ratio increases the yields of soluble protofibrils but Aβ40 retards the process. Solution mixtures of SEC-purified Aβ42 and Aβ40 monomers (total Aβ 40 μM) were prepared and incubated ...

3.6. The Aβ40/Aβ42 ratio has a negligible effect on protofibril molecular weight

In order to establish if the Aβ42/Aβ40 ratio affected the protofibril size, we took advantage of multi-angle light scattering in-line with SEC (SEC-MALS) to obtain molecular weight (Mw) values for each solution. The three solutions that yielded sufficient aggregates for evaluation were 40 μM Aβ42/0 μM Aβ40 (40/0), 32 μM Aβ42/8 μM Aβ40 (32/8), and 20 μM Aβ42/20 μM Aβ40 (20/20) (Figure 7A). Mw values were obtained every 0.167 mL and remained fairly constant across the Superdex 75 void peaks for all three solutions (Figure 7A), with little difference between varying Aβ ratios. Weight-averaging of the “slice” data through the protofibril concentration peaks (full width half max) for each solution gave Mw values of 2.95 × 106 Da, 3.13 × 106 Da, and 2.92 × 106 Da for 40/0, 32/8, and 20/20 solutions respectively. The Mw determinations were representative of approximately 650–700 monomer units within the Aβ aggregates, which is on the higher end of protofibril Mw and nearing protofilaments. Another size parameter, root mean square radius (Rg), was also obtained from the MALS data and confirmed there was little change in size between aggregates formed from the three solutions. Z-average Rg values were of 57 nm, 57 nm, and 58 nm for 40/0, 32/8, and 20/20 solutions respectively (data not shown). The control solutions containing Aβ42 alone at different concentrations were also evaluated with SEC-MALS. For this series, Mw values were found to be larger than those obtained for the series of Aβ42/Aβ40 mixtures, but did not vary much between the different Aβ42 concentrations (40 μM, 4.21 × 106 Da; 32 μM, 4.64 × 106 Da; and 20 μM, 5.25 × 106 Da) (data not shown). SEC- MALS-determined Rg values for the pure Aβ42 solutions were similar to the initial set of Aβ42/Aβ40 mixture with values of 55 nm, 57 nm, and 56 nm for 40, 32, and 20 μM respectively (data not shown). Overall, the findings revealed that the presence of Aβ40 in Aβ42/Aβ40 mixtures lowered the number of soluble protofibrils formed. However, the protofibrils that were formed had similar Mw and size to solutions of Aβ42 alone. This suppressive effect was greater than simply lowering the Aβ42 concentration in solutions of Aβ42 alone.

Figure 7
Increasing Aβ42/Aβ40 ratio increases protofibril yields but does not significantly protofibril molecular weight. Panel A. Molecular weights of the measureable protofibril peaks were obtained using SEC-MALS during the elution of the Aβ42/Aβ40 ...

ThT fluorescence in the SEC-eluted protofibril pools was similar between Aβ42/Aβ40 solution mixtures after normalization for concentration (data not shown). The comparable fluorescence per micromolar Aβ for the protofibrils was consistent with the similarity in Mw values between protofibrils formed by Aβ42 and Aβ42/Aβ40 mixtures. Despite these similarities, circular dichroism (CD) measurements uncovered subtle structural differences in protofibrils formed by Aβ42 and mixtures of Aβ42 and Aβ40 (Figure 8A). Protofibrils formed from a pure Aβ42 solution (40/0) contained significant β-sheet structure, while protofibrils isolated from the 32/8 and 20/20 Aβ mixtures exhibited less β-structure and some elements of α-helical structure. Interestingly, this same lessening of β-structure was observed in pure Aβ42 solutions when the concentration was lowered. Again, protofibrils formed in the 40 μM Aβ42 solution were mostly β-sheet, while the lower two concentrations (32 μM and 20 μM Aβ42) had elements of α-helix (Figure 8B). Cumulatively, a decrease in solution Aβ42 concentration effected aggregation rate, and the yield and secondary structure of protofibrils. However, the presence of Aβ40 enhances these effects.

Figure 8
Increasing Aβ42/Aβ40 ratio enhances protofibril β-structure. Peak protofibril fractions from the SEC-purified Aβ42 and Aβ40 solution mixtures in Figure 6A or different Aβ42 concentrations solutions in Figure ...

