Abstract
Background
α-synuclein (α-syn) deposition in the mid-brain region is one of the hallmark pathologies of Parkinson's disease (PD). The key steps involve the transformation of α-synuclein into a toxic oligomer and insoluble fibrillar aggregates.
Objective
To understand the role of curcumin-glucoside in the prevention of α-syn aggregation, a mechanistic approach.
Methods
In the present study, we synthesized a novel molecule, curcumin-glucoside (Curc-gluc), to improve the water solubility and partition coefficient, making the molecule with high bioavailability. The present study is focused on understanding the α-syn aggregation kinetics in the presence and absence of Curc-gluc, curcumin (Cur), copper (Cu), and iron (Fe) through Thioflavin T analysis, circular dichroism (CD), mathematical approach by self-association kinetics, and molecular docking models.
Results
The results indicated that Curc-gluc potentially prevented synuclein aggregation compared to Curc alone and inhibited Cu and Fe-induced synuclein aggregation. CD studies indicated that Curc-gluc favored α-helix formation. The docking studies indicated that Curc-gluc derivatives interacted with various chains of α-syn fibrils, namely G-chain, A-chain, I-Chain, and E-chain and the Pi-Pi interaction indicate that, Curc-gluc shows the most favorable binding affinity with the α-syn fibrils with 60.7222 kcal/mol of -CDOCKER_ ENERGY and 89.6516 kcal/mol of -CDOCKER INTERACTION_ ENERGY. The Absorption, Distribution, Metabolism, Excretion (ADME) analysis indicated that Curc-mono and di-gluc have the highest solubility, bioavailability, and tissue distribution compared to Curc alone.
Conclusions
The studies indicated that Curc-gluc prevented α-syn aggregation by favoring α-helix, binding to α-syn, and preventing aggregation, and had high bioavailability.
Keywords
Introduction
Parkinson's disease (PD) is the second most prevalent neurodegenerative disorder among brain disorders.1–5 Currently, PD affects 1 to 2 people per 1000, and the prevalence increases with age and affects 1% of the population above 60 years. 5% to 10% of patients have a genetic predisposition.1,6,7 The incidence and prevalence of PD do increase with advancing age; the condition is more common in men than women. 7 The etiology and pathology of PD are complex, but there are also no early biomarkers or precise molecular pathways for drug discovery.8–12 Our current focus is to understand the role of α-synuclein (α-syn) in PD pathology. The α-syn is a presynaptic protein composed of 140 amino acids with a molecular weight of 15kD, which contains three domains: N-terminal (residues 1–87), NAC (residues 61–95), and C-terminal (residues 96–140).12,13 α-syn binds to the phospholipids in the plasma membrane and will be in alpha-helix form.14,15 When the α-syn gets detached from the membrane, it will be misfolded.14–16 This Misfolded form goes from monomer to oligomer to protofibrils to mature aggregate fibrils, and these aggregates promote cell dysfunction. 17 Also, misfolded α-syn translates into (I) nucleus and (II) mitochondria.18–21 In the nucleus, α-syn is shown to bind to chromatin, leading to DNA strand breaks, promoting genomic instability and alterations in gene expression. In mitochondria, α-syn favors dysfunction through four pathways: (a) inhibits electron transport chain, (b) reduces the production of ATP, (c) binds to metals like iron (Fe) and copper (Cu) and creates altered metal homeostasis and, (d) induces oxidative stress and leads to cell dysfunction.18,21 Earlier studies have indicated that the prevention of α-syn aggregation by small molecules is a promising therapeutic intervention; however, success in drug discovery is still not yet apparent.19–24 Curcumin (Curc) is one such molecule that has the potential to become a drug candidate.25–28 However, the challenges associated with Curc include crossing the blood-brain barrier and solubility, and solutions are needed in this direction.29–31 Previously, our group 30 prepared a curcumin glucoside derivative, trying to improve its solubility and bioavailability. We synthesized novel Curcumin-glucoside (Curc-gluc), such as curcumin mono-glucoside, Curcumin di-glucoside, Curcumin mono-glucoside tetraacetate, and Curcumin di-glucoside tetraacetate. The present study focuses on the molecular understanding of small, novel molecules, such as Curc-gluc, in preventing α-syn aggregation through various approaches.
Methods
Materials
α-syn was procured from r-PEP, USA, Curcumin-Glucoside from CFTRI, Mysore in 2006, Thioflavin-T, FeCl3, CuCl2, were from Sigma-Aldrich, USA.
