We employed MD simulations to study the behavior of Aβ fibrils under different conditions. MD simulations enable the modeling of individual atom and molecule movement, offering insights into their dynamics. Our study examined on both in the absence of EF and applied EF conditions. The first simulation in ambient conditions was performed, then those with influence of applied EFs. Initially, static EFs with varied intensities (0.20, 0.30, 0.35, and 0.40 V/nm) were applied, guided by prior computational studies, aiming to obtain preliminary data on Aβ fibril responses. Additionally, simulation duration extended up to 1.5 µs (1500 ns) for static EF simulations at the intensity of 0.40 V/nm (see supporting information-SI).First, before commencing the simulations involving an oscillating EF, it was necessary to determine the appropriate frequency value for our system (see Figs. S1–S5). After conducting dielectric spectrum analysis and comprehensively review of prior research, we decided to conduct 100 ns test simulations at the same intensity, with frequencies ranging from 0.1 to 1 GHz (Fig. S1 and S2). Interestingly, the measured RMSD on the Cα atoms was much higher at 0.2 GHz than at 1.0 GHz (Fig. S3). Additionally, the number of internal hydrogen bonds within PFC in these cases also indicated that the frequency of 0.2 GHz is optimal (Fig. S4). Therefore, we ultimately decided to employ the oscillating EF at this specific frequency. This finding challenges the assumption that the destructive impact of EF should be stronger for the amyloid fibrils when the frequency is high22,47. To ensure robustness and avoid inappropriate conclusions, we replicated all our simulations for each case.Electrostatic properties of 2MVXBefore conducting the MD simulations, we performed an in-depth analysis of the charge distribution and dipole moment orientations within the PFC using the PyMol program. As illustrated in Fig. 1a and b, our analysis reveals a notably higher electron charge density on the N-terminal domain of the one subunit, when compared to the C-terminal. This observation is further validated by the dipole moment calculations, assigning Kollman charges48 to the atoms (Fig. 1b,c).Fig. 1Initial electrostatics and dipole moment calculations of PFC; electron density maps (a) and total dipole moment (b) of the one subunit. The total dipole moment of 2MVX (c).Remarkably, the two subunits within 2MVX are interconnected very precisely in the terms of charge distribution. The negative pole (N-ter) of one subunit precisely aligned with the positive pole of the second subunit (C-ter). This arrangement resulted in a highly symmetrical charge distribution within the PFC (Fig. 1c). Unlike individual subunits, the PFC exhibited near nonpolarity due to this balanced charge distribution. These findings suggest that the specific charge arrangement contributes to 2MVX’s stability. It imparts a degree of resistance to external EF or other environmental perturbations. In summary, understanding the electrostatic properties of 2MVX enhanced our knowledge of its stability and potential responses to the external factors.In this study, we also analyzed the dipole moments and electrostatic properties of various amyloid fibrils and compared them with globular proteins. Globular proteins generally exhibit more uniform charge distributions, while fibrillar proteins show distinct charge patterns due to their anisotropic structures. This difference in electrostatic properties is crucial for understanding their respective roles in biological systems. For a detailed comparison of these properties, refer to the SI (Fig. S7).Impact of static and oscillating EFs to the structure of PFCThe molecular shape of proteins plays a crucial role in determining their functional properties, with various functional attributes of peptides relying on these specific conformations. β-sheet structures serve as the fundamental building blocks of amyloid fibrils49. Proteins rich in β-sheet motifs are prone to aggregation and stacking, forming long-lasting, insoluble fibrils—a hallmark of amyloid plaques. These β-sheet structures within amyloid plaques are remarkably stable, resisting breakdown by enzymes in the cell, such as proteases50,51. Because the β-sheets in amyloid plaques are often densely packed (as in Fig. 1c), it might be difficult for proteases to access the peptide bonds52. In addition, the insolubility of amyloid fibrils can cause an obstacle, i.e., the dense packing of the β-sheets can also limit the access of water molecules, which are necessary for most proteases to function effectively. That’s why the β-sheet structures are major contributors to the persistence of amyloid deposits in the brain.To find out how much β-sheet have changed into non-β structures, a visual examination of the obtained trajectories was conducted first to identify any conformational alterations (Fig. 2), then we conducted secondary structure analysis through the obtained trajectories using DSSP tool53. DSSP tool helps to calculate the secondary structures along