3.7. There is little to no incorporation of Aβ40 into protofibrils formed from Aβ40/Aβ42 solution mixtures

The extent of protofibril formation was governed by Aβ42/Aβ40 ratio in solution mixtures. However, it was not known what the relative amounts of each peptide that were incorporated into SEC-isolated protofibril pools. A C-terminal-selective Aβ ELISA was developed in order to distinguish and quantitate differences in Aβ42 and Aβ40 concentrations. The ELISA was utilized to determine the Aβ42/Aβ40 ratio in the solution mixtures in Figure 5A and the peak protofibril fractions from the SEC elutions in Figure 6A. The ELISA-determined ratios are presented in Table 1 and underscored the difference between the initially prepared Aβ42/Aβ40 aggregation solution mixtures and which peptide ends up in the protofibril fraction. The initial aggregation solutions and protofibrils that contained only Aβ42 (40/0) were only detected by the Aβ42 capture antibody (Ab 2.1.3) and the samples that contained only Aβ40 (0/40) were only detected by the Aβ40 capture antibody (Ab 13.1.1). These findings confirmed the C-terminal selectivity of the ELISA (Table 1). Ratios of >10 or <0.01 were presented in Table 1 for clarity but the footnote indicates that only one peptide was detected in each pure solution.

Table 1
Measured Aβ42/Aβ40 ratios in mixed monomer solutions and in SEC-isolated protofibrils. Initial Aβ42/Aβ40 solution mixtures and SEC-eluted protofibril peak fractions were collected and analyzed by C-terminal-selective Aβ ...

The ELISA data indicated that the initial aggregation solution mixtures had Aβ42/Aβ40 ratios close to the expected values to the original dry peptide mixtures (Table 1). However, the predominant Aβ peptide found in the protofibrils formed from Aβ42/Aβ40 solution mixtures was Aβ42. Particularly emphasizing this point was the 20 μM Aβ42/20 μM Aβ40 (20/20) mixture, wherein SEC-isolated protofibrils contained virtually all Aβ42 (Table 1). C-terminal-selective ELISA analysis of the initial 20/20 mixture was close to an even mix between the two peptides. Our findings indicate that that much of the protofibril formation in Aβ42/Aβ40 mixtures occurred by Aβ42 with very little Aβ40 incorporation. While Aβ40 may impact Aβ42 aggregation and possibly alter the aggregate structure, it does not extensively incorporate within the oligomeric assembly.

3.8. Aβ40 has a lesser effect on protofibril formation when Aβ42 concentration remains fixed

In order to directly assess the impact of Aβ40 on Aβ42 protofibril formation, solution mixtures of newly SEC-purified Aβ42 and Aβ40 monomers were prepared. Aβ42 concentration was maintained at 20 μM, while Aβ40 concentration varied. In this paradigm, Aβ40 had only a small effect on Aβ42 protofibril formation. ThT fluorescence was moderately reduced in the solution mixtures containing Aβ40 after a 24 h incubation (data not shown), but only small changes were observed in the protofibril peak after SEC separation (Figure 9A). Accordingly, monomer levels increased in the Superdex 75 chromatograms as the Aβ40 concentration increased in the solution aggregation mixtures (Figure 9B). While the protofibril / monomer ratio decreased as the Aβ40 concentration increased (Figure 9C), a closer look at the SEC protofibril peak integration (nmol) indicated that the effect of Aβ40 on the amounts of Aβ42 protofibrils was slight (Figure 9D). However, there was an observable decrease in the ThT fluorescence in the collected SEC peak protofibril fractions from aggregation solutions that contained Aβ42 plus Aβ40 (Figure 9E). Further analysis with CD of the same peak SEC protofibril fractions suggested that the decrease in ThT binding/fluorescence of Aβ42 protofibrils formed in the presence of Aβ40 may reflect a change in the β-sheet character of the Aβ42 protofibrils (Figure 9F). Additional elements of secondary structure beyond β-sheet were noted in the spectra when the Aβ40 concentration increased. This was particularly evident after SEC-isolation and analysis of the 20 μM Aβ42 / 56 μM Aβ40 aggregation solution (Figure 9F). The overall findings demonstrate that decreased Aβ42 concentration and the presence of Aβ40 may work together to prevent protofibril formation, but Aβ42 concentration has a much more significant impact at the ratios examined in this study.

Figure 9
Protofibril formation at fixed Aβ42 concentration is not dramatically effected by inclusion of increasing Aβ40 concentration in the mixture. Solution mixtures containing SEC-purified Aβ42 monomers (20 μM) and increasing ...