Aggregation kinetics
The volume of 0.3 ml in tris- buffer (pH 7.4) of α-syn was taken in 10 mM glass tubes with micro-stir bars and was agitated at 37°C, with and without metal ions. The fibril formation of the experiment was conducted based on the Thioflavin T assay.30,31 Thioflavin T assay was conducted for the following samples; a) Syn sample alone; b) Syn + Curc sample; c) Syn + Curc-gluc; d) Syn + Cu and Curc; e) Syn + Fe and Curc; f) Syn + Cu and Curc-gluc; g) Syn + Fe and Curc-gluc; h) Syn preformed fibrils; i) Syn preformed fibrils and Curc, and j) Synuclein preformed fibrils and Curc-gluc. α-syn aggregation was studied with/without metal ions in the presence/absence of Curc and Curc-gluc as a function of different time kinetics, and the experiments were repeated five times, and the aggregation data was subjected to statistical analysis. In brief, for every 5 h, a 10-liter aliquot was taken out of the incubated sample, and 100-liter glycine–NaOH buffer with a pH of 8.2, was added. fluorescence spectra for Thio-t, which specifically binds to α-syn, were recorded at an excitation wavelength of 446 nm and emission spectra at 470–650 nm. Thio-T studies followed the procedures based on Bharathi et al..30,31 Measurements were in semi-micro quartz cuvettes with a 1-cm excitation light path using a Fluoromax-3 spectrofluorometer (Horiba Instruments, USA). The studies were conducted at a 10-band pass and 700 V. The light source was a 150 W xenon lamp.
Statistical methods
We analyzed the α-syn aggregation data using two methods: generalized least squares (GLS) and linear mixed-effect models. GLS provides a powerful tool for analyzing repeated measurement data when the dataset is unbalanced. This approach is beneficial in contexts where traditional repeated measures of ANOVA fail due to assumption violations arising from unbalanced data structures. GLS is an extension of ordinary least squares regression, designed to account for heteroscedasticity (non-constant variance) and correlations among the error terms, which can occur in repeated measures data. In this kind of repeated design, measurements taken on the same subject are likely to be correlated, and the variance can change over time or between treatments. Since original concentration data are not normally distributed, a log transformation was applied to achieve normality or linearity of residuals. The models have been fitted using the Restricted Maximum Likelihood method (REML). The compound symmetry correlation structure if all observations within the same ID have been used and if all observations within the same ID have the exact correlation.32,33
Liner mixed effects (LME) models
The mixed effects model is a robust approach that accounts for dependencies and variations within the data, as measurements are taken on the same subjects over time. Mixed effects models incorporate both fixed effects (which describe the population-level effects) and random effects (which account for individual differences among subjects or groups). The models have been fitted with the REML method. The dependent variable in the model is log concentration, and time (Hours), treatment, and their interaction are included as fixed effects, with random intercepts for time.
Effect sizes for the LME models were reported using omega squared.32,33 The following cut-off values were used to determine the effect sizes. Small Effect Size: ω2 around 0.01 Medium Effect Size: ω2 around 0.06 Large Effect Size: ω2 around 0.14
These values are based on benchmarks suggested by Cohen's criteria. 17 However, in the context of mixed models, calculating omega squared can be more complex due to the presence of random effects and the partitioning of variance across different levels of the model. Moreover, specific cutoffs for small, medium, and significant effects might not apply universally across all fields of study or types of outcomes, as the context can considerably influence what constitutes a “large” or “small” effect.
