Destabilization of Aβ fibrils by omega-3 polyunsaturated fatty acids: a molecular dynamics study

Abstract The senile plaques of neurotoxic aggregates of Aβ protein, deposited extraneuronally, mark the pathological hallmark of Alzheimer’s disease (AD). The natural compounds such as omega-3 (ω-3) polyunsaturated fatty acids (PUFAs), which can access blood–brain barrier, are believed to be potential disruptors of preformed Aβ fibrils to cure AD with unknown mechanism. Herein, we present the destabilization potential of three ω-3 PUFAs, viz. Eicosapentaenoic acid (EPA), Docosahexaenoic acid (HXA), and α-linolenic acid (LNL) by molecular dynamics simulation. After an initial testing of 300 ns, EPA and HXA have been considered further for extended production run time, 500 ns. The increased value of root mean square deviation (RMSD), radius of gyration, and solvent-accessible surface area (SASA), the reduced number of H-bonds and β-sheet content, and disruption of salt bridges and hydrophobic contacts establish the binding of these ligands to Aβ fibril leading to destabilization. The polar head was found to interact with positively charged lysine (K28) residue in the fibril. However, the hydrophobicity of the long aliphatic tail competes with the intrinsic hydrophobic interactions of Aβ fibril. This amphiphilic nature of EPA and HXA led to the breaking of inherent hydrophobic contacts and formation of new bonds between the tail of PUFA and hydrophobic residues of Aβ fibril, leading to the destabilization of fibril. The Molecular Mechanics Poisson–Boltzmann Surface Area (MM-PBSA) results explain the binding of EPA and HXA to Aβ fibril by interacting with different residues. The destabilization potential of EPA and HXA establishes them as promising drug leads to cure AD, and encourages prospecting of other fatty acids for therapeutic intervention in AD. Communicated by Ramaswamy H. Sarma


Introduction
Misfolded proteins that agglomerates abnormally (Laskowska et al., 2019) to form insoluble aggregates is the key characteristic of various neurodegenerative disorders, such as Alzheimer's disease (AD), Parkinson's disease (PD), and Huntington's disease (HD), which hinders the normal functioning of the brain (Shamsi et al., 2017). A multi-faceted chronic, progressive, and geriatric disease with unclear root causes and mechanisms describes AD (Morley et al., 2018). This is the most prevalent form of dementia across the global population and increasing at an uncontrolled rate. Dementia is described as cognitive impairment causing memory loss leading to poor judgment, inability to process daily information, and social withdrawal, affecting the normal life of a person. The devitalizing effects of dementia manifests into personal, societal, and global healthcare burden, adhered with AD (Grøntvedt et al., 2018). The amyloid precursor protein (APP) upon enzymatic cleavage leads to the release of Ab monomers (37-42 residue long), which upon aggregation forms neurotoxic Ab fibrils (Ye, 2013). The clinical signature of AD lies within extraneuronal insoluble senile plaques consisting of Ab fibrils (Ono, 2018) and intraneuronal hyperphosphorylated tau protein tangles (NFTs) ( Simi c et al., 2016). The Ab fibril is of greater clinical importance as it is the foremost step in aggregation and also directs the formation of NFTs further, as explained by amyloid cascade hypothesis (Findeis, 2007).
The concern surrounding Ab agglomeration and deposition can be dealt with different therapeutic approaches. These approaches are broadly classified as inhibition of fibrillation of Ab monomers and destabilization of preformed Ab fibrils (Cao et al., 2018). The failure of drugs in clinical trials, employing an anti-aggregation strategy, encouraged research on destabilization as an alternate therapeutic approach (Mehta et al., 2017). The destabilization of Ab fibrils serves the dual purpose of deforming aggregated fibrils and inhibiting the formation of higher order aggregates further (Rivi ere et al., 2009). The disruption of preformed Ab fibrils can be carried out by various small molecules (Shuaib & Goyal, 2018), nanoparticles (Sudhakar & Mani, 2019), peptide-based prodrug (Paul et al., 2018), and natural compounds (Habtemariam, 2019). Amongst these various possible therapeutics, natural compounds have been a preferred choice owing to their biocompatibility and intoxicity to human system (Andrade et al., 2019). Multivariate health benefits such as antioxidant, anti-cancerous, and neuroprotection make these natural compounds even more suited as drugs for a wide array of human diseases, including AD (Deshpande et al., 2019). However, there is a dire need to investigate the drugability, efficacy, and nutraceutical applicability of these compounds, to be used as drugs.
The dietary interventions of healthy foods are reported to delay, arrest, and reverse the progression of cognitive impairment in humans (Bhatti et al., 2020). Numerous epidemiological studies have also fostered the positive correlation between high adherence to Mediterranean diet (MeDi) with lower risk of cognitive decline, lower incidences of AD, enhanced learning, and memory function and lower mortality rate in AD (Karstens et al., 2019). MeDi is rich in polyphenols, flavonoids, and good amount of protein, due to inclusion of various fruits, vegetables, nuts, and fish. MeDi also provides a great source of healthy essential fatty acids (FAs) as well due to the presence of fish. Essential FAs are the ones that cannot be synthesized by humans and hence demands dietary consumption (N. Kaur et al., 2014) from plants or other natural sources. The omega-3 (x-3) FAs are one such essential FAs which are long-chain polyunsaturated FAs (PUFAs), having double bonds from third carbon from the methyl end. The prime precursor for x-3 FAs in humans is a-linolenic acid (LNL, C18:3) (Tocher et al., 2019). The LNL is plant-based PUFA found majorly in chia seeds, flaxseeds, soybean seeds, walnuts, and almonds (Abete et al., 2009) which gets metabolized inside the human body to Eicosapentaenoic acid (EPA, C20:5) and Docosahexaenoic acid (HXA, C22:6) but with limited percentage of conversion (Brenna, 2002). This limited conversion and less bioavailability can be compensated by dietary supplementation of EPA and HXA. The rich sources of EPA and HXA are mainly marine fish or fish oil (Derbyshire, 2019), macroalgae, and microalgae (Santigosa et al., 2020).