4. Discussion

Aβ assembly occurs via a nucleation-dependent polymerization process whereby the rate-limiting nucleation step is preceded by a lag phase and followed by rapid polymerization and ultimately fibril formation [53]. The length of the lag is inversely correlated with concentration and it can be abolished by the introduction of a pre-formed fibril seeds. The lag phase duration is highly variable even in homogeneous aqueous solutions and is sensitive to many conditions [28]. Protofibril formation is closely tied to the nucleation step in the Aβ aggregation pathway. Our investigation of Aβ42/Aβ40 mixtures allowed us to follow those early assembly steps and the effect of each peptide on the process.

Measurements of human plasma Aβ levels have yielded concentration ranges of 30–230 pM for Aβ40 and 6–31 pM for Aβ42 [11, 24, 54, 55]. The reported Aβ42/Aβ40 ratio from these studies ranges from 0.06–0.38 for control patients with the most common ratio centering about 0.15. While most techniques do not have the appropriate sensitivity at these low pM concentrations, the current study investigated Aβ42/Aβ40 ratios that might recapitulate those found in normal or AD physiology. Comparisons were then made to solutions composed of purely Aβ42 or Aβ40.

As the Aβ42/Aβ40 ratio was modulated in the prepared mixtures, we observed several phenomena. One, the extent of protofibril formation increased at higher Aβ42/Aβ40 ratios based on SEC elution profiles. Two, while the size (RH) and Mw of the SEC-isolated protofibrils from the mixtures increased only modestly or not at all in solutions with different ratios, the samples formed from higher Aβ42/Aβ40 solutions exhibited greater ThT fluorescence per μM Aβ suggesting that they had a more developed or defined β-sheet structure. Three, CD analysis confirmed a more developed β-sheet structure at higher Aβ42 concentration in the absence of Aβ40. Four, control experiments with pure Aβ42 solutions suggested that while Aβ42 concentration was a factor in the mixtures, Aβ40 appears to act as a modulator of Aβ42 aggregation. These findings highlighted the important role of Aβ42 concentration in both triggering protofibril formation and defining the structure and properties of the protofibrils.

The most straightforward explanation for Aβ40 retarding Aβ42 aggregation is a direct association with Aβ42 monomers. However, direct measurement of the protofibril fractions by C-terminal-selective ELISA indicated that the protofibrils contained predominately Aβ42 with negligible Aβ40 incorporation. This argues against an association-inhibition mechanism for Aβ40, at least at the nucleation stage. However, it is possible that trace amounts of Aβ40 monomers may have the capacity to physically interact with Aβ42 and interfere with aggregation at substoichiometric ratios. The introduction of one monomeric “cap” to a growing Aβ42 oligomeric chain could retard the process. There is previous evidence that once Aβ42 protofibrils are formed, they may be able to undergo additional growth via addition of Aβ40 monomers leading to a mixed Aβ42/Aβ40 protofibril [38]. A similar type of cross-deposition of Aβ42 and Aβ40 onto sonicated fibril seeds prepared from either peptide has also been demonstrated [41]. While further studies will be necessary to confirm that a single protofibril can be composed of both Aβ42 and Aβ40, the earlier published findings indicate that interactions between the two peptides may be more productive once a conformational template is present. Overall, the results demonstrate that Aβ42 self-assembly dominates the process at the nucleation stage, but Aβ40 may impede early aggregate (protofibril) formation. This report provides evidence of important roles for both Aβ42 and Aβ40 in nascent soluble Aβ aggregate formation and reinforces the critical consequence of increasing Aβ42/Aβ40 ratio.

Highlights

  • Protofibrils are rapidly formed from HFIP-treated, lyophilized dry Aβ peptide.
  • Aβ protofibril formation increases at higher Aβ42 / Aβ40 ratio.
  • Both Aβ42 concentration and the presence of Aβ40 modulate protofibril formation.
  • Protofibrils formed by mixtures of Aβ42 & Aβ40 are exclusively Aβ42.
  • Despite negligible incorporation, Aβ40 can alter Aβ42 protofibril structure.

Acknowledgments

This work was supported by Award Number R15AG033913 from the National Institute on Aging (MRN). We would also like to thank the National Science Foundation for the Major Research Instrumentation Award (#1337638) that funded acquisition of the Jasco 1500 Circular Dichroism Spectrometer for the University of Missouri-St. Louis.

