Introduction

Bacterial biofilm related multi-drug resistance is the leading cause of healthcare-associated infections and an enormous burden to the medical industry1. Bacterial biofilms are further reinforced by functional bacterial amyloids which develop the structural scaffold in these robust biofilms and deliver a resilient survival strategy for the bacteria along with promoting obstinate antibiotic resistance2,3. Interestingly, a level of complexity is added as functional amyloids are also associated with several other non-scaffold roles, including gene regulation, modulating host immune responses and emanating the cross-seeding of disease related amyloids worsening the fate of neurodegenerative diseases4,5.

Staphylococcus aureus is one of the major reasons of healthcare-associated infections and is known for its ability to develop resistance to antibiotics6. Phenol-soluble modulins (PSMs) are small amphipathic peptides and the key virulence factors governing the pathogenic potential of staphylococcal infections. The core family of PSMs released by Staphylococcus aureus are categorized by length; the α-type PSMs are shorter with a composition of ∼20–25 amino acids, whereas the β-type PSMs are longer and contain ∼44 amino acids. Some of the α-type PSMs can also exhibit cytolytic properties and are able to lyse eukaryotic cells7. It is well established that the α-type PSMs form amyloids which majorly provide architectural support and protection to the bacterial cells from the host immune responses by facilitating biofilm formation8,9,10. Additionally, recent studies have confirmed that functional amyloids of PSMs act as catalytic agents promoting the degradation of β-lactam antibiotics. The role of functional amyloids (particularly α-type PSMs) as biofilm strengthening agents and the PSM-amyloid-catalyzed degradation of antibiotics authenticate the possible antibiotic resistance mechanisms associated with the amyloid matrix in bacterial biofilms11.

In this study, we focused on the α-PSM family, primarily PSM α1 which is considered as a major contributor to α-PSM fibrillation and studied the individual ability of PSM α1 to form fibrillar amyloids. PSM α1 strengthens staphylococcal pathogenesis, primarily by facilitating the formation of an extracellular, ordered fibrillar structure with amyloid-like properties12. This fibrillar architecture stabilizes the staphylococcal biofilms making it resilient to antibiotic treatment. Gaining a clear understanding of the factors governing PSM α1 amyloid transformation and molecules targeting the same will lead to the designing of improved strategies to combat biofilm-associated infections and will also pave the way for the development of a novel repertoire of antivirulence therapeutics.

Phytocompounds have shown an astounding potential in treating amyloid disorders by binding to the hydrophobic amyloid core of intrinsically disordered peptides and proteins13,14,15,16. Extant literature has described the antimicrobial and other diverse pharmacological activities of chalcones which can be utilized extensively for the development of novel therapeutic agents17,18. Our study puts forward a novel report where the anti-amyloid activity of the phytocompound, Isobavachalcone (IBC), a naturally occurring chalcone, isolated from the seeds of Psoralea corylifolia has been demonstrated to inhibit the amyloidogenic transitions of PSM α1. IBC has already been advocated as a potent anti-cancer, anti-microbial, anti-inflammatory, antioxidant and neuroprotective agent in previous reports19,20.

Standard spectrofluorimetric techniques and high-end microscopy revealed that PSM α1 forms strong amyloid fibrils as was reported previously by Marinelli et al.12. Treatment of the monomeric form of PSM α1 with IBC showed successful inhibition of the amyloid transformation and revealed interesting insights about the concentration-dependent increase in the inhibitory effect. The inhibitory effect was also confirmed visually comparing TEM images of treated and untreated samples. The heteromolecular association and residue level interactions of PSM α1 with IBC were obtained using molecular docking and molecular dynamics simulations to understand the effect of IBC on PSM aggregates. The collective data suggested the role of IBC in inhibition of the amyloidogenic structural transitions of both PSM α1 monomers and aggregates. These findings may help in precisely understanding the role of active hydroxyl-bearing prenylated chalcones that can ultimately help in the tailored designing of bioactive chalcone derivatives for prophylaxis and therapies targeting the functional amyloids consolidating the staphylococcal biofilms.

