The Alzheimer’s Disease “non-amyloidogenic” p3 peptide revisited: a case for Amyloid-α

Abstract

Amyloid-β (Aβ) is an intrinsically disordered peptide thought to play an important role in Alzheimer’s Disease. It has been the target of most AD therapeutic efforts, which have repeatedly failed in clinical trials. A more predominant peptidic fragment, formed through alternative processing of the Amyloid Precursor Protein, is the p3 peptide. p3 has received little attention, which is possibly due to the prevailing view in the AD field that it is “non-amyloidogenic”. By probing the self-assembly of this peptide, we found that p3 aggregates to form oligomers and fibrils, and when compared with Aβ, displays enhanced aggregation rates. Our findings highlight the solubilizing effect of the N-terminus of Aβ, and the favorable formation of structures formed through C-terminal hydrophobic peptide interfaces. Our findings suggest a reevaluation of the current therapeutic approaches targeting only the β-secretase pathway of AD, given that the α- secretase pathway is also amyloidogenic.

Alzheimer’s Disease (AD) is a debilitating neurological disease characterized by amyloid plaques and neurofibrillary tau tangles. It currently affects almost 50 million people worldwide.¹ Although amyloid plaques may contain up to 900 unique proteins, with approximately 200 found consistently,^2,3 the target of most therapeutic efforts to date has been the Amyloid-β peptide (Aβ).^4,5 Aβ has been the subject of thousands of publications annually since 2000, a figure that is exponentially rising (Fig. S1). Yet, in 2020, all Phase III drugs designed to reduce Aβ production or aggregation have failed,⁶ and we are far from an effective treatment for AD.^5–9 Thus, it is critical to investigate the contribution of other plaque-associated peptides to amyloid deposition in the brain.

In light of recent findings that fragments of Aβ exhibit enhanced aggregation propensity,^10,11 we sought to understand the predominant cleavage product of the Amyloid Precursor Protein (APP),¹² cleaved by α- rather than β- secretase. This peptide, termed p3, has received remarkably little attention and it is largely regarded as benign. However, its biophysical and biological properties remain unclear and inconsistent, as discussed in our recent review¹³ and summarized in Table S1. Some have described p3 as neuroprotective,¹⁴ while others have demonstrated that p3 exhibits significant cytotoxicity.^15,16 Whereas p3 is often referred to as soluble and “non-amyloidogenic”,^17–22 several studies found that p3 formed “amorphous aggregates” and “lattice-like” networks,^19,23 and, possibly, amyloid fibrils.^24,25 One study, conducted by Vandersteen et. al.²⁵ revealed what were described as fibrillar fragments, shorter and dissimilar to those formed by Aβ. Despite the aforementioned studies revealing fibrillar-like morphologies for p3, potentially indicative of amyloidogenicity, the field still recognizes p3 production as an alternative neuroprotective, “non-amyloidogenic” pathway of APP.^17,19–22 To deconvolute these conflicting claims, we sought out to investigate the aggregation propensity of p3. We hypothesized, based on the hydrophobicity and large regions of predicted amyloidogenicity,¹³ that p3 could aggregate to form oligomers and fibrils.

To probe our hypothesis, that despite lacking the first 16 amino acids, p3 shares amyloidogenic properties with Aβ, p3 and Aβ (Fig. 1A) were synthesized and purified using our published protocols²⁶ (Fig. S2–S3). To evaluate fibrillogenicity, p3 was incubated under fibril-forming conditions and compared alongside Aβ. TEM analysis of p3 fibrils (Fig. 1B; S9A–B) revealed long, twisted fibrillar structures that closely resemble the fibrils formed by Aβ (Fig 1C). For both p3 and Aβ, the resultant fibril morphologies bear semblance to published quiescently-formed Aβ fibrils.²⁷ Thus, the absence of the hydrophilic, disordered N-terminus did not preclude fibril formation in p3, contrary to the previous claims that p3 is “non-amyloidogenic”.^18,19 In addition, p3 fibrils exhibited characteristic green birefringence typical for cross-β-sheet amyloid structures,^28,29 under polarized light upon incubation with Congo Red (CR) (Fig. 1D), again indistinguishable from Aβ (Fig. 1E). The presence of β-sheet-rich structures formed by p3 revealed by the CR assay are in agreement with the circular dichroism results obtained by Ali et. al.³⁰

Figure 1. Fibrillogenicity of p3 compared with Aβ.