Abbreviations used

AD
Alzheimer’s disease
amyloid-β protein
aCSF
artificial cerebrospinal fluid
DLS
dynamic light scattering
HFIP
hexafluoroisopropanol
RH
hydrodynamic radius
Rg
root mean square radius
MALS
mult-angle light scattering
SEC
size exclusion chromatography
ThT
thioflavin T

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

1. Reitz C, Mayeux R. Biochem Pharmacol. 2014;88:640–651. [PMC free article] [PubMed]
2. Alzheimer's Association. Alzheimers Dement. 2015;11:332–384. [PubMed]
3. Thies W, Bleiler L. Alzheimers Dement. 2013;9:208–245. [PubMed]
4. Prince M, Wimo A, Guerchet M, Ali GC, Wu YT, Prina M. Alzheimer's Disease International, World Alzheimer Report 2015, The global impact of dementia, an analysis of prevalence, incidence, cost and trends
5. Masters CL, Selkoe DJ. Cold Spring Harb Perspect Med. 2012;2:a006262. [PMC free article] [PubMed]
6. Dickson DW. J Clin Invest. 2004;114:23–27. [PMC free article] [PubMed]
7. De Strooper B, Vassar R, Golde T. Nat Rev Neurol. 2010;6:99–107. [PMC free article] [PubMed]
8. Selkoe DJ. Cold Spring Harb Perspect Biol. 2011;3:a004457. [PMC free article] [PubMed]
9. Selkoe DJ. Nat Cell Biol. 2004;6:1054–1061. [PubMed]
10. Nixon RA. J Cell Sci. 2007;120:4081–4091. [PubMed]
11. Hansson O, Zetterberg H, Vanmechelen E, Vanderstichele H, Andreasson U, Londos E, Wallin A, Minthon L, Blennow K. Neurobiol Aging. 2010;31:357–367. [PubMed]
12. Vigo-Pelfrey C, Lee D, Keim P, Lieberburg I, Schenk DB. J Neurochem. 1993;61:1965–1968. [PubMed]
13. Jarrett JT, Berger EP, Lansbury PT., Jr Biochemistry. 1993;32:4693–4697. [PubMed]
14. Scheuner D, Eckman C, Jensen M, Song X, Citron M, Suzuki N, Bird TD, Hardy J, Hutton M, Kukull W, Larson E, Levy-Lahad E, Viitanen M, Peskind E, Poorkaj P, Schellenberg G, Tanzi R, Wasco W, Lannfelt L, Selkoe D, Younkin S. Nature Med. 1996;2:864–870. [PubMed]
15. Hardy J. Trends Neurosci. 1997;20:154–159. [PubMed]
16. Tanzi RE. Cold Spring Harb Perspect Med. 2012;2:a006296. [PMC free article] [PubMed]
17. Gravina SA, Ho L, Eckman CB, Long KE, Otvos L, Jr, Younkin LH, Suzuki N, Younkin SG. J Biol Chem. 1995;270:7013–7016. [PubMed]
18. Roher AE, Lowenson JD, Clarke S, Wolkow C, Wang R, Cotter RJ, Reardon IM, Zurcher-Neely HA, Heinrikson RL, Ball MJ, Greenberg BD. J Biol Chem. 1993;268:3072–3083. [PubMed]
19. Bentahir M, Nyabi O, Verhamme J, Tolia A, Horre K, Wiltfang J, Esselmann H, De Strooper B. J Neurochem. 2006;96:732–742. [PubMed]
20. Deng Y, Tarassishin L, Kallhoff V, Peethumnongsin E, Wu L, Li YM, Zheng H. J Neurosci. 2006;26:3845–3854. [PubMed]
21. Kim J, Onstead L, Randle S, Price R, Smithson L, Zwizinski C, Dickson DW, Golde T, McGowan E. J Neurosci. 2007;27:627–633. [PubMed]
22. Findeis MA. Pharmacol Ther. 2007;116:266–286. [PubMed]
23. Tarawneh R, Holtzman DM. Neuropharmacology. 2010;59:310–322. [PMC free article] [PubMed]
24. Graff-Radford NR, Crook JE, Lucas J, Boeve BF, Knopman DS, Ivnik RJ, Smith GE, Younkin LH, Petersen RC, Younkin SG. Arch Neurol. 2007;64:354–362. [PubMed]
25. Fagan AM, Holtzman DM. Biomark Med. 2010;4:51–63. [PMC free article] [PubMed]
26. Bitan G, Kirkitadze MD, Lomakin A, Vollers SS, Benedek GB, Teplow DB. Proc Natl Acad Sci USA. 2003;100:330–335. [PubMed]