Results

Isobavachalcone stably binds and interacts with the amyloid core of PSM α1

The top docking clusters of PSM α1 bound to IBC, demonstrated the engagement of IBC in the PSM α1 structure near the amyloid core of the peptide (Fig. 1). The most favorable binding pose of PSM α1-IBC complex showed a stable binding with a docking score of − 5.9 kcal/mol. The complex is stabilized by hydrogen bonding, alkyl and hydrophobic interactions. The chemical structure of IBC and its heteromolecular associations with PSM α1 have been demonstrated in Fig. 1B and Figure S1. IBC stably interacts with the amino acids encompassing the amyloid core of PSM α1 (highlighted in Fig. 1A), particularly with Ile 7, Ile 8, Lys 9, Val 10, Ile 11, Lys 12 and Ser 13. Earlier studies have reported that this segment of PSMα1-IIKVIK is well-conserved even in the naturally occurring truncations of PSM α1 and also constitutes the core of its amyloid fibrils. This segment is known to independently form steric zippers comprised of the canonical cross-β fibril architecture8. Any molecule that binds and disrupts the association of this spine region of PSM α1, may further inhibit the nucleation/seeding of aggregating monomers, abrogating the amyloid transformation of PSM α1. The interaction of IBC with this amyloid core of PSM α1 indicates a potent amyloid modulatory potential of IBC. The choice of a potent therapeutic agent for the treatment of biofilm-associated infections is a pertinent problem21. Molecules like IBC which may interfere with the assembly of functional amyloids by selectively interacting with their amyloid cores, can inhibit the assembly of these amyloids as mature fibrils in the biofilms, subsequently waning the biofilms and making them susceptible to antibiotic treatment.

Figure 1
figure 1

(A) Sequence of PSM α1 (B) IBC binding near the amyloid core of PSM α1 and interactions with the amyloidogenic residues of PSM α1.

Full size image

Hydrophobicity and aromaticity play an important role in the amyloid aggregation of peptides and proteins22,23. Earlier studies have also reported that any inhibitor molecule that stably occupies the hydrophobic patches of amyloidogenic peptides or proteins helps in inhibiting their amyloid transformation24,25. This led to the subsequent study of putative inhibitory potential of IBC using in vitro amyloid-specific dye-binding assays and the amyloid transformation of PSM α1 monomers in the presence of IBC.

Isobavachalcone alters the amyloid transformation kinetics of PSM α1 and interferes with cross-beta aggregate formation

To study the amyloid transformation of PSM α1 in the presence of increasing concentrations of IBC, ThT kinetics assay was performed. PSM α1 was allowed to fibrillate in a potassium phosphate buffer (pH 7.4) at a concentration of 200 µM with constant agitation with and without the addition of IBC. Using ThT fluorescence, we observed the dose dependent inhibitory effect of the IBC (1:0.25,1:0.5, 1:1) with respect to untreated PSM α1. The ThT kinetics of PSM α1 amyloid fibrillation, in the absence of IBC, trailed a characteristic sigmoidal curve with a distinct lag phase of 28 h, an exponential phase up to 50 h, and reaching saturation after 55 h. In the presence of IBC, the modulatory effect was evident even at the lowest concentration of 50 µM, where the sigmoidal growth kinetics of PSM α1 aggregation was seen to be completely altered showing no significant increase in ThT binding; the effect was more dramatic at higher concentrations (Fig. 2). The kinetics of PSM α1 in presence of IBC at the highest concentration (200 µM) depicted more than 90% decrease in ThT binding and the kinetic parameters could not be determined as the aggregation did not follow a sigmoidal kinetics due to a complete subduing of ThT fluorescence. At the concentration of 100 µM of IBC, a prolonged lag phase of 42 h (Fig. 2, Table ST1) was observed which indicates that IBC interferes with the primary nucleation event of PSM α1 amyloid transformation. IBC successfully delayed and reduced the overall ThT binding even at sub-stoichiometric ratios at the saturation phase of PSM α1 amyloid fibrillation but a more prominent inhibitory effect was observed at equimolar concentration (1:1) which completely subdued the ThT fluorescence. Any probable ThT fluorescence quenching effect of IBC was ruled out after we confirmed no spectral overlaps or subduing of the ThT fluorescence in the presence of IBC in the reaction buffer (Figure S3).

Figure 2
figure 2

(A) ThT kinetics assay depicting the amyloid transformation of PSM α1 aggregating alone (black traces) and in presence of IBC (colored traces). (B) ThT maxima and lag time of PSM α1 amyloid transformation in the presence and absence of IBC (100 µM).