The critical role of x-3 PUFAs particularly EPA and HXA, in memory and learning (Fonteh et al., 2020), anti-inflammation in central nervous system (CNS) disorders (Bogie et al., 2020;Lay e et al., 2018), neurotrophicity (Benatti et al., 2004), prevention of psychosis in adults (Mongan et al., 2021), and lower incidence of AD (van Lent et al., 2021) is exemplary. Being the major constituent of the neuronal cells, PUFAs are involved in membrane fluidity and optimal synaptic communication between cells and neuroprotection (Luo et al., 2018). The low levels of these PUFAs are reported to have implications in overall neuropsychiatric health including AD (Aarti & Priyanka, 2018;Bove et al., 2018;Song et al., 2016), thereby causing cognitive decline and commencing dementia. The fish consumption has been recommended as a part of MeDi owing to its higher contents of EPA and HXA, for maintaining cognitive resilience and memory-associated functions . There is a crucial evidence on permeation of EPA and HXA through blood-brain barrier (BBB) that helps in maintaining membrane integrity (Ouellet et al., 2009). Passage through BBB, EPA, and HXA also reinforces Ab clearance from brain to blood, thereby obviating Ab accumulation in brain . These findings emphasizes on consumption of fish, walnuts, and almonds, rich in EPA and HXA for overall good health of BBB and brain (Ren et al., 2017). In addition to that, diet supplementation with EPA and HXA was also found supportive in reducing the amyloid burden in older rats (Lim, 2005) and conversely their deficiency invokes depression and loss of memory (Bove et al., 2018). These PUFAs were also reported to induce microglial phagocytosis of Ab peptides (Hjorth et al., 2013). In vitro and cytotoxic studies on SH-SY5Y cell lines revealed the inhibitory potential of HXA by preventing Ab 1-42 polymerization (Miwa et al., 2013). A recent report on direct interaction of EPA and HXA with Ab peptides led to 16-84% inhibition in fibrillogenesis of the peptides (El Shatshat et al., 2019). These PUFAs also have stimulatory effect on insulin-degrading enzyme (IDE) leading to degradation of Ab peptide (Grimm et al., 2016). The reduction of amyloidogenic processing of APP by stimulating b and c secretase by HXA (Grimm et al., 2011) consonances well with the beneficial role of these x-3 fatty acid as possible drug leads for the treatment of AD.
The computational studies not only traverse atomic-scale mechanisms by means of molecular dynamics (MD) simulations but also validates results obtained by in vitro studies as well (Nasica-Labouze et al., 2015). A simulation study that reported adsorption of Ab peptides over the cell membrane in the presence of PUFAs causes weaker interactions between the peptides itself. This leads to the inhibition of aggregation of Ab peptides, thereby restricting the progression of AD (Ntarakas et al., 2019). A recent study on inhibition of aggregation prone fragment (KLVFFA) of Ab peptide by HXA helps in understanding its inhibitory potential (Zhou et al., 2018). The beneficial effects of PUFAs in improving AD by various routes have been provided; however, the direct interaction and destabilization of Ab fibrils by EPA and HXA is still not properly understood. The unavailability of enough reports instigates the need for more computational-based studies to gain insight into the plausible destabilization mechanism of Ab fibrils by PUFAs.
In the present work, the focus has been led upon investigating the atomistic details of the destabilization mechanism of Ab fibrils by x-3 PUFAs. The present study was conducted by MD simulation, which facilitates drug designing in lieu to formulate a clinically approved drug to treat AD. The substantial increase in root mean square deviation (RMSD) and radius of gyration (R g ) demarcates disorganization of Ab fibril structure in the presence of EPA and HXA as compared to LNL when each of the complex systems was simulated for 300 ns. To acquire more in-depth details, the Ab-EPA and Ab-HXA complexes have been subjected to longer simulation time of 500 ns. The disruption of hydrogen bonds (H-bonds), salt bridges, and breaking of hydrophobic contacts confirms destabilization of the Ab fibril in the presence of EPA and HXA. These PUFAs were observed to insert themselves in the hydrophobic cavity of the fibril owing to their amphipathic property. The insertion in the cavity is due to flexibility of PUFAs and the chemical interaction involving different residues in Ab fibril highlights amphiphilic nature of EPA and HXA. The hydrophilic carboxylic head is found to interact with positively charged lysine (K) residue, whereas the hydrophobic carbon tail manages to form bonds with other hydrophobic residues in the fibril well testified by Molecular Mechanics Poisson-Boltzmann Surface Area (MM-PBSA) analysis. These interactions of PUFAs with different residues in Ab fibril competes with native interactions of Ab fibril leading to its destabilization. The results thus obtained after MD simulation establishes both EPA and HXA as potent disruptors of Ab fibril and thus a promising drug lead for the treatment of AD.