Mathematical model
The α-syn aggregation kinetics is studied using thioflavin-T (Thio-T) as it identifies the structures that have cross-β β-structural features. 34 The Thio-T exhibits an enhancement when it binds to aggregates and displays low fluorescence in the presence of α-syn monomers. The aggregation kinetics of α-syn is in sigmoid form, clearly defining a flat Thio-T with monomers up to 15 h, and sharply increased with oligomer/ protofibril formation, and later reached saturation when mature fibrils are formed. In the presence of Curc, the monomer phase is enhanced up to 25 h, and then a lower sharp oligomer peak and flattened fibril formation are observed. The magnitude of the difference is 50% between α-syn and Curc. In the presence of Curc-gluc, the monomer phase persists for up to 35 h, followed by a small oligomer peak and flattened aggregates, with a peak magnitude difference of around 75%. The above data indicate that the Curc-gluc has insight into the high potential of preventing the aggregation kinetics from monomer to oligomer/ protofibril to mature fibrils. Now, the data on dependent time-dependent self-association kinetics is transformed into a mathematical analysis. The data can be classified into 3 phases: (A) at t = 0, which consists of a lag phase; (B) t = 25, a linear phase with small aggregates; (C) t = constant, a plateau phase indicating fibril formation. The following mathematical approach is adopted to understand the kinetics of α-syn aggregation. The above data is analysed by using the model from our earlier published paper. 31
Circular dichroism (CD) studies
CD studies were done as explained in our studies. 30 In brief, CD far UV-CD studies are done for α-syn monomers in Tris buffer (pH 7.4) with the following combinations a) Syn sample alone; b) Syn + Curc sample; c) Syn + Curc glucoside; d) Syn + Cu and Curc; e) Syn + Fe and Curc; f) Syn + Cu and Curc-gluc; g) Syn + Fe and Curc-gluc; h) Syn preformed fibrils; i) Syn preformed fibrils and Curc, and j) Syn preformed fibrils and Curc-gluc. The CD spectra was recorded at 25°C using JASCO-J 700 spectropolarimeter with 1 nm slit width and 1 nm path length. Each spectrum was done with an average of four repetitions. The CD contributions from Curc, Curc-gluc, and tris buffer were subtracted to get the real spectra. The conformation of spectra was analyzed.35,36
Docking models
The present computational study involves the Curc-gluc derivatives (Figure 1) and α-syn interactions via docking analysis. The work was carried out with the α-syn fibril formed by complete full-length protein - Twister Polymorph (PDB ID- 6CU8) retrieved from the Protein Data Bank (PDB). 37 The details of the fibril structure and the selected Curc–gluc derivatives ligands are shown under the docking section. The fibril and ligands were later subjected to minimization by using energy minimization protocol using input force field CHARMm (Chemistry at Harvard Macromolecular Mechanics 38 a versatile and widely used molecular simulation program which is a General-purpose all-atom force field with comprehensive coverage for proteins, nucleic acids, and general organic molecules in Discovery Studio (version 3.5) software. 39 Further to minimization, a two-step computational study was carried out to study the putative binding mode of Curc–gluc derivatives, ligand pocket analysis, and molecular docking simulations. Initially, by using Discovery Studio 3.5, 39 the binding site analysis was carried out by using the binding site tools to identify binding sites of a receptor, which derive potential binding pocket sites from cavities in the structure of the receptor based on volume criteria. We determined various binding sites on the α-syn fibrils (PDB ID- 6CU8). The 11 binding sites that were derived from cavities in the structure of the receptor are studied. The binding sites found are displayed as a set of points and a transparent red sphere. The 11 pockets in the fibril, while in the fibril core, comprise residues like 50HGVATVAEKTKEQVTNVGGAVVTGVTAV77 and a pocket on the fibril surface surrounded by His50 and Glu57. Therefore, this region on the surface binding pocket, because of the buried binding pocket, is unlikely to act as a catalytic site for α-syn secondary nucleation. Likewise, docking exercises are conducted to critically evaluate and decipher the extent of binding of these ligands, which were retrieved from the PubChem database and designed 40 and used to study the interactions with α-syn and further identify their relative binding strengths by using CDOCKER, a docking algorithm. 37 CDOCKER is a grid-based molecular docking method where the receptor is held rigid while the ligands are allowed to flex during the refinement to specify the ligand placement in the active site using a binding site sphere. The 10 top-scoring Ligands with Highest CDOCKER_ ENERGY and -CDOCKER INTERACTION_ ENERGY were considered to study the binding affinity via Receptor ligand interactions. The molecular details of the docked α-syn - Ligand complexes images shown in figures and their docking interaction energies, interacting amino acid residues, and hydrogen Bond distance of α-syn interacting with the ligand are tabulated.
Structures of Curcumin-glucoside derivatives.
Prediction of physicochemical and ADMET properties
The ADMET predictor® is used for predicting physicochemical properties, where five compounds are analyzed.41,42 Lipinski's rule is used to analyze the compounds that respect the threshold values of these principles, such as a molecular weight of less than 500 g/mol, no more than five donor bonds, no more than 10 acceptor bonds, and a partition coefficient (LogP) no more than 5. If it complies, it will have adequate passive diffusion across cell membranes and effective absorption from the intestine to the blood. If more than 2 do not comply, poor absorption and permeability are expected. This rule is used for the prediction behavior of oral administration of drugs.