Full size image

It is noteworthy that the altered ThT kinetics in presence of IBC suggests an interference in the primary nucleation event which is evident by the reduction in aggregation rate and enhancement in the lag time with IBC treated PSM α1 samples. These findings indicate that IBC binds effectively to the conformations populated at the early stages of PSM α1 fibrillation and prevents the formation of cross-beta structures, depicted by the lower fluorescence intensity of the ThT binding to PSM α1 samples in presence of IBC.

Isobavachalcone interferes with an early event of PSM α1 amyloid fibrillation stemming morphologically distinct aggregates of lower order

The subdued ThT fluorescence and kinetic attenuation of higher order aggregation in presence of IBC encouraged us to characterize the morphology of PSM α1 aggregates using high-end microscopic techniques. Transmission electron microscopy (TEM) images of end-stage aggregates (derived at 55 h) depicted mature inter-twined, distinct amyloid fibrils in control PSM α1 samples, aggregating in the absence of IBC (Fig. 3A). In contrast, the time-matched samples incubated with highest concentration of IBC (200 µM) revealed mostly amorphous, unstructured aggregates (Figure S2).

Figure 3
figure 3

TEM micrographs showing the end stage aggregates of (A) PSM α1 aggregating alone. (B) PSM α1 aggregates formed in the presence of lowest concentration of IBC (50 μM IBC).

Full size image

Interestingly, even at the lowest concentration of IBC, the fibrillar load significantly reduced. A striking reduction in the formation of mature amyloid fibrils was seen in the IBC-treated samples and only some short, disrupted and feeble fiber-like structures were seen (Fig. 3B). This indicates that the prenylated chalcone, IBC is a potent inhibitor of PSM α1 amyloid transformation and effectively directs the formation of off pathway aggregates of lower order. Previous studies have suggested that the formation of such off pathway aggregates completely revoke the amyloid fibrillation of aggregating monomers26. The inhibition of PSM α1 amyloid transformation by IBC may help in the significant reduction of S. aureus biofilm formation which are strengthened by PSM amyloids. These findings motivated us to decipher the effect of IBC on actual modulation of particle size distribution of aggregates formed during PSM α1 amyloid transformation.

Dynamic light scattering (DLS) was employed to assess the size distribution of PSM α1 aggregates formed in the presence and absence of IBC. The samples of PSM α1 aggregating alone, revealed a progressive increase in particle size with distinct populations calculated by DLS at the start (0 h) and end-stage or saturation phase (55 h). The PSM α1 samples observed at 0 h have a diameter within the range of 2 to 4 nm (Fig. 4A) whereas the PSM α1 aggregates observed during the saturation phase at 55 h exhibited notably larger diameter values surpassing 9000 nm (Fig. 4B), demonstrating the presence of higher-order aggregates corroborating the formation of mature amyloid fibrils. Conversely, the PSM α1 samples incubated with IBC displayed a remarkable reduction in the formation of higher order aggregates. The samples treated with IBC displayed a formation of lower order aggregates with a diameter ranging from 60 to 114 nm (Fig. 4C). Evidently, the gradual emergence of higher-order particle sizes in the absence of IBC, confirms the on-pathway nature of PSM α1 amyloid transformation, a phenomenon likely revoked in the presence of IBC. Furthermore, these findings validate the microscopic observations of varied PSM α1 aggregates obtained in case of IBC treated samples. This interesting attenuation of higher order aggregate formation motivated us to monitor the α-helix to β-sheet transition of PSM α1 in the presence of IBC.

Figure 4
figure 4

Dynamic light Scattering (DLS) data illustrating the particle size distribution in PSM aggregates formed with and without IBC. (A) Diameter of the PSM α1 species before incubating the samples for aggregation (initial 0 h). (B,C) depict particle size variations without IBC and after incubation with equimolar concentration of IBC respectively (end-stage aggregates at 55 h). Secondary structure analysis of the end-stage aggregates by CD. (D) PSM α1 shows a predominantly α-helical signature at 0 h (black traces), after amyloid transformation there is an increase in the β-sheet content corroborated by the characteristic β-sheet CD signature in the absence of IBC. (E) In the presence of even the lowest concentration of IBC (50 µM), the CD signature indicates the loss of β-sheet architecture but retention of a residual helical content.