Selection of PUFAs and Ab 17-42 fibril for molecular dynamics simulation
Three-dimensional structure of Ab 17-42 fibril (PDB ID: 2BEG) (Luhrs et al., 2005) was taken from the protein data bank (PDB) and was used as the receptor protein in the current study. The Ab fibril is a pentamer structure spanning from residues 17-42, which form the characteristic b-sheet, the prerequisite for fibrillation, and stable organization of amyloid structures. The characteristic b-sheet structure is further divided into three regions-b1 region (18-26 residues), turn region (27-30 residues), and b2 region (31-42 residues) as depicted in Figure 1(a) as derived from visual molecular dynamics (VMD) package (Humphrey et al., 1996).
There are five chains in 2BEG, designated as A, B, C, D, and E, having the same primary sequence, with A and E being the terminal chains. 2BEG as the representative model for Ab fibril, employed in numerous studies owing to its well-established and stable structure (Gupta & Dasmahapatra, 2019;Kuang et al., 2015;Lemkul & Bevan, 2010;Saini et al., 2019). The ligands that are investigated in the study are x-3 PUFAs, viz. EPA, HXA, and LNL. The PDB structures for all of these PUFAs have been taken from the ligand summary page as EPA.pdb, HXA.pdb, and LNL.pdb, respectively, and drawn by using ChemDraw (Cousins, 2011) as shown in Figure 1 The definition for all the systems, for MD simulation, has been depicted in Table 1 as S0, S1, S2, and S3 where S0 represents the control system, i.e. without ligand.

Molecular dynamics simulation of PUFAs with Ab fibril
The interaction between PUFAs and Ab fibril (x-3-Ab) has been studied by MD simulation using GROMACS v5.1.1 (Van Der Spoel et al., 2005), with GROMOS 96 54a7 (Huang et al., 2011;Oostenbrink et al., 2004;Somavarapu & Kepp, 2015) force field. The same force field has been employed in various other studies as well for system definition (Barale et al., 2019;Du et al., 2015;Lemkul & Bevan, 2010;Liu et al., 2011;Oostenbrink et al., 2004;Orteca et al., 2018;Saini et al., 2017;Tavanti et al., 2018). The ligand topologies obtained from rscb pdb database have been submitted to PRODRG server (Sch€ uttelkopf & van Aalten, 2004) to generate gromos defined coordinates, which are GROMACS readable. The PRODRG obtained topologies for EPA, HXA, and LNL have been given in Section 2.1, 2.2, and 2.3, respectively, in Supplementary Information. The topologies thus obtained for PUFAs in .gro format were then combined with Ab fibril by means of random insertion to form protein-ligand complexes: C1 (Ab-EPA), C2 (Ab-HXA), C3 (Ab-LNL), and C0 as (Ab-Water) ( Table 1). All the complexes were solvated in a 7 Â 7 Â 7 nm 3 cubic box with water, represented by the SPC/ E water model (Berendsen et al., 1987). Subsequently, adequate numbers of sodium ions were added to neutralize the system. Following this, the energy minimization by the steepest descent method was carried out, equilibrated further by NVT followed by NPT ensemble at 300 K and 1 atm pressure, respectively (Bussi et al., 2007;Nos e & Klein, 1983). The equilibrated system obtained thereafter, was submitted for production run for 300 ns with a time step of 2 fs. The periodic boundary conditions were applied in all the directions for all the systems. To check for the statistical significance and reproducibility of the results, three sets for each system have been simulated, and the data are represented as an average over three sets. Further, the simulation time has been increased to 500 ns for the selected PUFAs amongst investigated molecules to gain insights into the mechanism of the interaction between protein and ligand. The visualizations of MD trajectories were analyzed using VMD.

Analysis of MD-generated trajectories
We have calculated various parameters using different tools defined within the GROMACS v5.1.1 package. The global structural stability of all the x-3-Ab complexes was determined by Ca-RMSD by g_rms. R g (radius of gyration) value was calculated by g_gyrate tool, and solvent-accessible surface area (SASA) has also been used to enumerate detailed structural changes in Ab fibril by the PUFAs. These observations led to the elimination of the least potent PUFA, thereby extending the run time for other PUFAs only. The extended simulation for two PUFA molecules for 500 ns helps in capturing the time-dependent changes. The destabilization potential of the screened PUFA molecule was further accessed by evaluating the secondary structure content by the DSSP tool. The destruction of hydrogen bonds determined by g_hbond tool gives an idea on the decrease in the structural strength of Ab fibril in the presence of the ligand. The salt bridges were determined for all the neighboring pairs for all the chains by g_mindist module of GROMACS package. The intra-and interchain hydrophobic contacts for different residue pairs were estimated for all the peptide chains. The interchain distance for neighboring terminal chains have been calculated by g_mdmat to gain knowledge about the possible interactions via a distance matrix.

