- Top of page
- Materials and Methods
β-amyloid peptide (Aβ) is one of the main protein components of senile plaques associated with Alzheimer's disease (AD). Aβ readily aggregates to forms fibrils and other aggregated species that have been shown to be toxic in a number of studies. In particular, soluble oligomeric forms are closely related to neurotoxicity. However, the relationship between neurotoxicity and the size of Aβ aggregates or oligomers is still under investigation. In this article, we show that different Aβ incubation conditions in vitro can affect the rate of Aβ fibril formation, the conformation and stability of intermediates in the aggregation pathway, and toxicity of aggregated species formed. When gently agitated, Aβ aggregates faster than Aβ prepared under quiescent conditions, forming fibrils. The morphology of fibrils formed at the end of aggregation with or without agitation, as observed in electron micrographs, is somewhat different. Interestingly, intermediates or oligomers formed during Aβ aggregation differ greatly under agitated and quiescent conditions. Unfolding studies in guanidine hydrochloride indicate that fibrils formed under quiescent conditions are more stable to unfolding in detergent than aggregation associated oligomers or Aβ fibrils formed with agitation. In addition, Aβ fibrils formed under quiescent conditions were less toxic to differentiated SH-SY5Y cells than the Aβ aggregation associated oligomers or fibrils formed with agitation. These results highlight differences between Aβ aggregation intermediates formed under different conditions and provide insight into the structure and stability of toxic Aβ oligomers.
Alzheimer's disease (AD) is a progressive, neurodegenerative disease of the central nervous system and the leading cause of dementia in aging population. One of the histopathological hallmarks of AD is the formation of neuritic plaques, the major protein component of which is β-amyloid peptide (Aβ). Two variants of Aβ, Aβ1-40 and Aβ1-42, are the most abundant forms found in AD (Bayer et al. 2001; Turner et al. 2003).
Accumulating data show that while Aβ aggregates readily into amyloid fibrils, the Aβ species in soluble oligomeric forms rather than fibrils are closely associated with the neurotoxicity in vivo and in vitro. Toxicity has been attributed to Aβ-derived diffusible ligands (ADDLs) composed of Aβ1-42, with molecular weights between 17 and 42 kDa (Lambert et al. 1998; Klein et al. 2001) and heterogeneous globular species (also described as ADDLs) with hydrodynamic radii between 3 and 8 nm (Chromy et al. 2003). Spherical aggregates, not described as ADDLs, with radii over 15 nm have also been reported to be toxic (Hoshi et al. 2003). Toxic protofibrils of Aβ1-40 have been described with hydrodynamic radii ranging from 9 nm to over 300 nm (Walsh et al. 1999; Ward et al. 2000; Wang et al. 2002). When investigators have compared toxicity of fibril and oligomer Aβ species, they have found that the oligomeric species were more toxic (Dahlgren et al. 2002). While some investigators believe that there is a common structure associated with amyloid toxicity (Kayed et al. 2003), a careful characterization of such a common structure is still not available. Part of the challenge in elucidating the relationship between Aβ structure and toxicity is the difficulty in isolating pure samples of Aβ oligomers for characterization, although use of urea (Kim et al. 2004), low temperatures (Yong et al. 2002), or photocross-linkers have aided in their identification (Bitan and Teplow 2004). One approach has been to carefully characterize Aβ monomers and fibrils, and from those structures infer the nature of a toxic aggregation intermediate. However, it is becoming increasingly clear that even the molecular level structure of Aβ fibrils changes with aggregation conditions (Stine et al. 2003; Petkova et al. 2005, 2006; Shivaprasad and Wetzel 2006; among others).
In work presented here, we examined the differences in Aβ aggregation intermediates and final structures formed when only a simple modification in Aβ aggregation conditions was made, the presence or absence of agitation during aggregation. We show that while the final structures in the Aβ aggregation pathway are comparable by measures such as electron microscopy, the toxicity of fibrils formed are significantly different. In addition, intermediates in the aggregation pathway show significantly different structural rearrangements. When guanidine hydrochloride unfolding was used as a simple measure of stability of different aggregated species, the Aβ aggregation intermediates formed under quiescent conditions, the most toxic Aβ species we observed, were the structures that changed conformation at the lowest concentrations of guanidine hydrochloride. These results highlight the differences in Aβ aggregation mechanisms when aggregation occurs with agitation or under quiescent conditions. In addition, our results provide insight into the structure of the toxic Aβ oligomers formed during aggregation under quiescent conditions.
- Top of page
- Materials and Methods
In the effort to develop new methods of preventing toxicity associated with Aβ of Alzheimer's disease, many investigators have sought to elucidate both the mechanism of Aβ aggregation (Lomakin et al. 1997; Esler et al. 2000; Pallitto and Murphy 2001; Kim et al. 2004; O'Nuallain et al. 2005; Sabate and Estelrich 2005; Pellarin and Caflisch 2006) and the relationship between Aβ structure and toxicity (Lambert et al. 1998; Ward et al. 2000; Walsh et al. 2002; Klein et al. 2001; Chromy et al. 2003; Hoshi et al. 2003). In the work we present here, we show that when Aβ is aggregated under different conditions, different mechanisms of aggregation appear to occur, and the oligomers formed during aggregation have different toxicities.