ADMET analysis
Software ADMET Predictor® version 10.4 (Simulations Plus Inc., Lancaster, CA, USA) is used as an in-silico simulation tool. All compounds are subjected to an ADMET analysis to gain insight into the biopharmaceutics and pharmacokinetic properties, which includes the prediction of several properties such as absorption, distribution, metabolism, excretion, and toxicity Performed, and its derivatives are simulated with a human oral dose of 10 mg for 36 h.41,43 For each compound, various ADMET properties are assessed: fraction absorbed (Fa%), bioavailability (F%), Human volume of distribution (Vd), and clearance (L/h). These evaluations offer exciting details about the potential behavior and suitability of the compounds for further development.
Results
Aggregation kinetics
In the present study, we analyzed the aggregation of α-syn by Thioflavin assay using the following combinations, a) Synuclein sample alone; b) Syn + Curc sample; c) Syn + Curc-gluc; d) Syn + Cu and Curc; e) Syn + Fe and Curc; f) Syn + Cu and Curc-gluc; g) Syn + Fe and Curc-gluc; h) Syn preformed fibrils; i) Syn preformed fibrils and Curc, and j) Syn preformed fibrils and Curc-gluc. The Syn self-aggregates, and the aggregation curve passes an S-shape with an initial slow aggregation pattern, a steep aggregation pattern, and a plateau stage. The Curc-gluc effectively reduced the aggregation kinetics compared to Curc. Cu effectively enhanced the Syn aggregation over Fe. Curc-gluc effectively reduced the Cu-induced aggregation over Curc. A similar observation was noticed where Curc-gluc effectively reduced the Fe-induced aggregation over Curc (Figures 2–4). We analyzed the above Syn aggregation kinetics data using the mathematical self-analysis pathway and then advanced statistical analysis. Further, our studies on Syn preformed fibrils indicated a Thioflavin T value of 5.9 ± 1.1, Syn preformed fibrils + Curc Thioflavin T value is 3.9 ± 0.8, and Syn preformed fibrils + Curc-gluc Thioflavin T value is 1.8 ± 0.5. The data indicated that Curc-gluc effectively disintegrated the Syn-preformed fibrils over Curc.
Log concentration versus time (hours) α-syn in the presence of Curc and Curc-gluc.
Log concentration versus time (hours) α-syn in the presence and absence of Cu.
Log concentration versus time (hours) α-syn in the presence an absence of Fe.
Mathematical self-analysis pathway of syn aggregation kinetics
We further developed the self-association pathway of α-syn based on state I (soluble monomer), state II (oligomer), state I (soluble monomer), state II (oligomer), and state III (aggregation). In principle, Thio-T binds more to aggregates and less to oligomers, and Thio-T binds more to aggregates and less to oligomers, but it doesn’t bind to monomers. Using this principle, the above graphs have been classified into self-association databases. The self-analysis parameters are taken from the following graphs. Lag Phase Linear Phase Plateau Phase (I). Lag Phase (i) Initial set value of monomer: It is indicated as the analysis monomer set point and undergoes self-analysis. It is denoted as mn1. Here t = 0 (t = time) (ii) End point of monomer: It is indicated as monomer end point and undergoes self-analysis. It is denoted as mnx. Here (mn = Thio-t values and x = time) Linear Phase: A point is indicated as ol1-olx in this phase α-syn is converted into oligomers, and this phase is further divided into a set point for oligomers and an aggregate endpoint. (i) Setpoint for oligomers: It is indicated as an oligomer's set point and undergoes self-analysis. It is denoted as ol1. Here t = 0 (t = time) (ii) End point of the oligomer: It is indicated as the oligomer endpoint and undergoes self-analysis. It is denoted as olx. Here (ol = Thio-T values and x = time) Plateau Phase: It is indicated as ag1-agx in this phase α-syn is converted into fibrils/aggregation and this phase is further divided into ag1 is aggregation set point and agx is aggregation end point. (i) Set point of fibrils: It is indicated as a fibril set point and undergoes self-analysis. It is denoted as ag1. Here t = 0 (t = time) (ii) End point of fibrils: it is indicated as fibrils endpoint and undergo self-analysis. It is denoted as agx. Here (agl = Thio-T values and x = time)
The mathematical analysis of aggregation data indicates that the aggregation of α-syn visualized over time of 0 to 50 h using Thio-T is classified mathematically into 3 phases. (I) a lag phase with a time 0 (II) complete linear phase where the aggregates keep increasing linearly (III) a plateau phase where aggregation parameters complete, keeping t constant. In Figure 2(a), we analyzed the kinetics of α-syn aggregation alone and found a lag time of 18 h, and the linear phase started at 19 h, and the plateau phase is from 65 h. In Figure 2(b)
Statistical analysis
The α-syn aggregation data is analyzed statistically by two models, the GLS and LME models. The First experimental data analysis used α-syn aggregation and prevention of aggregation by Curc and Curc-gluc.