Binding mode analysis by MM-PBSA
The binding free energy (BFE) of the Ab fibril with the selected PUFA was calculated by MM-PBSA (Genheden & Ryde, 2015;Gohlke et al., 2003;Homeyer & Gohlke, 2012). The package was used within the GROMACS environment by g_mmpbsa tool, which calculated the overall binding energy and individual residue contribution (Kumari et al., 2014). MM-PBSA calculates the BFE from the MD trajectory based on the following equation: where DE MM is the molecular mechanics contribution in vacuum, DG psolv is the polar contribution of solvation energy, calculated by the Poisson-Boltzmann equation, and DG npsolv is the nonpolar contribution toward binding, determined by SASA model. The high computational demand limits the calculation of entropy contribution. Moreover, the relative change in entropy contribution (DS) toward free energy calculation is negligible and thus can be ignored (Gupta & Dasmahapatra, 2020;Kanchi & Dasmahapatra, 2019;A. Kaur et al., 2020;Zou et al., 2019). The calculation of BFE in the present study was done on the last 20 ns trajectory with Dt ¼ 100 ps to reduce the computational time. The energetic contribution of individual residues was also quantified to identify key residues actively involved in binding with PUFAs. The solute and solvent dielectric constants were taken as 4 and 80, respectively.

Molecular dynamics simulation of PUFAs with Ab fibril
The simulation studies were conducted for 300 ns for all the three PUFA compounds to test their destabilization potency.
The system definition has been tabulated in Table 1 in the Methods section. The global stability parameters, such as Ca-RMSD, Rg, and SASA, have been calculated for all the complex systems, which demarcated a trend of EPA and HXA being destabilizers of equal strength as compared to LNL. The Ca-RMSD value for Ab fibril has been calculated for the entire protein, and for the terminal chain E, separately. The higher value is observed from the time trajectory of Ca-RMSD for the entire protein (Figure 2(a)) and for the chain E (Figure 2(b)) for Ab-EPA and Ab-HXA systems as compared to the Ab-Water. This increase in RMSD clearly establishes the destabilization of the Ab fibril in the presence of EPA and HXA. However, LNL is found to be not much effective as a potential disruptor. The corresponding average Ca-RMSD values for the entire protein and for the chain E has been plotted for all the systems under investigation as probability density functions (PDFs) in Figure 2(c) and (d).
The increase in average RMSD value of 0.967 nm for Ab-EPA system cause shift in the curve as compared to Ab-Water system. From Figure 2(c), it is clearly observed that Ab-LNL system records sharpest peak with average value of 0.810 nm close to 0.821 nm of Ab-Water system. The marginal increased RMSD value of 0.875 nm in Ab-HXA system directs toward further investigation. Similarly, from Figure 2(d), increased RMSD value for chain E highlights on the potential of EPA and HXA as better destabilizer as compared to LNL. The demonstrating structural instability of Ab fibril is also reported by other natural compounds which supports our findings (Saini et al., 2019).
The higher Rg value for Ab-EPA and Ab-HXA as compared to the Ab-Water system (Figure 3(a,c)) indicates the decreased compactness of the fibrillar structure in the presence of these PUFAs. The trend of SASA (Figure 3(b,d)) may 2BEG.pdb (Ab fibril) EPA.pdb (Eicosapentanic acid) not be enough to claim destabilization potential of HXA. However, a marginal increase of SASA by EPA and decrease in the presence of LNL motivates further testing of EPA and HXA, discarding LNL at this stage itself. A similar increase in Rg and SASA values of Ab protofibril by a flavonoid derivative  and arginine containing short peptides (Barale et al., 2019), respectively, leading to destabilization of Ab protofibril corroborates well with our findings. These observations helped in eliminating LNL for extended run and thereby restricting further investigations to EPA and HXA only to look for the better destabilizer amongst the two. The extension of production runs up to 500 ns for three systems-Ab-Water (C0), Ab-EPA (C1), and Ab-HXA (C2)-has been visualized and computed with several other parameters to gain atomistic details on the destabilization mechanism of these PUFAs. The VMD visualization of the final configuration of C0, C1, and C2 system has been presented in Figure  4(a-c), respectively. The snapshots (Figure 4(a-c)) of the trajectory for Ab fibril showed binding of the PUFA molecules with Ab fibril leading to high degree of perturbations in the fibril indicating destabilization potential of EPA and HXA.
The binding possibility can be explained from the inherent conformational flexibility of the n-3 PUFA molecules, which introduce fluidity in the biomembrane upon incorporation. Such modification in the microenvironment of the biomembrane leads to execution of wide range of biological functions (Law et al., 2005). The presence of different side chain groups in Ab fibril and amphiphilic PUFA molecule makes the conformational rearrangement is such a way that facilitates effective interactions between these two (Shaikh et al., 2015). Both the ligands are observed in linear conformation with disoriented Ab fibril, establishing the destabilization of the fibril. A similar disorganized conformation of Ab fibril was observed in the presence of caffeine (Gupta & Dasmahapatra, 2019), which supports the current observation. Both the EPA and HXA molecules were observed to access hydrophobic cavity of the Ab fibril via entry from the terminal chain, which is in parity with the interaction behavior of other phenolic compounds . The ability to insert themselves in the fibrillar cavity is associated with intrinsic flexibility of these molecules (Rustan & Drevon, 2005;Stillwell & Wassall, 2003). In addition to the final trajectory, the snapshots at 300 ns and 400 ns have been illustrated in Figure S1 (Section 1), Supplementary Information ( Figure S1a and b for C0, Figure S1c and d for C1 and Figure  S1e and f for C2 system, respectively). The Ab fibril in the presence of water showed a little twisting at the ends of the    The C a -RMSD, R g , and SASA have been analyzed to check the integrity of the Ab fibril in the presence of EPA and HXA for an extended production run of 500 ns. The time trajectory and corresponding average Ca-RMSD value for the entire protein, chain E and turn region has been plotted for all the three systems in Figure 5(a-c) and (d-f), respectively. The average value for Ca-RMSD for the entire protein as depicted in Figure 5(d) was recorded as 0.760, 1.016, and 0.918 nm for C0, C1, and C2 systems, respectively. A similar increase in Ca-RMSD value has been observed for the terminal chain E (Figure 5(e)), and the turn region ( Figure 5(f)) of the Ab fibril in the presence of EPA and HXA as compared to the Ab-Water system.