As seen in Figure 1, kinetics of Aβ aggregation, regardless of the presence or absence of agitation, differ in scale, but appear similar in form. The characteristic lag in development of extended β-sheet structure to which Congo red binds is typical of a nucleation mechanism of aggregation that a number of investigators propose for Aβ fibril formation (Lomakin et al. 1997; Pallitto and Murphy 2001; Pellarin and Caflisch 2006). Increased rates of aggregation with agitation relative to Aβ aggregation under quiescent conditions has also been observed by others (Tycko 2006; Petkova et al. 2006). In this study, we provide agitation by rotation at 18 rpm on a laboratory rotator. Others have used gentle agitation and sonication to mix during Aβ aggregation and found that the higher energy mixing increased the rate of aggregation (Petkova et al. 2006). While Congo red binding suggests that aggregation with and without agitation may be similar, data obtained from size exclusion chromatography, CD spectroscopy, and electron microscopy (Figs. 2–4) suggest that there are significant differences in the mechanism of aggregation with and without agitation of Aβ.
SEC chromatograms indicate that, throughout aggregation under conditions with or without agitation, species that elute at volumes consistent with monomer and dimer are present in the same ratio of peak areas in all chromatograms. This indicates these two species were in equilibrium with each other under all these conditions. Others have made similar observations of Aβ aggregation (Pallitto and Murphy 2001). The amounts of monomer and dimer relative to other species decreased with time, corresponding to increases in fibril formation as indicated by Congo red binding (Table 1). Different investigators have suggested that monomer or dimer are the building blocks for fibril nucleation or growth (Garzon-Rodriguez et al. 1997; Tseng et al. 1999; Roher et al. 2000; Hwang et al. 2004).
CD data indicate that Aβ was largely β-sheet prior to aggregation and did not undergo significant secondary structural rearrangements when aggregated with agitation. When aggregated under quiescent conditions, however, a dramatic shift in structure was observed at intermediate times (∼4–8 h after aggregation is initiated), which corresponded to the presence of an intermediate-sized species in chromatograms (approximate molecular weight as estimated from elution volume on SEC of 53 kDa) and large globular species in electron micrographs. None of these intermediate species were observed when aggregation proceeded with agitation.
A number of investigators have observed Aβ aggregation intermediates, formed either at low temperature or via a particular preparation method, sometimes referred to as Aβ derived diffusible ligands (ADDLs) (Lambert et al. 1998; Klein et al. 2001), that have molecular weights within the range we observed via SEC. A number of reports indicate that these ADDLs are the (or one of the) toxic Aβ oligomers (Lambert et al. 1998). Others have described spherical species of various sizes that are Aβ aggregation intermediates with sizes ranging from 3 to 20 nm, and some up to 100 nm depending on starting peptide and preparation method (Harper et al. 1999; Westlind-Danielsson and Arnerup 2001; Hoshi et al. 2003). Toxicity of the spherical Aβ species has also been reported (Hoshi et al. 2003). The ∼53-kDa species observed via SEC may be distinct from the spherical species observed via electron microscopy; however, it is not possible to tell from our data. Others have observed the formation of Aβ micelles (Lomakin et al. 1996; Yong et al. 2002; Kim and Lee 2004; Kim et al. 2004; Sabate and Estelrich 2005) which might appear as a large globular species via EM, but could change apparent molecular size during chromatography. Alternately, it is probable that via SEC we would not be able to detect a globular protein aggregate of 20 nm diameter as it would be removed from our sample during our sample pretreatment, nor would we be able to detect 53-kDa proteins via EM as they would be below the resolution of the microscope used.
That we observe dramatic structural rearrangements with aggregation under quiescent conditions but did not observe the same structural rearrangements when aggregation occurred with agitation is not surprising. If Aβ aggregation is nucleation dependent as others propose (Lomakin et al. 1997; Ramírez-Alvarado et al. 2000; Wogulis et al. 2005), then with agitation, one would expect more rapid collisions of molecules and faster nucleation, both in solution and against the walls of the tube in which aggregation took place. The faster rate of nucleation would mean that Aβ would have less time to undergo any intramolecular structural rearrangements that might be energetically favorable and increase stability of the fibril formed. Without agitation, the rate of peptide collision with other peptides would be slower, allowing more time for intramolecular rearrangement of the peptide before combining with other peptides to form an aggregation intermediate. In short, agitation would favor intermolecular rearrangements to bury hydrophobic surfaces or form hydrogen bonds that would lead to aggregation, while quiescent conditions would favor intramolecular rearrangements to bury hydrophobic surfaces. These intramolecular rearrangements could lead to formation of aggregation intermediates not seen when intermolecular interactions predominate.