GLS model for first experimental data
GLS model analysis of α-syn aggregation kinetics pattern.
Residual standard error: 0.1850189.
Degrees of Freedom: 53 total, 47 residual.
Linear mixed effect model α-syn aggregation kinetics pattern.
Signif. Codes: 0 ‘ *** ’ 0.001 ‘ ** ’ 0.01 ‘ * ’ 0.05 ‘ . ’ 0.1‘ ’ 1.
Linear Mixed Effect Model α-syn aggregation kinetics pattern.
-one side CIs: upper bound fixed at [1.00].
Linear Mixed Effect Model α-syn aggregation kinetics pattern.
Degrees-of-freedom method: Kenward-Roger.
p value adjustment: Tukey method for comparing a family of 3 estimates.
Generalized least squares fit by REML for α-syn aggregation kinetics.
Model: log_ concentrations ∼ Time_ Hours * Treatment
Data: syn_Ex2_Data
LME model output, ANOVA results and 95% CI variables, and the effect sizes for the first set of experimental data
Linear mixed model fit by REML. T-test use Satterthwaite's method.
Signif. Codes: 0 ‘ *** ' 0.001 ‘ ** ' 0.01 ‘ * ' 0.05 ‘ . ' 0.1 ‘ ' 1.
Analysis of deviance data (Type III Wald chisquare tests).
Response: log_ Concentration.
- One-side CIs: upper bound fixed at [1.00].
Post hoc comparisons for α-syn aggregation kinetics.
Time Hours = 21.7:
Degrees-of-freedom method: Kenward-Roger.
p value adjustment: Tukey method for comparing a family of 6 estimates.
Log concentration vs time (hours).
Generalized least squares fit by REML
Model: log- concentration ∼ time hours * treatment
Data: syn_Ex3_ Data
LME Model for α-syn aggregation kinetics.
Lim = near mixed model fit by REML. T-test use Satterthwaite's method
Formula: Log-concentration ∼ Time Hours * treatment + (1 | Time Hours).
Data: Syn_Ex3_ Data
REML criterion at convergence: 43
Signif. Code: 0 ‘ *** ' 0.001 ‘ ** ' 0.01 ‘ * ' 0.05 ‘ . ' 0.1 ‘ ' 1.
Analysis of derivatives table (Type III Wald chisquare test).
Response: log_ Concentration
- One-side Cis: upper bound fixed at [1.00].
Post hoc comparisons for α-syn aggregation kinetics.
Post-hoc comparisons with Tukey test based on the mixed effect model.
The post-hoc analysis of estimated marginal means (EMMs) from the above linear mixed effects model. It provides pairwise comparisons of treatment levels at the meantime point (Time Hours = 27.5) with a Tukey adjustment for multiple comparisons. Tukey's method controls the family-wise error rate, reducing the chance of type I errors due to multiple testing. The Kenward-Roger method was used to approximate the degrees of freedom for the tests. This method is often used to provide a more accurate inference when the sample size is small, or the design is unbalanced. The confidence level for the confidence intervals around the EMMs is set at 95%.
Post-hoc comparisons
The post hoc analysis with Tukey's adjustment indicates that at the meantime point evaluated (Time_Hours = 27.5), there are statistically significant differences in the response variable between all treatment pairs. α-syn is associated with a higher log concentration than both Syn + Curc and Syn + Curc-gluc, and Syn + Curc is associated with a higher log concentration than Syn + Curc-gluc. The results suggest that treatment effects are significantly different from each other at the current point in the study.
Estimated marginal means
Treatments: i) Syn: The EMM of the response variable at Time Hours = 27.5 is 3.65, with a standard error (SE) of 0.045, ii) Syn_Curc: The EMM at the same time point is 3.02, with a SE of 0.044 and iii) Syn_Curc-gluc: The EMM is 2.56, with a SE of 0.044.