The current observations resonate well with high Ca-RMSD value on account of destabilization of Ab fibril by a natural novel drug, wgx-50 (Fan et al., 2015). The color scheme considered for three systems is blue for Ab-Water, red for Ab-EPA, and green for Ab-HXA system, which is maintained in other parametric analyses as well, if not mentioned otherwise. The observations from Figure 4 and Figure 5 emphasize on representation of chain E as the most affected chain of the pentamer. Henceforth, various parameters have been depicted for chain E separately to gauge the extent of destabilization.
The Rg value, indicative of the flexibility, and SASA value, giving an account of hydrophobic and hydrophilic composition of a protein structure (Berhanu & Hansmann, 2012), have been determined. These two parameters have been plotted as time trajectory and interval plots with 95% confidence interval (CI) around mean value in Figure 6. Herein, Figure 6(a) and (c) represents Rg, and Figure 6(b) and (d) denotes SASA value for all the three systems. The mean Rg value of 1.377 nm, 1.388 nm, and 1.368 nm for C0, C1, and C2 system have been plotted in Figure 6(a) and (c) indicates consideration of EPA as a potential destabilizer. The increased SASA value observed for Ab-EPA system from Ab-Water system (see Figure 6(b,d)) compliments the selection of EPA as a fibrillar disruptor. However, a clear and detailed picture about role of HXA has been discussed in the next sections.
The increased RMSD, R g , and SASA values demonstrating structural instability of Ab fibril is also reported by other natural compounds which supports our findings (Saini et al., 2019). Conclusively, higher values observed for Ca-RMSD, Rg, and SASA translates the potency of both PUFA molecules (EPA and HXA) as a promising destabilizing agent of Ab fibrils, thereby curing AD. All these global stability parameters demand further in-depth analysis to gain insight on the molecular details of the interaction between PUFAs and Ab fibril.

Secondary structure determination
The secondary structure plays a decisive role in the overall stability, longevity, and functionality of the protein. The stability of Ab fibrils has been attributed to their higher b-sheet content, which directs the formation of an organized b-fibrillar structure. The DSSP (Kabsch & Sander, 1983) tool delivers the time evolution profile of possible secondary structure configurations achievable by Ab fibril. The percentage content of the different possible secondary structures for all the three systems C0, C1, and C2 has been tabulated in Table 2. The coil content for C1 and C2 was recorded as 45% and 42%, which is considerably higher than C0 (37%). Contrarily, a marked decrease in b-sheet and b-bridge content for C1 and C2 systems as compared to C0 (Table 2) implicates the loss of the structural stability and integrity of the Ab fibril.
The secondary structure profiling for Ab-Water, Ab-EPA, and Ab-HXA has been shown in Figure 7(a-c), wherein a prominent decrease in b-sheet along with rise in coil and turn content are recorded for the C1 and C2 systems as compared to the C0. The significant decrease in the b-sheet content ensures an obstruction to the formation of higher order aggregates barring neurotoxicity of Ab fibrils.
A similar reduction in b-sheet content of Ab fibril leading to its destabilization in the presence of caffeine (Gupta & Dasmahapatra, 2019), ellagic acid (Gupta & Dasmahapatra, 2020), D744 (a fluorinated derivative of curcumin (Saini et al., 2017)), and norepinephrine (Zou et al., 2019) is found to be coherent with the current findings. The loss in b-sheet content and remarkable increase in the coil content due to PUFAs makes them potential anti-amyloidogenic drug lead.

Backbone stability
The integrity of the fibrillar backbone, gauged by the probability of bond formation owing to the distance between Ca of K28 residues of neighboring chains, was determined. The same has been plotted as time trajectory (Figure 8(a,b)) and as corresponding bar charts (Figure 8(c) and (d)) based on the mean value. The average K28-K28 distance for neighboring terminal chains (viz. A-B and D-E) was found to be increased beyond the cutoff limit of bond formation in the presence of EPA and HXA. Figure 8(a) and (c) depicts striking increase in K28-K28 distance from 0.49 nm (Ab-Water) to 0.65 and 0.64 nm for Ab-EPA and Ab-HXA for chain A-B, respectively. A similar noticeable increase for chain D-E has been observed from time trajectory and bar charts from Figure  8(b) and (d), respectively, as 0.85 nm for Ab-EPA from 0.39 nm from Ab-Water (Figure 8(b,d)). Such an enhancement in the distance between these residues conciliates bond formation and hence compromising on the stability of the Ab fibril.

Effect of PUFAs on various bonds in Ab fibril
There are various types of intra-and intermolecular bonds that imparts stability to any protein structure at different levels (viz. primary to quaternary). These bonds broadly encompass H-bonds, covalent bonds, salt bridges, and hydrophobic contacts. Herein, we present a detailed analysis of the change in these molecular interactions in the presence of EPA and HXA.