It has recently been suggested that amyloid fibrils form through a micelle intermediate when relatively unstable β-sheet-forming peptides form the building blocks of the amyloid, while no micelle intermediate would be seen if a stable β-sheet-forming peptide were used as fibril precursors (Pellarin and Caflisch 2006). It is possible that under quiescent conditions, Aβ forms a somewhat unstable monomer, but with agitation, a more stable β-sheet monomer or dimer is formed.
Work coming out of the laboratories of Tycko (Petkova et al. 2005, 2006) and Wetzel (Whittemore et al. 2005; Shivaprasad and Wetzel 2006; Williams et al. 2006) demonstrates that Aβ fibrils formed with agitation and those formed under quiescent conditions have different molecular level structures. The presence of intramolecular salt bridges, the number of layers of β-sheets that form a protofibril unit, and peptide residues that make up β-sheet segments may all be different in fibrils formed under the different conditions. Thus it is not surprising that our data suggest that different mechanisms of aggregation govern fibril formation with and without agitation.
The relative stability of species formed during aggregation in the presence and absence of agitation can be inferred from data collected on conformation change of Aβ in a denaturant (Fig. 5; Table 2). When Aβ was aggregated with agitation, fibrils underwent conformation change in denaturant more readily than monomer. The trend in change in protein conformation with denaturant concentration (or slope of changes in free energy with denaturant concentration “m”) is traditionally interpreted as a measure of buried surface (Myers et al. 1995; Pace et al. 1996), suggesting that Aβ fibrils formed during agitation have more buried surface than Aβ monomers. When Aβ was aggregated under quiescent conditions, aggregation intermediates underwent conformation change more easily than either fresh or fibril Aβ, suggesting the intermediate had the most buried surface. While this interpretation of our results is in disagreement with an earlier report (Kremer et al. 2000), it is consistent with data that comes from our laboratory of the fluorescence of the hydrophobic probe, 1-anilino-8-naphthalene sulfonic acid (ANS), indicating that the aggregation intermediate formed under quiescent conditions was less hydrophobic or bound less ANS than either monomer or fibril (unpubl. results).
We examined biological activity or toxicity of fibrils and aggregation intermediates formed with and without agitation (Fig. 6). Toxicity of fibrils was greater when formed with agitation, while toxicity of aggregation intermediates formed between 4 and 8 h after initiation of aggregation under quiescent conditions were more toxic than other species formed. Fibrils formed with agitation and aggregation intermediates formed under quiescent conditions changed structure more readily in denaturant than other species examined (Fig. 5). Thus, for samples and conditions considered here, the difference in apparent stability in denaturant correlated with toxicity. A number of investigators have suggested that Aβ membrane interactions via Aβ conformational changes are important in the mechanism of Aβ toxicity (McLaurin and Chakrabartty 1997; Terzi et al. 1997; Demuro et al. 2005). Behavior of Aβ in a membrane and in a denaturant might be analogous. Thus, one could speculate that toxicity of an Aβ species may be related to its ability to change structure within the cell membrane.
There are several alternative explanations of the data presented. Others have reported that Aβ oligomers with structures similar to those we describe forming during aggregation under quiescent conditions are more toxic than fibrils (Dahlgren et al. 2002; Hoshi et al. 2003). However, some investigators have observed that fibrils formed under quiescent conditions, then sonicated, are more toxic than fibrils formed with agitation (Petkova et al. 2005). When comparing toxicity of fibrils formed via different methods, it has been very difficult to characterize concentration or number of fibrils in a solution. The average length of fibrils has been shown to change with aggregation condition (Walsh et al. 1999; Ward et al. 2000). The number of fibrils might also change with aggregation conditions (or with sonication). With agitation, one would expect larger numbers of fibrils or fibril nuclei than would be formed without agitation. Larger number of fibril ends or fibril nuclei would be expected to be associated with greater toxicity (Wogulis et al. 2005). Consequently, differences in toxicity observed may simply be due to differences in oligomer or fibril concentrations in solution.
In summary, we report structure and toxicity of Aβ species formed during aggregation via different methods. We show that fibrils formed via different methods do not have the same toxicity nor the same apparent stability to denaturants, nor do they appear to be formed via the same mechanism or have the same intermediate aggregation species. Aggregation intermediates formed under quiescent conditions appear to be similar to Aβ oligomer species reported by others such as ADDLs or spherical globules. They also may be the micelles suggested by others to be part of the Aβ aggregation pathway from which nuclei form. The same aggregation intermediates had low β-sheet content, low stability in denaturants, and high toxicity. This work provides insight into the structure of toxic Aβ oligomers and highlights differences in aggregation mechanisms of Aβ fibrils formed with and without agitation.