Pairwise comparisons (contrasts)
Syn versus Syn_Curc: The difference in EMMs between Syn and Syn + Curc is 0.628, with an SE of 0.050. This difference is statistically significant (p < 0.0001), indicating that at Time Hours = 27.5, Synuclein has a significantly higher response than Syn + Curc.
Syn versus Syn_Curc-gluc: The difference in EMMs between Synuclein and Syn_Curc-gluc is 1.084, with a SE of 0.050. This is also statistically significant (p < 0.0001), suggesting a significantly higher response for Synuclein compared to Synuclein + Curc-gluc at the same time point.
Syn_Curc versus Syn_Curc-gluc: The difference in EMMs between Syn_Curc and Syn_Curc-gluc is 0.456, with a SE of 0.049. This difference is statistically significant (p < 0.0001), indicating that Syn + Curc has a significantly higher response than Syn + Curc-gluc at Time Hours = 27.5.
A similar interpretation needed to be applied for the second and third post-hoc experimental data results, depending on the p-values and the estimated coefficient values.
Second experimental data analysis: GLS model
The data showcases the output of the GLS model for the second experimental data set. From that, we can see the following effects, i) The treatment of treatment Syn_Cu is associated with a significant increase in log concentration (by 0.196 + 0.021 = 0.217) units per each 1-h increase in the time, compared to the reference treatment “Syn Alone”, ii)In contrast to that, the treatment of treatment Syn_Curc-gluc is associated with a significant decrease in log concentration (by −0.663–0.015=-0.678) units per each 1-h increase in the time, compared to the reference treatment “Syn Alone”, iii)The treatment Syn_Curc is also associated with a significant decrease in log concentration (by −0.204–0.016=-0.220) units for each 1-h increment, compared to the reference group “Syn Alone and iv). The other treatment effects (Syn_Cu_Curc and Syn_Cu_Curc-gluc) do not significantly change over time.
LME model output, ANOVA results and 95% CI variables and the effect sizes for second experimental data
The results represent the outcome for the LME model, ANOVA table, 95% CI for variables and the effect sizes using omega squared values for the second set of experimental data. According to the ANOVA table, we can see that the overall effect of time, treatment, and interaction effects are all significant with p-values <0.000. For the reference treatment (“Syn Alone”), log concentration is expected to increase by 0.0293 units with each time unit. For treatment Syn + Cu + Curc- gluc at time 0, log concentration is expected to be 0.338 units lower than the reference Syn only. For treatment Syn + Cu at time 0, log concentration is expected to be 0.279 units higher than the reference treatment “Syn Alone. The treatment Syn_Cu is associated with a significant increase in log concentration (by 0.279 + 0.015 = 0.294) units per each 1-h increase in the time, compared to the reference treatment “Syn Alone”. In contrast to that, the treatment Syn_Curc-gluc is associated with a significant decrease in log concentration (by −0.663–0.015=-0.678) units per each 1-h increase in the time, compared to the reference treatment “Syn Alone”. The treatment Syn_Curc is also associated with a significant decrease in log concentration (by −0.204–0.015=-0.219) units for each 1-h increment in the time, compared to the reference group “Syn Alone”. The other treatment effects (Syn_Cu_Curc and Syn_Cu_Curc-gluc) do not significantly change over time.
Post hoc comparisons
Post-hoc comparisons indicate the potential role of Curc-gluc in the prevention of α-syn aggregation.
GLS model for second experimental data
The result shows the output for the GLS model for the third experimental data set. From the data, we can see the following effects, i)The treatment of Syn with Fe is associated with a significant increase in log concentration (by 0.119 + 0.016 = 0.135) units per each 1-h increase in the time, compared to the reference treatment “Syn Alone”, ii) In contrast, the treatment of Syn_with Curc-gluc is associated with a significant decrease in log concentration (by −0.663–0.015=-0.678) units per each 1-h increase in the time, compared to the reference treatment Syn only, and iii) The treatment of Syn with Curc is also associated with a significant decrease in log concentration (by −0.204–0.015=-0.219) units for each 1-h increment in the time, compared to the reference group α-syn Only.