Hydrogen-bonds: The theoretical (Zheng et al., 2007) and experimental (Sunde et al., 1997) studies outline the importance of an extensive H-bond network contributing toward the stability of Ab fibrils. We have calculated the average number of H-bonds for the entire protein, b1, and turn region, which showed a decreasing tendency in the presence of EPA and HXA molecule (refer Table 3). The average number of H-bonds for interpeptide chain D-E was observed to be 10.1 and 10.4 for C1 and C2 system, which is less as compared to 11.3 for C0. A similar decrease of 31.74% and 29.06% in C1 and C2 systems, respectively, from C0 (Table 3) for intrachain E peptide, testifies the degradation of H-bonds in Ab fibril upon interaction with PUFAs. These observations are in close agreement with the destabilization of Ab fibril by wgx-50, a novel compound from Sichuan pepper, which also showed a drastic decrease in both intra-and interchain H-bonds (Fan et al., 2015). The H-bond disruption of Ab fibril in the presence of PUFAs would prohibit further fibrillar growth and elongation.
Salt-bridge formation: Electrostatic interactions between oppositely charged residues in any protein structure are called salt bridges (Donald et al., 2011;Musafia et al., 1995). Salt bridges impart the stability to the protein and influence its structural conformation, pattern recognition, function, degradation pattern, and catalytic activity (Ban et al., 2019). In the present study, we have calculated the average distance between intra and inter D23 and K28 residues for the terminal chain E for all the three systems (viz. C0, C1, and C2). The same plots have been drawn here also, with time trajectory and interval plots with 95% CI around mean value in Figure 9(a-d), respectively. The average distance for these residues was found to be increased in the presence of Ab-EPA (Figure 9(a,c)); however, not much changes were observed in Ab-HXA (Figure 9(b,d)), indicating the ineffectiveness of HXA toward breaking of salt bridges between D23 and K28 residues in the Ab fibril. The inter-residual distance between D23 of chain D and K28 of neighboring chain E was measured as 0.72 nm for Ab-EPA, which is considerably pronounced as compared to 0.28 nm in Ab-Water system (Figure 9(a,c)). A similar increase to 0.37 nm for intrapeptide chain E (D23-K28) for Ab-EPA system from 0.27 nm in Ab-Water system (Figure 9(b,d)) was measured.
Such increased distances indicate bond breaking, and hence misalignment of Ab fibril around the terminal chain. The results thus obtained are found to be concurrent with the breaking of salt bridges of Ab fibril in the presence of copper ions , C-60 fullerene (Andujar et al., 2012), and surfactin (Verma et al., 2016) molecules, leading to the destabilization of Ab fibril. The salt bridge disruption in destabilizing the entire Ab fibril in the presence of brazilin (Du et al., 2015) corroborates with the current findings.
Interchain distance matrix: We have determined the residue-residue contact between terminal chain D and E by contact map analysis. As depicted in Figure 10(a), there is a portrayal of contacts throughout chain D and E for Ab-Water system. These residual contacts were fairly compromised for Ab-EPA and Ab-HXA systems (Figure 10(b,c)), which can be viewed from presence of blank spaces in the matrix.
The illustration of contact map analysis for the Ab-HXA systems provides a well-founded idea on the reduced contacts between chain D and E, thereby moving E chain apart from the parent fibrillar structure. Evidently, the contacts between terminal chains are disrupted, which confirms the destabilization of the Ab fibril. These observations are found in parity with the interchain matrix observed in the presence of flavonoid derivative .
Hydrophobic contacts: The major contributors toward the folding pattern and stability of the protein are electrostatic and hydrophobic interactions. The burying of hydrophobic residues inside the solvent inaccessible cavity provides remarkable high strength to amyloid fibrils (Pace

Secondary structure (%)
Ab-Water Ab-EPA Ab-HXA Coil 37 ± 0.57 45 ± 0.57 42 ± 0.57 b-Sheet 38 ± 0.57 37 ± 0.57 33 ± 0.57 b-Bridge 5 ± 0.57 4 ± 0.57 3 ± 0.57 Bend 14 ± 0.57 12 ± 0.57 16 ± 1 Turn 3 ± 0.57 2 ± 0.00 3 ± 0.00 a-helix 0 ± 0.00 0 ± 0.00 0 ± 0.00 3-helix 0 ± 0.00 1 ± 0.57 0 ± 0.00 Chain_Separator 3 ± 0.00 3 ± 0.00 3 ± 0.00 et al., 2011). According to the principle of Principle of Amyloid Self-Assembly (PASA), the fibrillar structure, which possesses the maximum number of inter-and intrapeptide hydrophobic interactions, is the most stable and favorable fibril structure during fibrillation process (Tarus et al., 2006). There have been various hydrophobic residual contacts that play a crucial role in the structural organization and strengthening of the Ab fibril. The hydrophobic contact pairs reported for the current pentamer Ab structure are A21-V36, L34-V36, and F19-G38 (Luhrs et al., 2005), interacting in both intra-and interpeptide manner. We have calculated mean distance between these residues as a function of time to estimate the probability of bond formation between them. The inter-residual distance between A21 of chain C and V36  of chain D (Figure 11(a,c)) was estimated as 0.52 nm for Ab-Water system, which was increased to 0.72 nm in the presence of EPA, and to 0.75 nm when HXA was inserted, impeding A21-V36 bond formation. A considerable increase in the distance between L34 and V36 residue was also recorded for chain D and E (Figure 11(b,d)). The L34-V36 distance for