A) Fibril structure³⁵ and sequence of Aβ(1–40), with p3(17–40) indicated in purple. Highlighted regions (17–22 and 30–40) indicate the amyloidogenic regions of Aβ, calculated by AGGRESCAN.¹³ TEM images of fibrillar B) p3 and C) Aβ. Polarized light microscopy images of CR stained D) p3 and E) Aβ fibrils. F) Dot blots stained for OC binding for Aβ and p3. All peptide samples were prepared at 20 μM and incubated for 7 days at 37 °C in PBS (pH 7.4).

Our results agree with the findings of Sawaya et. al. that large hydrophobic residue patches frequently associate to form “steric zippers”.³¹ A hydrophobic steric zipper is plausible in the case of p3 given that the two amyloidogenic, hydrophobic patches of Aβ (LVFFAE and AIIGLMVGGVV)¹³ are also found in p3 (Fig. 1A). To further investigate if fibrillar Aβ and p3 have conformational similarities, the OC antibody, which recognizes conformation-specific epitopes of amyloid fibrils,³² was employed (Fig. 1F). The dot blots reveal OC binding for both p3 and Aβ across all timepoints, confirming the presence of conformational similarities in the fibril structures for both peptides, again highlighting the importance of the hydrophobic C-terminal regions for fibril formation.³¹ Together, these findings of shared fibril morphology with Aβ, CR birefringence, and OC-binding, establish that p3 is amyloidogenic as defined by the Nomenclature Committee of the International Society of Amyloidosis.^33,34

We next studied fibril formation kinetics of p3 via the Thioflavin T (ThT) and TAMRA-quenching assays.³⁶ The ThT assay (Fig. 2A) revealed that the fibrilization of p3 was significantly more rapid than for Aβ. The p3 kinetic profile was absent of a characteristic sigmoidal growth profile beginning with a “lag phase”, that is typically seen for Aβ. This indicates that the nucleation phase of the fibril formation mechanism was expedited for p3. Moreover, a ThT-monitored seeding assay revealed that Aβ fibrilization can be enhanced by seeding of pre-formed p3 fibrils (Fig. S10).

Figure 2. Kinetics of Fibril Formation.

A) ThT (20μM) monitored aggregation kinetics of Aβ and p3. B) TAMRA quenching assay. Peptides were prepared at 20 μM and incubated at 37 °C in PBS (pH 7.4) with continuous shaking. Each data point is an average of four replicates with error bars representing standard deviation.

To validate the kinetic trends seen in the ThT assays, TAMRA dye was conjugated to the N-termini of p3 and Aβ (Fig. S6–S7), and the fluorescence signal suppression, indicative of aggregation,³⁶ was monitored. As shown in Fig. 2B, TAMRA-p3 fluorescence decayed more rapidly than TAMRA-Aβ, further revealing that p3 forms fibrils more rapidly than Aβ. The ion-rich 1–16 region of Aβ therefore appears to provide a solubilizing function for Aβ, which attenuates amyloid fibril formation. This may also explain why p3 is a major component of preamyloid plaques found in the brains of those with Down Syndrome.¹⁹ Fibril formation of TAMRA-p3 was confirmed with TEM (Fig. S8–S9). Together, these results indicate that p3 may play an important role in amyloid deposition in the brains of those with AD.