LME model for third experimental data
The data represents the outcome for the LME model, ANOVA table, 95% CI for variables, and the effect sizes using omega squared values for the third set of experimental data. According to the ANOVA table, we can see that the overall effect of time, treatment, and interaction effects are all significant with p-values <0.0001. The treatment of Syn with Fe and Curc is associated with a significant increase in log concentration (by 0.079–0.013 = 0.066) units per each 1-h increase in the time, compared to the reference treatment “Syn alone”. In contrast, the treatment of Syn_Curc-gluc is associated with a significant decrease in log concentration (by −0.664–0.015=-0.679) units per each 1-h increase in time, compared to the reference treatment Syn only. The Treatment of Syn_Curc is also associated with a significant decrease in log concentration (by −0.204–0.015=-0.219) units for each 1-h increment in the time, compared to the reference group Syn Only. The Post-hoc Comparisons indicate the potential performance of Curc-gluc in preventing α-syn aggregation.
Circular dichroism studies
CD studies of α-syn were conducted at 192–260 nm wavelengths. These spectra indicated a robust negative peak at 195–200 nm, indicating random coil conformation. The data analysis (Table 13) indicated 5% of α-helix, 30% of β-sheet, and 60% of random coil for α-syn alone. When Curc is added to syn and incubated for 48 h, the spectral data analysis data indicated 35% of α-helix, 25% of β-sheet, and 40% of the random coil. But when Curc-gluc was added to the α-syn, the spectral data analysis indicated 65% of α-helix, 15% of β-sheet, and 20% of a random coil. We analyzed spectral data in the presence and absence of metals like Cu and Fe. These metals are added to Syn and incubated for 48 h. The Syn-Cu complex data analysis alone is indicated as 2% of α-helix, 60% of β- β-sheet, and 38% of a random coil. When Curc is added to syn-Cu complex, the secondary conformational data analysis indicated α-helix of 36%, β-sheet as 20%, and a random coil as 44%. When Curc-gluc is added to the Syn-Cu complex, the data analysis indicated an α-helix of 55%, β-sheet as 15%, and a random coil as 30%. The data analysis of the Syn-Fe complex alone indicates- α-helix of 3%, β-sheet as 50%, and random coil as 47%. When Curc is added, the data analysis indicated anis the Syn-Fe complex, the data analysis indicated an α-helix of 30%, β-sheet as 20%, and the random coil as 50%. When Curc-gluc is added to the Syn-Fe complex, the data analysis is indicated as α-helix of 65%, β-sheet as 10%, and random coil is 25%. The preformed fibrils are treated with Curc-gluc and the secondary formational data indicated the following, i) preformed fibrils have 60% β-sheet, random coil 30%, and unstructured feature 10%, ii) Curc-gluc brought changes in secondary conformation in preformed fibrils, 20% β-sheet, 60% α-helix, and 20% random coil, and iii) Curc alone brought the following secondary conformation changes in Syn formed fibrils- 40% β-sheet, 25% α-helix, and 35% random coil. The data indicates that Curc-gluc could be able to disintegrate preformed fibrils and favor conformational change from β-sheet to α- α-helix.
Circular dichroism studies of α-syn.
Docking studies
In silico docking interaction of α-syn fibrils with Curc–gluc derivatives-binding affinity study (Figures 5 and 6). The ten ten-scoring Ligands with the Highest CDOCKER_ ENERGY and -CDOCKER INTERACTION_ ENERGY were considered to study the binding affinity via receptor-ligand interactions. The molecular details of the docked α-syn-Ligand complexes shown in the figures and their docking interaction energies, interacting amino acid residues, and hydrogen bond distance of α-syn interacting with the ligand are tabulated in Tables 14 and 15.
(a) Alpha-synuclein fibril (PDB ID- 6CU8) formed by full length protein - Twister Polymorph (b) showing the Identified 11 binding sites of a receptor derived from cavities in the structure of the receptor of alpha-synuclein fibril (PDB ID- 6CU8).
(a) α-syn-Curcumin Mono-Glucoside Tetraacetate binding pattern. (b) α-syn interaction pattern with Curcumin Di-Glucoside Tetraacetate. (c) α-syn interaction pattern with Curcumin Mono-Glucoside. (d) α-syn interaction pattern with Curcumin Di- Glucoside.
Binding affinity of the selected Curcumin–Glucoside derivatives ligands with the α-synuclein (syn) using PDB ID: 6CU8. With binding energies -CDOCKER_ ENERGY and–CDOCKER INTERACTION_ ENERGY indicating the strength of interactions.
Summary of interactions of four Curcumin–Glucoside derivatives with α-synuclein Fibrils.