Ab-Water system was recorded as 0.40 nm, which was increased   to 0.50 and 0.62 in Ab-EPA and Ab-HXA systems, respectively, hindering the bond formation. The shift in the PDFs curves toward higher values of A21-V36 and L34-V36 distance observed for Ab-EPA and Ab-HXA systems in Figure  11(c,d) clearly highlights on the destabilization potential of both the PUFAs. A distinct differentiation from time trajectory and interval plots can also be viewed in chain A-B and B-C for A21-V36 residual contact for C1 and C2 system (Supplementary Information, Figure S2a-d, Section 1), which supports the destabilization of Ab fibril. Furthermore, the average distance of F19-G38 for chain D and E (both intra and inter pairs) has been analyzed and plotted in Figure 12(a-d). The mean distance for neighboring chains C-D (Figure 12(a)) and D-E (Figure 12(b)) pairs was observed to be 0.86 and 0.61 nm for C1 and 0.66 and 0.62 nm for C2 system which is considerably higher as compared to 0.55 nm and 0.43 nm in C0 system. A simultaneous increase in the intraresidual distance for chain D-D and E-E can also be viewed in Figure 12(c,d), respectively. The corresponding time trajectory to the average F19-G38 distance for all these four selections has been plotted in Supplementary Information, Figure S3a-d, Section 1, which well compliments with the above findings.
The average mean distance for the other end of Ab fibril (viz. A-B, B-C, A-A, and C-C) for F19-G38 pair has been plotted in Supplementary Information, Figure S4a-d, Section-1 with corresponding time trajectory in Figure S5a-d, Section 1). Both Figure S4 (a-d) and S5 (a-d) show that there is no bond formation between these residues, which is attributed to an increased distance between F19-G38, thereby disarraying the pentamer structure. All these observations clearly manifest the obstruction in bond formation, which disarrange the organized Ab structure. Similar findings were also recorded in case of chemically synthesized drugs having anti-cholinesterase activity tested in vitro and in silico, causing disruption of preformed Ab fibrils (Brogi et al., 2014). This increased distance between hydrophobic residues for the terminal chains implies breaking of these bonds in the presence of PUFAs, leading to the disruption of the entire fibrillar structure of Ab fibrils.

Binding free energy analysis between Ab fibril and PUFAs
The integration of the MM-PBSA method with MD simulation studies has been done with an objective to gain more insights on the binding affinity of the neighboring terminal chains in the presence of PUFAs. The detailed interchain binding free energy (BFE) analysis for chain A-B and D-E for C0, C1, and C2 has been tabulated in Table 4. The lowest G binding observed for chain A-B (-454.5 ± 8.4 kcal/mol) for Ab-Water system mentioned in Table 4 clearly indicates the highest affinity of these two chains. On the contrary, the higher G binding of -87.5 ± 7.9 kcal/mol and -20.9 ± 8.9 kcal/ mol for Ab-EPA and Ab-HXA, respectively, specifies the lesser binding affinity between chains A-B in presence of these PUFAs. A similar level of separation of terminal chain E from the fibril is attributed to the higher G binding value for Ab-EPA (-151.7 ± 10.3 kcal/mol) and Ab-HXA (-63.7 ± 11.6 kcal/mol) system as compared to -189 ± 10.8 kcal/mol for Ab-Water (Table 4). The estimation of various other associated energy terms, such as, DE MM , DG psolv , and DG npsolv , provide more precision on binding by specific amino acid residues involved in binding (Karami et al., 2017). The DE vdW or chain A-B for Ab-EPA and Ab-HXA systems has been recorded as -243.7 ± 1.6 kcal/mol and -228.2 ± 1.8 kcal/mol, respectively. The DE elec for chain A-B has been calculated as -66.7 ± 2.3 kcal/mol and -3.3 ± 2.3 kcal/mol for Ab-EPA and Ab-HXA, respectively. The notable lower value for DE vdW as compared to DE elec explains the dominance of van der Waals contribution over electrostatic interactions for chain A-B in the presence of both the binders (viz. EPA and HXA). The similar observation has been made for chain D-E as well in presence of both the ligands. The dominance of van der Waals interactions is attributed to the involvement of hydrophobic residues in binding of PUFAs with Ab fibril, which corroborates well with the disruption of hydrophobic interactions, as explained in the above section. The van der Waals interactions dominating over electrostatic ones have also been reported for binding of AZD2184, a potential PET tracer with Ab fibril (2BEG), attributed to the hydrophobic interaction (Kuang et al., 2015).
Further, the individual contribution for the entire Ab fibril has been determined and illustrated in Figure 13(a,b) for EPA and HXA, respectively. This has been done to elucidate the key residues involved in the binding of PUFAs to the Ab fibril. Figure 13(a) unveils the major involvement of L17 and K28 form all the five chains of the pentamer highlighted with different symbols and color. The contribution from V18 and I32 from chain A, L34 from chain B, and V36 from chain C indicates the hydrophobic interactions pertaining to the hydrophobic residues involved. Similarly, for HXA binding Figure 13(b) demarcates L17 and K28 from all five chains. Additionally, the contribution form V18, F19, and I31 from Chain A; V18, I32, and L34 from chain B; and L34 from chain C highlights the van der Waals interaction. These key residues explain the dual contribution of nonpolar and electrostatic interactions toward Ab-EPA and Ab-HXA complex formation (Liu et al., 2012).