Full size image

Isobavachalcone inhibits the amyloidogenic structural transitions of PSM α1

PSM α1 monomers prior to aggregation (at 0 h of ThT kinetics) show a predominant α-helical structure, but after amyloid transformation there is an increase in the β-sheet content corroborated by the characteristic β-sheet CD signature in the absence of IBC (Fig. 4D). Interestingly, even at the lowest concentration of IBC (50 µM), the CD signature indicates the loss of β-sheet rich structure along with a residual helical content (as depicted by the CD signature in Fig. 4E). The percentage secondary structure content of PSM aggregates obtained at the end-stage of amyloid fibrillation depicts a prominent increase in the β-sheet content but in the presence of even the lowest concentration of IBC a comparatively lower β-sheet content and higher random coil content was observed in the aggregates obtained at the saturation phase of the kinetics (Figure S5). This inhibition of the conformational conversion of α-helical native structure to β-sheet rich amyloid aggregates of PSM α1, confirms the amyloid inhibitory potential of the prenylated chalcone, IBC. This inhibitory action of IBC on the conformational conversion of PSM α1 would further help in weakening the staphylococcal biofilms by inhibiting the assembly of PSM α1 monomers as cross-beta sheet rich functional amyloids. These findings clearly suggest the anti-amyloid and subsequent anti-biofilm ability of IBC. This modulatory effect of IBC on PSM α1 monomers undergoing amyloid transformation motivated us to study if the pre-formed PSM α1 aggregates can be remodeled by IBC.

Modelling the PSM aggregate and choice of potential ligands

The understanding of the aggregation of PSM α1 in the presence of isobavachalcone required us to use an aggregate of PSM α1. The lack of a 3D structure for PSM amyloid aggregates prompted us to model the PSM aggregate in order to find out about the effect of IBC binding. A trimeric PSM aggregate was modeled using InterEvdock. The top 10 consensus-based InterEvDock poses were generated. InterEvScores employs co-evolutionary information to rank the conformers. Essentially, the top ranked conformer was utilized for the study (Fig. 5). It was observed that the amino acids involved in amyloid formation as ranked by InterEvdock is similar to the consensus predicted from AmylPred2 (PSM sequence-IIKVIK).

Figure 5
figure 5

Cartoon view of modeled PSM-A trimer structure. Each chain is depicted in different colors (A) Side view (B) Top view (C) Secondary Structure Information calculated from PDBsum (D) Possible active site cleft observed on the PSM α1 monomeric structure, (E) PSM-IBC Complex.

Full size image

Defibrillation of PSM amyloid

Changes in the conformation of the modeled PSM-A were analyzed using various built-in modules in GROMACS. Deviation in the structural features was monitored using Root Mean Square Deviation (RMSD) for the backbone atoms. Modeled PSM-A (black line) showed a minor fluctuation in the initial period. However, the system was completely equilibrated and stabilized after 55 ns (Fig. 6A). In case of PSM-IBC (red line) the system was in constant fluctuation. The patterns were analyzed at residual level using RMSF graph. It was evident that the initial perturbation in the PSM-A system is due to the terminal N-terminal residue and the unwinding of the alpha-helix to turns. RMSF graphs also indicate a higher deviation between residues 5–12 in chain B and C in PSM-A system. This might me a clear indication of the formation of nucleation point, thereby initiating a turn or twist in the nearby residue over the course of simulation. This could be confirmed by secondary structure analysis and identifying the hotspot residue from the post-MD analysis (discussed in later sections).

Figure 6
figure 6

Top: (A) RMSD Plot of PSM-A compared with PSM-IBC. (B) Number of Hydrogen Bonds observed during the period of simulation. Black line indicates PSM-A and Red line indicates PSM-IBC complex throughout the manuscript. Bottom: (A) Rgyr (B) SASA Plot (C) Secondary Structure Analysis using DSSP. Hotspots of of transition in (top) PSM-A and (bottom) PSM-IBC are marked from R1–R4.

Full size image

Generally, a decrease in the number of hydrogen bonds (HBs) in the system is considered as a clear indication of defibrillation of the amyloid. Particularly, in the PSM-IBC system considered for the study, the number of HBs decreased gradually over the period of simulation (Fig. 6B). Initial increase in the HBs is attributed to interaction between the PSM and ligand considered for inhibiting aggregation. Among the inhibitors, IBC exhibited good inhibitory action as the number of HBs dropped below 30 over the simulation period which is at par with the experimental results.