the whole simulation trajectories and gives percentage predictions for each type of secondary. A diagram of the results averaged over four replicates is presented in Fig. 3.The findings indicate a noticeable reduction in the β-sheet content under both EF conditions. However, the nature of the field appears to influence the denaturation of the Aβ plaque differently31. While the intensities are similar, sharp reductions in β-sheet structures were observed when applying an electric field at a frequency of 0.20 GHz. Furthermore, at an intensity of 0.40 V/nm, the protein exhibited a structural transformation into non-β conformations characterized by coils, bends, and turns, lacking a distinct functional conformation19. In line with findings from previous studies, the emergence of α-helical structures is also evident in our observations24.Fig. 2The final snapshots out of 500 ns simulations are shown under ambient conditions in a green frame, with static EFs in blue frames, and with oscillating EFs in red frames. The final snapshot from the trajectory extended to 1500 ns at 0.40 V/nm static EF is shown in the dashed blue frame. The frequency of 0.2 GHz was used for all oscillating EFs.After 500 ns of simulation, the static EF separates the pentamers that make up the PFC, but the connections between the individual peptides remain (cf. Figs. 2 and 3). Therefore, reduction in β-sheet conformation are not significant. The similar output data was repeated in all other replica simulations. As mentioned in Kalita et al.,22 the presence of a static field leads to a prevalence of the parallel β-sheet structure, favoring it over the antiparallel β-sheet configuration. Obviously, this phenomenon can be rationalized by the tendency of individual peptides to align themselves parallel to the field direction and adopt conformations characterized by higher dipole moments.In alignment with this observation, our investigation reveals a similar peptide orientation response to the external static EF. Initially, the subunits’ dipole moments are turned in an opposing orientation, corresponding to a state of minimal dipole moment. However, when subjected to the influence of the EF the interconnected subunits undergo spatial separation (rapture). This spatial change leads to a shift in the dipole moments, resulting in a parallel orientation each other, i.e. their positive pole (N-ter) and negative pole (C-ter) become on the same side (Fig. 2, blue dashed borders). Considering the observations depicted in Fig. 1, the final structures obtained from static EF simulations (Fig. 2, blue dashed borders) would exhibit exceedingly high electrostatic energies and dipole moments. Consequently, these structures would be incapable of existing under ambient conditions. However, the scenario differs when subjected to astatic EF: the peptides themselves tend to attain higher dipole moment states and can occupy metastable states characterized by specific potential depths under the influence of the field. These metastable states, in turn, can prevent (trap) the peptides from further degradation.Fig. 3Secondary structure analysis diagram of PFC calculated by DSSP tool. The diagram shows all the results of static EF, oscillating EF and ambient condition. Each given percentage is the average of the results of four replicate simulations.To explore this possibility, we extended the duration of simulations at 0.40 V/nm intensity up to 1.5 µs. As illustrated in Fig. 3, even after 1 µs, the β-sheets are still remained. Thus, we can conclude that the structure observed at 500 ns has not undergone significant alterations even in longer simulation time (see Fig. 2).In the presence of oscillating EF, the scenario is completely different. In this case, no conformations trapped to metastable states under the field are formed, and therefore, PFC is completely destroyed even into individual peptides in a short period of time.Changes in dipole moment and dielectric propertiesTo gain insights into the mechanisms by which static and oscillating EFs disrupt the PFC, we investigated their effects at an intensity of 0.40 V/nm. This intensity was chosen because it induced the highest disruptive effects on the PFC as observed in previous results. At the 0.20 GHz frequency, a single cycle of the electric field oscillation takes 5 ns. When exposed to the initial half-cycle (2.5 ns) of the oscillating EF, charged amino acids within the peptides experience propulsive force induced by EF.This force tends to align their charges with the applied field (polarization), leading to a temporary increase in the overall dipole moment of the PFC. During the subsequent half-cycle (another 2.5 ns), the direction of the EF reverses. This reversal EF forces the previously aligned charged amino acids to reorient, resulting in a significant perturbation of their positions (shaking). This situation is confirmed by the results obtained by calculating the dipole moment (Fig. 4a). The contrastingly behavior of the dipole moment under the oscillating EF supports this mechanism.Fig. 