Evidence is growing that the toxicity of Aβ is not from the accumulation of insoluble fibrils, but rather from the soluble oligomeric intermediates.³⁷ Oligomers have been shown to disrupt neuronal function,³⁸ long-term potentiation,³⁹ neuronal microRNA expression,⁴⁰ and memory.^41,42 Oligomers vary in size and shape and are difficult to characterize given their transient nature. Following the trapping protocol by Ahmed et. al.,⁴³ we trapped and imaged p3 oligomers, which are indistinguishable from those shown in TEM images of Aβ published by Ahmed. et. al.⁴³ The TEM images displayed in Fig. 3A–B (Fig. S11) reveal round particles ranging from 10–85 nm in diameter for both p3 and Aβ. No elongated, fibrillar species were identified in the oligomeric TEM images, as reflected in the histograms shown in Fig. 3C. The average particle diameter for p3 was determined to be 26.6 ± 11.2 which is within error of the value calculated for Aβ, 23.4 ± 11.5 nm (Fig. 3C). The histograms shown in Fig. 3C reveal very similar particle size distributions for both peptides. However, the histogram for Aβ exhibits some tailing, representative of larger structures (60–100 nm) not prevalently seen for p3. The formation of large non-spherical, elongated structures seen for Aβ but not p3, may be due to micelle formation, promoted by the amiphipathic, surfactant-like build of Aβ,⁴⁴ a property not shared with p3 (Fig. 1A). To the best of our knowledge, these are the first TEM images of oligomers formed by p3, refuting the previous claim that p3 does not form oligomers.⁴⁵ Our findings that p3 forms oligomers similar to those formed by Aβ agree with the MD simulations conducted by Miller et. al. indicating that p3 oligomers can form both parallel and antiparallel β-sheets.⁴⁶ Moreover, since several studies have demonstrated cellular toxicity for both monomeric and fibrillar p3,^15,16,23,47 we probed the toxicity of oligomeric p3 (Fig. S12). We found that at 50 μM, p3 reduced SH-SY5Y cell viability to approximately 80%, in comparison to 50% for Aβ-treated cells (Fig. S12). We attribute the higher viability of p3-treated cells to the rapidity of p3 fibrilization, as shown in Fig. 2.

Figure 3. Oligomer Characterization.

TEM images of 20 μM A) Aβ and B) p3 incubated at 4°C in PBS (pH 7.4) for 6h. C) Histograms displaying size distribution of oligomers formed by Aβ and p3 calculated from A-B and Fig. S11.

To further characterize early stages of self-assembly of p3, photo-induced crosslinking⁴⁸ was utilized to provide a snapshot of the short-lived, metastable oligomers. Two modified variants of p3 were synthesized given that tyrosine is required for crosslinking⁴⁹ (Fig. 4A): p3_F19Y, p3_F20Y. The SDS-PAGE gel shown in Fig. 4B reveals monomeric, dimeric, trimeric and tetrameric oligomers resulting from covalent cross-linking for both p3_F19Y and p3_F20Y. Aβ also formed similar assembly sizes in addition to faint signals corresponding to higher order structures (20–30 kDa), which may be related to the presence of larger oligomers formed by Aβ but not p3, as seen in TEM (Fig. 3; S11). The presence of low-N oligomers identified with TEM and SDS-PAGE demonstrate that p3 can form transient intermediates broadly similar in size and shape as Aβ. These findings are important in the context of AD because cellular toxicity of Aβ oligomers has been attributed to an excess of hydrophobic surface exposure of oligomers in cell membranes.⁵⁰ Given that p3 is almost entirely hydrophobic (Fig. 1A), p3 oligomers could potentially play an important role in amyloid-related neurotoxicity.

Figure 4. Oligomer Characterization.

A) Sequences of p3 peptides with Phe → Tyr substitutions. B) SDS-PAGE gel of crosslinked samples of Aβ, p3, p3_F19Y, p3_F20Y (20 μM). Control, non-crosslinked samples are shown in Fig. S13. Crystal structures displaying possible trimeric (C⁵¹,D⁵²) and tetrameric (E)⁵³ assemblies of p3 peptidic fragments, with representative centroid-to-centroid distances between phenylalanines from neighboring units indicated.

ThT kinetic curves for both p3_F19Y and p3_F20Y are shown in Fig. S14, indicating that tyrosine substitution did not prevent fibril formation. Interestingly, the kinetic profile for p3_F20Y closely matched that of p3, while p3_F19Y exhibited a slightly attenuated onset of aggregation. This may indicate a favorable interaction between F19 and another residue, reinforced by the discovery that a F19P mutation abolishes Aβ aggregation.⁵⁴ Moreover, our findings are supported by Bitan et. al. showing that truncations of residues 1–9 of Aβ do not inhibit oligomer formation.⁴⁹ Aβ also formed similar assembly sizes in addition to faint signals corresponding to high-N structures (20–30 kDa).