The individual residues involved in binding of EPA to Abfibril are L17 and V18 from b1 region, K28 from turn region and I32, L34, and V36 from b2 region, from all the five chains. Similarly, the contribution from L17, V18, and F19 (b1 region), K28 (turn region) and I31 and L34 (b2 region) from Ab fibril in binding to HXA has been observed. The residues involved in binding synchronize with the breaking of native salt bridges and various hydrophobic contacts present in the Ab fibril. The involvement of L17, V18, and F19 residues in Ab-HXA complex formation is in close agreement for Ab peptide and HXA binding, leading to inhibition of fibrillation (Zhou et al., 2018). The presence of same residues were also reported in binding of PUFAs with Ab fibril (El Shatshat et al., 2019) studied by docking, clearly testifies our results. The binding pattern of different aromatic PUFAs, such as curcumin derivatives (Orteca et al., 2018) and biophenols , to Ab fibril supports the current observation. Our results well corroborate with the binding pattern of oligoproline (Kanchi & Dasmahapatra, 2019) with Ab fibril and di-triazole derivate (A. Kaur et al., 2020) with the terminal chain leading to the destabilization of the fibrillary structure. The MM-PBSA results reveal the mechanistic details on the binding of PUFAs to the Ab fibril by interacting with different residues over the entire chain length. The access to fibrillar cavity by EPA and HXA led to destabilization of Ab fibril, establishing it as a promising therapeutic agent for the treatment of AD.

Conclusions
The destabilization of amyloid fibrils is considered to be one of the most promising therapeutic strategies for AD. The natural compounds are believed to be suitable drug candidates owing to their biocompatibility and dietary consumption. The present work focuses on the screening of three major x-3 fatty acids (PUFAs) for their destabilization potential on Ab fibrils. The anti-amyloidogenic properties of three PUFAs, viz. EPA, HXA, and LNL, have been studied by MD simulation. After initial screening of PUFAs for 300 ns, EPA and HXA were considered for further extended run up to 500 ns. The VMD representation of the MD trajectory thus obtained for both Ab-EPA and Ab-HXA systems showed maximum perturbations and deflections of the terminal chains of Ab fibril due to the insertion of these PUFAs in the fibrillar cavity. The increased values of RMSD, Rg, and SASA in the presence of EPA and HXA as compared to Ab-Water system indicate a disoriented structure of Ab fibril. The increased distance between D23-K28 inhibits the formation of inter and intrachain salt bridges by EPA, without significant effect of HXA. However, the destabilization effect of both PUFAs was verified by loss in the number of H-bonds and b-sheet content. The increase in average distance between hydrophobic residues (viz. K28-K28, A21-V36, and F19-G38) prevents sustenance of these interactions, thus promoting destabilization of Ab fibril in the presence of PUFAs. This explains the destabilization potential of both EPA and HXA on Ab fibril. The BFE analysis by MM-PBSA tool demarcates less binding affinity of the terminal chains A-B and D-E in the presence of EPA and HXA compared to the control system, pertinent to high binding energy. The involvement of positively charged residue (K28) and hydrophobic residues (viz. L17, V18, F19, I31, I32, L34, and V36) collectively elucidates electrostatic and van der Waals interactions between PUFAs and Ab fibril, owing to amphiphilic nature of the PUFAs. The electrostatic interaction is observed between polar carboxylic head and K28 residue, whereas the long non-polar carbon tail interacts with hydrophobic residues (L17, I31, I32, L34, and V36). The minor dominance of van der Waals interactions over electrostatic interaction can be assessed from the higher contribution of hydrophobic residues involved in binding. This led to the insertion of these two PUFA molecules in the fibrillar cavity, thereby assisting the destabilization of the pentamer. The deflection of terminal chains, A and E from the parent pentamer structure is observed on account of the decreased binding affinity, attributed to the increase in the BFE as measured by MM-PBSA method.
The deformation of Ab fibril by x-3 PUFAs not only curbs the neurotoxicity concerns but also inhibits the formation of higher order toxic aggregates upon interacting with upcoming monomers. The present work provides the mechanistic details on the significant role of the PUFAs in destabilizing preformed Ab fibril, thereby providing a promising cure to AD. The better interaction and destabilization of Ab fibril due to amphiphilic fatty acid makes them promising drugs to cure AD, as they can access fibrillar cavity much better than other small molecules. These findings pave the pathway for the investigation of other natural amphiphilic compounds to measure their destabilization potency. Such compounds can be utilized for better targeting of Ab fibrils, thereby providing suitable treatment for AD.

Supplementary information
The Supplementary Information has been divided in two sections, viz. Section 1 and Section 2. Section 1 contains five figures - Figures S1, S2 Time trajectory of F19-G38 Distance for Ab in water, EPA and HXA ligands for (a) chain A-B, (b) chain B-C, (c) chain A-A, and (d) chain C-C. The PRODRG obtained topologies for the ligands: EPA, HXA and LNL has been given in Section 2 as 2.1, 2.2, and 2.3 respectively.