Among several neurological disorders, AD is one of the most common forms of dementia. Even though the causes of the Alzheimer’s onset and progression are still under debate, according to the amyloid cascade hypothesis [1
], the amyloidogenic process that leads to the formation of extracellular aggregates of Aβ peptides is considered one of the main markers of Alzheimer’s occurrence and severity. Until now, only two strategies are used to provide symptomatic relief to AD patients: acetylcholinesterase inhibitors, to maintain the level of acetylcholine in the brain, and N-methyl-D-aspartate receptor antagonists, to prevent excitotoxicity [59
]. Unfortunately, serious side effects and poor effectiveness in some phases of the disease have been detected [40
]. Between different lengths of ordered and disordered amyloid peptides, the Aβ42
fibril is known to be the most toxic due to its tendency to self-assembly into ordered structures [5
]. Moreover, this structure presents two different structural rearrangements, S-shaped and the U-shaped [54
]. In the last years, several studies have investigated the different behaviour of these polymorphisms and the S-shaped form has demonstrated greater conformational and mechanical stability than the U-shaped form [11
]. Therefore, the S-shaped structure represents the primary target for pharmacological treatments, aimed to reduce the amyloidogenic process and interfere with the amyloid aggregates’ stability. In this context, the search for destabilizers of Aβ fibrils may provide fruitful insights in the open research for treatments targeting AD. Natural compounds have shown promising effects, proving to be effective in many in vitro and in vivo studies with minimal side effects and increased blood brain barrier permeability [35
]. However, the molecular mechanism of action of these compounds is still unclear and several computational studies have tried to characterize their effects on different amyloid aggregates [31
]. In this work, a combination of ensemble docking and MD simulations has been applied to evaluate the influence of 57 promising compounds on preformed S-shape Aβ42
]. We identify three different mechanisms of action for the best ten natural compounds: (I) inter-chain destabilization, (II) pocket distortion and (III) pocket stabilization. In particular, 6-shogaol and oleuropein (mechanism I) are able to disrupt the protein ordered structure docking between adjacent fibril chains; curcumin, gossypin and piceatannol (mechanism II) dock into a binding pocket identified by the amyloid S-shape, affecting the whole protein conformation; the other ligands (mechanism III), instead, preserve or slightly influence the conformational state of the amyloid fibril. In this way, we find out that only 6-shogaol, oleuropein, curcumin, gossypin and piceatannol appreciably affect the protein stability, reducing the percentual content of beta sheets, the order parameter value and the inter-chain interaction area if compared to the wild type structure. It is worth remarking that the ligands belonging to mechanisms I and II have demonstrated similar conformational effects on the amyloid fibril, inducing similar reductions of beta-sheet structure content, order parameter and inter-chain interaction area. Therefore, all the identified destabilizing compounds, i.e., 6-shogaol, curcumin, gossypin, oleuropein and piceatannol, are underlined for further investigations. Moreover, compound shared features may be used for determining a pharmacophore model to rationally design novel compounds, hopefully characterized by a more effective destabilizing strength on Aβ toxic assemblies. To remark on the importance of the selected compounds’ chemical features, it is worth mentioning that brazilin, a modulator of the amyloid fibril conformation [62
], shares five common features with the here characterized destabilizing ligands belonging to classes I and II (Figure S5
The remaining compounds seem to stabilize the binding pocket, maintaining an ordered structure of the amyloid aggregate. Concerning previous literature [31
], the present research expresses some aspects of novelty. In particular, this work considers the S-shaped polymorphism as the ligand target. Previous computational works have mostly studied a different amyloid polymorphism, namely the U-shaped one [47
], which might be less stable than the S-shaped one [11
]. Moreover, this work also provides a comprehensive comparative investigation on a considerable number of natural inhibitors, investigating their binding and action mechanisms.
Most of the selected compounds have shown antioxidant and anti-inflammatory properties in vivo [3
]. Recent studies have remarked the destabilization action of curcumin on Aβ40
], also with advanced amyloid accumulation [63
]. Furthermore, it has been observed that oleuropein acts against the formation of toxic oligomer and amyloid fibrils, favouring the formation of non-toxic aggregates and improving cognitive functions [3
]. It is worth mentioning that mechanism-III compounds have proven beneficial effects on AD onset and progression. Therefore, their mechanism of action probably alters the fibril structure in a different stage of the pathology or mostly affects other important factors of the disease, including the oxidative stress, the tau hyperphosphorylation, the α-secretase expression and the β-secretase activity.
Most ligands, except for 6-shogaol and oleuropein, interact with common residues in two main binding areas, identified by residues E11–F19 and I32–L34. Therefore, these residues seem crucial for the definition of the binding pocket and for the effective binding of the investigated compounds. Similar regions have been identified by previous studies about curcumin-like compounds in a complex with the U-shape polymorphism [51
]. Moreover, in vitro studies have shown the key role of F19 and F20 for efficient Aβ42
]. Finally, the pharmacophore modelling highlights common features between the destabilizing compounds, outlining the presence of an additional hydrophobic characteristic for the mechanism-I compounds. This feature could probably be related to the ability of 6-shogaol and oleuropein to fit between protein chains. All ligands share aromatic characteristics that have been seen to be important for interacting with amyloidogenic aggregates [74
]. Moreover, the presence of three H-bond acceptor features is common to all the destabilizing compounds, but not for the other ones, suggesting that this could be an important characteristic for an effective binding to the amyloid fibril and for the activity of the investigated chemicals.
As a limit of the present research, it is worth mentioning that molecular simulations employed in this work cannot represent the entire amyloidogenic process and the present study is focused on the action of natural compounds on preformed amyloid fibrils. Only experimental studies may access the proper time and length scales to correctly describe the whole fibrillogenic process. However, the present research represents a meaningful comparative investigation with atomic resolution, which helps the ligand screening workflow by elucidating binding and action mechanisms of a considerable number of existing natural compounds already considered in the AD research field. Future developments might consider combinations of different compounds, as well as the effect of different compound concentrations to better clarify potential cooperative or competitive mechanisms.