Other parameter like Rgyr and SASA were also analyzed for the system under consideration for the aggregation profiles. It was observed the compactness of the system kept increasing in PSM-A indicating aggregation whereas, the accessible surface area (SASA) reduced to a value of 45 nm2, much lower in comparison to the SASA values observed in PSM-IBC complex.

Alpha to beta transition aiding amyloid formation

Transition from the alpha-helix to beta-sheet in the system is observed using Dictionary of Secondary Structure of Protein Analysis (DSSP)27. Transition is observed in PSM-A from the initial stages of simulation (Fig. 6C). Apparently, the transition is initiated at the chain C from 5 ns and progresses to the chain B 20 ns onward (R1 & R2) (Fig. 6C). Comparison between Fig. 6C top and bottom reveals a sudden dip in the number of HBs in PSM-A, in contrast to PSM-IBC, and structural transition from alpha-helix to turns can be seen around 10 ns. However, the structure is stabilized with the formation of β-bridges later on, indicated by the constant HBs all over the simulation. Complete transition from α to β-bridge occurs at 15 ns and 20 ns in chains C and B respectively. Chain A unravels into turns and there is no structural transition observed thereafter. At the same time, transition of α-helix into turns and bends are significant in the PSM-IBC system. Loss of α-helix is also confirmed with the unraveling of the structure and steady decrease in the number of HBs over the simulation period [R3-R4]. Newly formed β-strands (PSM-A: Chain B) tend to be perturbing but become stable and mature over the course of simulation. A detailed figure indicating the transition of α to β over the time period of 100 ns in the PSM-IBC system is depicted in Fig. 7A.

Figure 7
figure 7

(A) MD Snapshots of PSM-IBC over the course of simulation. (B) Principal component analysis of PSM-A (left) and PSM-IBC (right).

Full size image

Principal component analysis

Defining the compactness of the protein–ligand complex based on the associated eigen vectors and the subspace occupied is a crucial part of the MD simulation in determining the state of the system considered. The PCA plot depicts a clear observation of successful aggregation in the modeled PSM-A system. On the other hand, PSM-IBC occupies a much more prominent conformational space (Fig. 7B).

Post MD-binding energy calculation

Post MD-Binding Energy calculation was performed using MM-PBSA approach as discussed in the methods. Only dominating conformation over > 70% was considered for the calculation. The various parameters mined from the MM-PBSA calculation of PSM-IBC complex is listed in the Table 1. In PSM-IBC other energy parameters are quite dominating, indicating higher affinity of binding. The higher binding was mainly due to the vdW energy and the Polar Solvation Energy. The distribution of various interaction energy parameters is illustrated in Figure S6.

Table 1 MM-PBSA free energy parameters calculated from dominant cluster in the 100 ns trajectory. All energy parameters are in kJ/mol.
Full size table

Free Energy Decomposition analysis was also performed with g_mmpbsa package to retrieve the crucial residue contribution in the amyloid patches for PSM-IBC. It was observed that the residues LYS6 (-5.7275 ± 0.0993), LYS9 (-4.0030 ± 0.0729), PHE10 (-10.0392 ± 0.1007), ASP13 (-6.5558 ± 0.1588) and LEU17 (-2.6741 ± 0.0646) was contributing significantly to the binding (all values are in kJ/mol).

Overall, there is similarity in the results obtained from the MD simulation for the aggregates with the experimental results. The decrease in the percentage composition of β structures observed in the CD spectra for the monomer in the presence of the potential amyloid inhibitor IBC, is also vindicated by the DSSP analysis of the secondary structural transitions of the PSM system with IBC. The initial decrease in the number of hydrogen bonds and stabilization in PSM-A confirms the structural rearrangements leading to nucleation. The state of the defibrillation, particularly at the point of nucleation, also shows the deviation in the PSM-IBC system from the RMSF graphs. In the event of a nucleation, the red line of the PSM-IBC system should have been above or at least at par with the PSM-A (black line). In the PSM-IBC system, defibrillation starts with the chain B which cascades a structural change in the chain C which defibrillates at a much faster rate than the chain B. On the other hand, the binding score of − 58.381 between PSM and IBC is a further validation of the efficacy of binding of IBC. On the whole, the experimental results concur with the results of the simulation in the aggregate system to validate the role that IBC might play in modulating the transitions of the PSM α1, which can have implications in the disruption of biofilms of S. aureus.