4Time dependent evolution of the total dipole moment (a) and the dielectric constant (b) with an intensity of 0.40 V/nm.Under the static EF, the dipole moment increases initially and then remains almost unchanged, indicating there’s no more significant change in conformation. In contrast, under the oscillating EF, the dipole moment exhibits significant fluctuations nearly from 500 up to 3500 Debye that coincide with each cycle of the electric field. This means, there must be significant changes in the conformation as many times as the dipole moment changes. In the presence of oscillating EF, those conformational changes are continuously occurring that ultimately leading to the complete destruction of PFC.According to the analysis of the above results, when the external field is applied in the form of oscillating EF, it has a greater effect on the protein. That is, PFC is more susceptible to oscillating EF than to static EF. In order to clarify this, we demonstrated the change of the dielectric constant of the PFC over time under the influence of both static and oscillating EFs (Fig. 4b). The dielectric constant characterizes protein’s response to an applied EF, indicating how easily considering protein can become polarized. A high dielectric constant signifies heightened responsiveness to the EF, suggesting that the protein is more prone to polarization54. As observed in Fig. 4b, the dielectric constant of the PFC remains significantly elevated under the influence of the oscillating EF. This sustained high value suggests a continuous and substantial response of the PFC to the rapidly changing EF direction in an oscillating EF. This ongoing response likely translates to the significant shaking and perturbation of the structure.Disruption internal connections2MVX is a supramolecular assembly that consist of parallel β-sheets. Amyloid-β fibrils are stabilized by a network of intermolecular hydrogen bonds55 (h-bonds) between the backbones of β-strands in adjacent peptide chains.Fig. 5Different effects of static and oscillating EFs on the number of internal hydrogen bonds in PFC.Specifically, the carbonyl oxygen (C=O) of one amino acid links up with the amide hydrogen (N–H) of another amino acid. This network of h-bonds acts like tiny bridges, holding the fiber structure together and preventing it from falling apart. Besides the stability, h-bonds play crucial rule in elongation of existing fibrils. When another amyloid-β peptides encounter a fibril structure, it can form h-bonds with the exposed β-strands on the fibril surface. These interactions, along with the hydrophobic side chains, stimulate the new peptide to align with the existing β-sheet structure, leading to the fibril elongation as more peptides are incorporated56. Therefore, reducing the number of internal h-bonds is the key to both disrupting and preventing the propagation of toxic fibrils.We calculated the average number of internal h-bonds formed in the protein to show how static EF and oscillating EF can disrupt internal h-bonds (Fig. 5). The GROMACS gmx hbond module was used to calculate possible hydrogen bonds between atoms within the protein. This approach effectively captures both intra-peptide and inter-peptide hydrogen bonds contributing to the overall stability of the fibril structure. The number of h-bonds was calculated from the last 50 ns of each trajectory and averaged among all the replicated simulations. Typically, an increase in field intensity correlates with a reduction in the number of internal h-bonds. The presence of static and oscillating EFs affects PFC differently, particularly leading to a sharp decrease in the average number of h-bonds under the influence of an oscillating EF.It is known that h-bonds are mainly formed between the highly polar group of amino acids in a protein. Both static and oscillating EF exert forces on the protein structure due to their interactions with those polar groups (permanent dipoles) within the protein. However, the nature of these forces differs between static EF and oscillating EF. A constant static EF might exert a pulling or pushing force on the charged atoms within the protein, potentially straining the existing hydrogen bonds. However, this strain might not be enough to overcome the overall stability of the network, leading to a limited disruption in h-bonds. Oscillating EF introduces time-varying EF, causing oscillations in the protein’s environment. The interaction of oscillating EF with polar groups (C=O or N–H) can lead to localized movements or vibrations, which result in the conversion of electrical energy into thermal motion. This localized increase of thermal motion can occur in specific regions of the protein where the charged groups are concentrated (Fig. 1a) or where interactions with the field are strongest. This movement can lead to changes in the protein’s conformation (see Fig. 6). In simple terms, the dynamic nature of oscillating EF can lead to effective disruption of h-bonds through repeated perturbations of the protein structure. We also provided