27. Kayed R, Head E, Thompson JL, McIntire TM, Milton SC, Cotman CW, Glabe CG. Science. 2003;300:486–489. [PubMed]
28. Harper JD, Wong SS, Lieber CM, Lansbury PT., Jr Biochemistry. 1999;38:8972–8980. [PubMed]
29. Walsh DM, Lomakin A, Benedek GB, Condron MM, Teplow DB. J Biol Chem. 1997;272:22364–22372. [PubMed]
30. Walsh DM, Hartley DM, Kusumoto Y, Fezoui Y, Condron MM, Lomakin A, Benedek GB, Selkoe DJ, Teplow DB. J Biol Chem. 1999;274:25945–25952. [PubMed]
31. Lambert MP, Barlow AK, Chromy BA, Edwards C, Freed R, Liosatos M, Morgan TE, Rozovsky I, Trommer B, Viola KL, Wals P, Zhang C, Finch CE, Drafft GA, Klein WL. Proc Natl Acad Sci USA. 1998;95:6448–6453. [PubMed]
32. Nichols MR, Moss MA, Reed DK, Lin WL, Mukhopadhyay R, Hoh JH, Rosenberry TL. Biochemistry. 2002;41:6115–6127. [PubMed]
33. Stine WBJ, Dahlgren KN, Krafft GA, LaDu MJ. J Biol Chem. 2003;278:11612–11622. [PubMed]
34. Kirschner DA, Inouye H, Duffy LK, Sinclair A, Lind M, Selkoe DJ. Proc Natl Acad Sci USA. 1987;84:6953–6957. [PubMed]
35. Westerman MA, Cooper-Blacketer D, Mariash A, Kotilinek L, Kawarabayashi T, Younkin LH, Carlson GA, Younkin SG, Ashe KH. J Neurosci. 2002;22:1858–1867. [PubMed]
36. Haass C, Selkoe DJ. Nat Rev Mol Cell Biol. 2007;8:101–112. [PubMed]
37. Walsh DM, Selkoe DJ. J Neurochem. 2007;101:1172–1184. [PubMed]
38. Jan A, Gokce O, Luthi-Carter R, Lashuel HA. J Biol Chem. 2008;283:28176–28189. [PMC free article] [PubMed]
39. Yan Y, Wang C. J Mol Biol. 2007;369:909–916. [PubMed]
40. Kuperstein I, Broersen K, Benilova I, Rozenski J, Jonckheere W, Debulpaep M, Vandersteen A, Segers-Nolten I, Van Der Werf K, Subramaniam V, Braeken D, Callewaert G, Bartic C, D'Hooge R, Martins IC, Rousseau F, Schymkowitz J, De Strooper B. Embo J. 2010;29:3408–3420. [PubMed]
41. Pauwels K, Williams TL, Morris KL, Jonckheere W, Vandersteen A, Kelly G, Schymkowitz J, Rousseau F, Pastore A, Serpell LC, Broersen K. J Biol Chem. 2012;287:5650–5660. [PMC free article] [PubMed]
42. Chang YJ, Chen YR. Febs J. 2014;281:2674–2687. [PubMed]
43. Lannfelt L, Moller C, Basun H, Osswald G, Sehlin D, Satlin A, Logovinsky V, Gellerfors P. Alzheimers Res Ther. 2014;6:16. [PMC free article] [PubMed]
44. Paranjape GS, Gouwens LK, Osborn DC, Nichols MR. ACS Chem Neurosci. 2012;3:302–311. [PMC free article] [PubMed]
45. Paranjape GS, Terrill SE, Gouwens LK, Ruck BM, Nichols MR. J Neuroimmune Pharmacol. 2013;8:312–322. [PMC free article] [PubMed]
46. Wyatt PJ. Anal Chim Acta. 1993;272:1–40.
47. Tanford C. Physical chemistry of macromolecules. John Wiley and Sons, Inc; New York: 1961.
48. Wisdom GB. Methods Mol Biol. 2005;295:127–130. [PubMed]
49. Nichols MR, Colvin BA, Hood EA, Paranjape GS, Osborn DC, Terrill-Usery SE. Biochemistry. 2015;54:2193–2204. [PubMed]
50. Lashuel HA, Grillo-Bosch D. Methods Mol Biol. 2005;299:19–33. [PubMed]
51. LeVine H. Protein Sci. 1993;2:404–410. [PubMed]
52. Krebs MR, Bromley EH, Donald AM. J Struct Biol. 2005;149:30–37. [PubMed]
53. Jarrett JT, Lansbury PT., Jr Cell. 1993;73:1055–1058. [PubMed]
54. Giedraitis V, Sundelof J, Irizarry MC, Garevik N, Hyman BT, Wahlund LO, Ingelsson M, Lannfelt L. Neurosci Lett. 2007;427:127–131. [PubMed]
55. Toledo JB, Shaw LM, Trojanowski JQ. Alzheimers Res Ther. 2013;5:8. [PMC free article] [PubMed]