The examination of cytotoxicity of such potent molecules (like the prenylated chalcone, IBC) against human cells is essential for evaluating their selectivity and safety. The cytotoxicity of IBC has already been tested against epidermal human keratinocytes (HaCaT cell line) as S. aureus infection is associated at the topical level. It has been reported that IBC was not able to reduce the cell viability of HaCaT cells even at high concentrations of 25 µg/mL28, indicative of the fact that IBC is a selective compound with substantially low cytotoxicity making it a potential therapeutic candidate. Moreover, IBC shows a wide range of pharmacological activities but due to its low bioavailability and hepatotoxicity, there can be more drug interactions20. Therefore, insights from more clinical researches are obligatory for its appropriate clinical application. However, it can be firmly stated that IBC serves as a privileged scaffold and can be considered as a promising lead for the tailored designing of effective anti-amyloid or anti-biofilm agents.

Discussion

Phenol soluble modulins are a family of amphipathic virulent peptides, which develop fibrillar amyloids involved in firming the architecture of the robust biofilms of S. aureus, a recurrent cause of biofilm-associated infections, predominantly those developing on indwelling medical devices. PSM α1 plays a major role in the fibrillation of α-PSMs giving rise to extracellular fibrils supporting the biofilm-associated antibiotic resistance. The amyloid fibrillation of PSM α1 is effectively inhibited by IBC which stably binds to the monomers and possibly interferes with the primary nucleation event. We suggest that IBC exerts its amyloid modulatory effect by reducing the alpha to beta transitions of PSM α1 and preserving the alpha-helicity along with the formation of unstructured species. The anti-amyloid activity of IBC is analogous to amyloid inhibitory phytocompounds which similarly help in the burial of hydrophobic patches of amyloidogenic proteins/peptides like PSMs subsequently mitigating the association of aggregating monomers. The antibacterial activity of IBC has been reported in some studies which have also suggested its promising efficacy as a medical device coating agent29. Some studies have explored a combination of IBC and curcumin and proposed that this cocktail when administrated together with gentamicin effectively augments its antimicrobial effect against S. aureus biofilms30. Besides, it will be interesting to study and discover the effect of IBC on other amyloid forming phenol soluble modulins and cytotoxic members of the PSM family which can shed light on the specificity and broad range efficacy of IBC. The potency of IBC in biofilm disruption by targeting amyloid fibrils can be studied more appropriately with the amyloid fibril structure of PSM α1. Unfortunately, the fibril structure of PSM α1 is not available and simulations by envisaging such higher order structures is not practicable. This study further warrants insights on the efficacy of IBC in dissociating the pre-formed fibrils by simulation studies performed with a proper amyloid fibrillar structure. This will help in understanding how effectively IBC can disaggregate the pre-formed fibrillar species of PSM α1 and if it can further disintegrate the biofilms of Staphylococcus aureus.

Systematic evaluations using biophysical assays and high-end microscopy have suggested the potency of IBC in inhibiting the amyloid transformation of PSM α1, emphasizing its importance in serving as a structural scaffold for designing novel anti-biofilm compounds targeting functional amyloids consolidating biofilms. The results of the simulation in the aggregate system, viz. the DSSP analysis, the RMSF deviation, the PCA and the free energy values, are also at par with the biophysical and microscopic visualizations, indicating the efficacy of IBC in modulating the amyloid transition of PSM α1. Newer IBC analogues could serve as a starting avenue for developing better therapeutic agents targeting these key virulence factors of Staphylococcus aureus and other similar antibiotic resistant microbial assemblies. Future researches should focus on corroborative studies supporting the amyloid inhibitory and subsequent anti-biofilm activity of IBC using in vivo models.

Methods

Molecular docking with PSM monomer

The PDB structure of PSM α1 was obtained from the RCSB database (5KHB) and was utilized for docking. The unnecessary atoms in the PDB structure were removed and UCSF chimera (available at http://www.cgl.ucsf.edu/chimera) was used for the processing, manipulation and visualization of the peptide. In this study, the phytochemicals and target peptide were considered as input in the Mcule server and the top scoring leads showing high docking score were selected. Further, for rigid docking, the Autodock tool was utilized. Initially, water molecules and heteroatoms were removed from the structure of PSM α1. Addition of Gasteiger charges and H-bond were carried out prior to docking. Top-scoring leads were selected for aggregation studies, based on their non-toxicity.