Furthermore, the enhanced staining for trimers and tetramers for p3 shown in Fig. 4B may reveal initial steps in the mechanism of p3 aggregation. Crystal structures of similar species are shown in Fig. 4C–E,^51–53 indicating possible structures of the prominent species shown in Fig. 4B.

In summary, we have demonstrated that the “non-amyloidogenic” p3 peptide, in fact does exhibit substantial amyloidogenic properties. Given this, and its similarity to Aβ, we propose to rename p3 to Amyloid-α (Aα). Our results revealed OC-positive, CR birefringent, p3 fibrils that formed more rapidly than Aβ. We also found that, contrary to previous literature,⁴⁵ p3 aggregated to form intermediate oligomers, that share a similar size distribution with Aβ. Overall, this work suggests that p3 may not be as innocuous as previously suggested, and further analysis is needed to understand the role of p3 in Alzheimer’s Disease.

Methods

Additional experimental methods can be found in the Supporting Information.

Peptide Synthesis

All peptides were synthesized by SPPS using Fmoc chemistry, following our previously published protocols⁵⁵ on Tentagel® S PHB resin (Rapp Polymere GmbH, cat no. RA1327).

Fibril growth

Lyophilized peptide (Aβ40 or p3) was dissolved in 20 mM NaOH and sonicated for 30 s, then diluted to 20 μM in PBS. The samples were incubated either (1) at 37 °C for 24 hours with mild agitation, or (2) at 37 °C quiescently for 7 days. For TEM imaging, 3 μL of sample aliquots were spotted onto freshly glow-discharged carbon-coated electron microscopy grids (Ted Pella, Catalog No. 01701-F). The grids were rinsed with milliQ water after 1 min incubation, followed by staining with 30 μL 1% uranyl acetate.

Congo Red

Since CR crystals can exhibit false-positive birefringence, a protocol from Nakka et. al.⁵⁶ was adapted. Either Aβ40 or p3 fibrils were grown quiescently as described above at 20 μM. Fibrils were centrifuged at 14000 g for 15 min and washed with MilliQ H₂O x2. 5 μL of the CR stock solution (7 mg of CR in 1 mL MilliQ H₂O) was added to 1 mL of MilliQ H₂O and then added to the pelleted peptides. The samples were incubated for 1 h with mild agitation and centrifuged at 14000 g for 15 min and washed with MilliQ H₂O x2. 30 μL were added to glass slides and air-dried overnight. Images were collected on a Leica Epifluorescence widefield microscope, with a polarizer.

Dot blot assay

Aβ40 or p3 fibrils (20 μM) were prepared as described above and kept at 37 °C. At each timepoint over the course of 7 days, 2 μL of sample was spotted on nitrocellulose. After the blots dried, they were blocked with 5% non-fat milk in TBST for 1 h at 25 °C. The samples were then washed with TBST for 5 min x3, and then incubated with OC antibody (1:1000 in 5% non-fat milk in TBST) overnight at 4 °C, followed by washing with TBST for 5 min x3. The membrane was then incubated with the secondary antibody (1:10,000 in 5% non-fat dry milk in TBST) for 1 h at 25 °C, and washed with TBST for 5 min x3 and then developed with the Opti-4CN Substrate kit (Bio-Rad, cat no. 1708235).

Oligomer Growth and Imaging

To trap the intermediate oligomers, lyophilized Aβ40 or p3 was dissolved in cold 20mM NaOH and sonicated in an ice bath for 30 s. Samples were diluted to either 20 μM in PBS and incubated at 4 °C for 6 h without agitation. TEM samples were prepared and imaged as described above.

ThT Assay and TAMRA Quenching Assays

ThT assay and TAMRA quenching were conducted as described previously.^57,58

Photochemically induced crosslinking of peptides

4 μL of 1 mM [Ru(bipy)₃]²⁺ and 4 μL of 20 mM ammonium persulfate were added to 32 μL aliquots of 20 μM Aβ40, p3, p3_F19Y and p3_F20Y in PBS. The samples were irradiated for 1.2 s with white light using our previously described setup.⁵⁵ Following irradiation, the samples were immediately quenched with 40 μL of loading buffer containing 5% 2-mercaptoethanol, and separated by SDS-PAGE gel electrophoresis (12% tris-tricine polyacrylamide) at 100 V for 2 h. The gels were developed by silver staining.