a video as SI, illustrating the amyloid fibril being disrupted by the 0.4 V/nm oscillating EF. This video provides a visual representation of the conformational changes and disruptions.Structural bioinformatics analysisTo gain a deeper understanding of the protein’s structural dynamics, we employed root mean square deviation (RMSD) analysis alongside principal component analysis (PCA), which are standard techniques in structural bioinformatics57 for studying protein motion in MD simulations. In this study, we utilized Pairwise RMSD, which offers a more comprehensive view compared to traditional RMSD58. Pairwise RMSD calculates the average difference in the positions of a set of atoms (often the backbone alpha carbon—Cα atoms) between two individual frames within a trajectory. By comparing every frame to every other frames, it generates a matrix that captures even subtle structural changes throughout the simulation. In contrast, traditional RMSD only measures the deviation of each frame from a single reference structure (usually the initial frame). Figure 6 shows the pairwise RMSDs calculated over Cα atoms. In this case, time interval (dt) is 100 ps, and 5000 frames are compared in each diagram. We can see that at low intensities in EF does not induce significant changes in the structure. As is clear that more yellow color appeared at the higher intensities of EF, e.g., 0.35 and 0.40 V/nm. Although the RMSD appears to be slightly higher under the static EF intensity of 0.40 V/nm, we can still see that there are frames whose root mean square deviations are not substantially different from each other. This means that the structural changes in the presence of static EF is not very extensive. In the presence of oscillating EF, the changes are much higher, especially after 400 ns at 0.40 V/nm, it can be seen that all the frames get completely different from each other. In our case, this means the structure undergoes completely disruption.Fig. 6Pairwise RMSD diagrams calculated for the protein’s Cα atoms at different EF intensities. 5000 frames were extracted from the 500 ns trajectory (the time frame dt = 0.10 ns). Each panel displays the RMSD values for 5000 frames. The color intensity from the dark violet to yellow reflects the magnitude of the RMSD.Based on the Pairwise RMSD analysis, oscillating EFs at 0.40 V/nm induce the most significant structural changes in the protein, potentially leading to complete unfolding. In contrast, the static EF at the same intensity appears to have a less pronounced effect, allowing for some degree of structural stability.To complement the insights from RMSD analysis, we performed Principal Component Analysis (PCA) on both the native PFC and the EF effected structures. PCA is a technique that allows us to explore the collective motions of the protein’s atoms, particularly the alpha carbons (Cα) in this case. The PCA results are presented in Fig. 7. At low static EF intensities (i.e., 0.20 and 0.35 V/nm), the PCA plots show minimal change compared to the native state (Fig. 7, green). The data occupy a similar region in the phase space, indicating that these low-intensity fields do not significantly alter the PFC’s structural flexibility or induce substantial conformational changes. In contrast, the PFC structures exposed to oscillating EFs exhibit a markedly different behavior. The data points out in these plots are more scattered, indicating a comparatively larger area occupied in the phase space (cf. static and oscillating results in Fig. 7). This signifies a greater degree of structural fluctuation and enhanced flexibility within the PFC under oscillating electric fields. Particularly noteworthy is the behavior observed under the high intensity of oscillating EF (0.40 V/nm). The data points in this case encompass the most extensive region in the phase space, suggesting the most significant conformational exploration and potential disruption of the PFC structure.Fig. 7The PCA results calculated for static EF (green), oscillating EF (red) and ambient (black) conditions. Each panel represents a “phase space,” which essentially depicts the range of conformations accessible to the PFC under different conditions. The first and second eigenvectors (directions of greatest variance) are plotted against each other, with each data point representing a snapshot (frame) from the MD simulation trajectory.It is worth mentioning that using non-polarizable force field parameters in our simulations may limit the accuracy of the hydration and interaction dynamics observed in our system. Employing polarizable force fields for both water and protein could potentially yield more accurate results by better capturing the electronic polarization effects within the environment. Despite this limitation, our current simulations provide consistent results within the framework utilized. In future studies, incorporating polarizable force fields in applied EF simulations could offer enhanced accuracy and more detailed data, further refining our understanding of the considering system.