Preparation of peptides (PSMs)

The peptide, PSM α1 with more than 95% purity was procured from GenPro Biotech. The peptides were dissolved in DMSO at a concentration of 10 mg/mL. Subsequently, the peptides were diluted in potassium phosphate buffer (10 mM) pH 7.4 up to a final concentration of 200 μM.

Amyloid aggregation experiments

The peptides were taken in the aggregation reaction at a concentration of 200 µM, in potassium phosphate buffer (10 mM, pH 7.4) containing 40 µM ThT with 0.02% (v/v) sodium azide to prevent microbial contamination. The samples were prepared in 96 well plates and the reaction volume was adjusted to 100 µL. The plates were incubated at a temperature of 37 °C and continuously agitated at 240 rpm in Varioskan Flash™ multimode plate reader. Of the two best leads, epigallocatechin gallate (EGCG) was eventually ruled out and IBC was chosen to be the subject of further investigations, based on the emerging reports of EGCG being a pan-assay interference compound31. Readings were taken after every 30 min in the presence and absence of increasing molar concentrations of IBC (1:0.25, 1:0.5 and 1:1 molar ratios). The amyloid transformation kinetics was monitored by ThT fluorescence assay.

Thioflavin T fluorescence assay

The kinetics of amyloid transformation of PSM α1 (in presence and absence of IBC) was carefully monitored by ThT fluorescence assay. Fluorescence measurements were done using a Varioskan Flash multimode plate reader (Thermo Fisher Scientific) and analysis was performed using the SkanIt Software 2.4.5 RE for Varioskan Flash. The excitation and emission wavelengths were kept at 450 nm and 485 nm, respectively and the slit width for excitation and emission were set at 5 nm. Analysis of the ThT fluorescence data was done as described previously26. Briefly, the average fluorescence data obtained from triplicate reaction sets of each molar concentration of IBC was fitted using Eq. 1 (Boltzman-sigmoidal kinetics).

$${\text{y}} = ({\text{A}}_{{2}} - {\text{A}}_{{1}} {/1} + {\text{e}}^{{({\text{t}} - {\text{t}}_{{{1}/{2}}} {\text{)/k}}}} ) + {\text{A}}$$
(1)

where A1 and A2 are the initial and final maximum saturated ThT fluorescence intensities respectively, t1/2 is the time at which half maximum fluorescence is attained and k is the rate constant. The nucleation or lag time was calculated as t1/2-2 k.

All measurements were carried out in triplicates wherein error bars represent standard deviation of each data point.

Transmission electron microscopy (TEM)

The morphology of PSM α1 aggregates (formed with and without the screened inhibitor, IBC) was detected using the JEOL JEM-1400plus transmission electron microscope, operated at an accelerating voltage of 120 kV at ACTREC, Khargar, Mumbai. About 30 μL of PSM α1 samples aggregating alone and with IBC, obtained at the end-stage of ThT aggregation kinetics (after two-fold dilution) were adsorbed onto 300 mesh copper grids precoated with a carbon film. After air drying for 5 min, excess fluid was removed using a filter paper and the samples were negatively stained with a drop of 2% solution of uranyl acetate, then the samples were air dried and observed further.

Dynamic light scattering

Dynamic light scattering (DLS) was employed to determine the particle size distribution within the peptide aggregates developed with and without the identified inhibitor. The size distribution of PSM α1 samples aggregating alone and in presence of IBC were measured using the NanoPlus DLS Particle Size and Zeta Potential Analyzer (from Otsuka Electronics, Japan) operated at the Central Sophisticated Instrumentation Facility, BITS Pilani KK Birla Goa Campus. For each data set, measurements were performed at 25 °C with a total of 3 runs and 70 accumulations conducted for each measurement. Samples collected at different time points (initial 0 h sample and end stage aggregates obtained at 55 h) were centrifuged at 10,000 g for 20 min and then transferred into a glass cuvette before each measurement.

Secondary structure analysis

Secondary structural changes in PSM α1 were observed at initial and end stage of aggregation. Aggregating samples in the presence and absence of IBC were taken out at the end stage and studied for secondary structural transformations from the far UV-CD spectra recorded in a Jasco CD Spectropolarimeter (Japan), at the Biophysics Facility of ACTREC, Khargar, Mumbai. Data collection was conducted at a resolution of 0.5 nm between data points, with the scanning process proceeding at a rate of 50 nm per minute. The data recordings employed a quartz cuvette (2 mm pathlength cuvette) and spectral data were gathered over a wavelength ranging from 190 to 260 nm and the band width was 1 nm.

Modeling of PSM aggregate system

Trying to establish a model for the aggregation, we considered a system containing three monomeric units as amyloid aggregates to investigate the effect of IBC on pre-formed aggregates. We again utilized the starting coordinates from PDB ID 5KHB as a monomeric unit, and later added individual monomeric units to the initial structure using InterEvDock32, which was further equilibrated before modeling. The final PSM-Aggregates (PSM-A) contained three monomeric units and caution was taken that the distance between the monomeric units is not greater than 4 Å.

Physicochemical properties of the top leads

The physicochemical and pharmacokinetic properties of the top leads binding to PSM α1 amyloidogenic patches were analyzed using SWISS-ADME server. In this study, SWISS ADME tool was implemented in initial virtual screening process due to its comprehensive analysis of physicochemical properties, pharmacokinetics, drug-likeness, and medicinal chemistry friendliness. It offers high accuracy in bioavailability and solubility. The toxicity profile was analyzed using the Mcule server, which is a comprehensive database (offering identification of hits) and also provides information about the ADMET properties of the bound ligand. Screened compounds were also subjected to Ghose filter and Muegge Filter.

Molecular docking with PSM aggregates

Modeled PSM-A in PDB format was fed into the Mcule server to screen potential targets again. Mcule provides representational docking results with scores and ADMET information33. To further confirm the interaction between the top-screened ligands, rigid-docking was performed using Autodock-Vina tool34. Initially, water molecules and heteroatoms were removed from the structure of PSM. Addition of Gasteiger charges and H-bond were carried out prior to docking. Grid was generated around the amyloid-forming regions, as reported from previous studies8. The docked structures were analyzed for the lowest energy conformer, Root Mean Square Deviation (RMSD) and hydrogen bond (HB) interaction. Peptide-ligand interaction profile was obtained using Protein Plus server and Protein Ligand interaction profiler [available at https://proteins.plus/ and https://projects.biotec.tu-dresden.de/plip-web/plip respectively].

Molecular dynamics

To unveil the binding efficacy of the selected ligand and to understand the molecular mechanism of ligand-induced dispersal of amyloid fibrils, a molecular dynamics study was performed with GROningen MAchine for Chemical Simulations (GROMACS 2020.4)35. Aggregation behaviour of the PSM was observed on stacks of PSM aggregating alone (PSM-A) and Ligand-bound protein stacks (PSM-IBC complex) using various modules available within the GROMACS package. Topology was generated using OPLS (Optimized Potentials for Liquid Simulations) force field36. Ligand topology was generated using Ligpargen server37. Period of simulation was calculated based on the convergence. System was solvated using TIP3P in a 1 nm dodecahedron box. System was neutralized prior to energy minimization. Both the systems were energy minimized for 50,000 steps to remove steric clashes within the system. NVT and NPT equilibration was performed prior to production run. Initially the production run was setup to 50 ns and was extended to confirm convergence of the system.

Trajectory analysis was performed using the built-in GROMACS modules as discussed in our previous studies14. Trajectory was pre-processed for artifacts and critical analysis elucidating the behavior of aggregation of the PSM was performed. Binding Free energy calculation of selected conformers extracted from MD trajectory was performed by MM-PBSA38. Cluster analysis was performed using gmx-cluster tool39. One representative conformation from dominant cluster over the wide range of trajectories was finally considered for analysis. Per residue decomposition was performed to determine the energetics of hot-spot amino acid contributing in the binding interface. We applied the Principal Component analysis (PCA) to understand the dynamics of the aggregates in greater depth. Prominent structures were generated using gmx-anaeig and Free Energy landscape was generated using gmx-sham and in-house python scripts. Graphical representation of conformers was generated using VMD and UCSF Chimera.