To further understand the binding mechanisms of the potential Curc–gluc derivatives to α-syn by using in silico docking techniques and to investigate the detailed interactions and underlying disruptive mechanisms of Curc–gluc derivatives molecules with α-syn 43 to 83α Twister Polymorph (PDB ID- 6CU8) that confer affinity and specificity. Our present study focused on Curc–gluc derivatives like Curcumin Mono-Glucoside Tetra Acetate, Curcumin Di-Glucoside tetra acetate, Curcumin Mono-Glucoside and Curcumin Di-Glucoside with α-syn Interactions via docking analysis. In molecular docking simulation studies, the ligand gets bound to the site's protein's binding pocket in different sites, thus providing valuable information on the binding site and mode interactions. The selected Curcumin–Glucoside derivatives and their molecular structure and details are shown in Figure 5 and Table 14. The Curcumin–Glucoside interactions with the amino acids at the binding site are shown in Table 14. The Curcumin–Glucoside interactions with the amino acids at the binding site are shown in Table 14. As we are concerned mostly with parallel β-sheet amyloids rather than α-syn itself, the α-syn43–83 regions (consisting of 40 residues) are chosen as our model. Our docking studies demonstrate that Curcumin–Glucoside derivatives molecules can probably destabilize α-syn prefibrillar tetramer by disrupting the β-sheet structure and destroying the intra- and inter-peptide E46-K80 salt bridges. All the Curcumin–Glucoside derivatives docked interacted with various chains of α-syn fibrils, namely G-chain, E-chain A-chain, C-Chain and I-Chain. The Summary of interactions of ligands with α-syn with amino acid from 55 to 79 are shown in Table 15. The interaction data indicated that Curcumin Di-Glucoside tetraacetate shows the most favorable binding affinity with the α-syn Fibrils with 60.7222 kcal/mol of -CDOCKER_ ENERGY and 89.6516 kcal/mol of -CDOCKER INTERACTION_ ENERGY and could form strong hydrogen bond interaction with the core functional amino acids such as GLU61, THR72, GLY73, GLU57, THR59 and VAL74. With A, C, G, E, I chain. The compounds also exhibit the Pi–Pi interaction. Similarly, with Curcumin Di-Glucoside binding energy of 10.714 kcal/mol of -CDOCKER_ ENERGY, 74.6959 kcal/mol of -CDOCKER INTERACTION_ ENERGY exhibiting the interactions with A, C, E, G, I Chains.
Curcumin Mono-Glucoside Tetra Acetate with 31.9804 kcal/mol of -CDOCKER_ ENERGY 73.9655 kcal/mol of -CDOCKER INTERACTION_ ENERGY with G, E, A chains and Curcumin Mono-Glucoside 10.1967kcal/mol of -CDOCKER_ ENERGY 62.1933kcal/mol of -CDOCKER INTERACTION_ ENERGY with G, A, E chains of the Fibril. A greater number of hydrogen bond interactions were found in the binding Curcumin Di-Glucoside derivatives with α-syn Fibrils residues. The binding energy comparative Bar chart is shown in Figure 7. Where Bar chart Showing the rectangular bars with lengths that are proportional to the values representing the docked binding energies of ligands - CDOCKER_ ENERGY and -CDOCKER INTERACTION_ ENERGY of the Curcumin–Glucoside derivatives. Whereas Curcumin Di-Glucoside tetraacetate shows the highest Binding Energy. These molecules were found to have a high binding affinity at three sites of α-syn tetramer. The electrostatic and hydrogen bonding interactions play dominant roles in Curcumin–Glucoside derivatives molecules, probably disrupting α-syn protofibrils. The aromatic π-stacking helps Curcumin–Glucoside derivatives molecules bind to α-syn protofibril interactively.
Bar chart showing the rectangular bars with lengths that are proportional to the values representing the docked binding energies of ligands - CDOCKER_ ENERGY and -CDOCKER INTERACTION_ ENERGY of the Curcumin–Glucoside derivatives. While Curcumin Di-Glucoside tetraacetate showing the Highest Binding Energy.
Thus, these interactions probably demonstrate that Curc–gluc derivatives molecules can reduce the β-sheet content of α-syn tetramer and may convert part of the β-sheet structure into the random coil or bend conformation. Since molecules interact across the β-sheet, probabilities of residues like GLU61 20, THR72, GLY73, GLU57, THR59, VAL74. The detailed interacting residue and all compounds’ corresponding binding energy values are listed in Tables 14 and 15 and Figure 5(a), b and Figure 6(a)-(d). It is known from the previous studies that the NMR study suggested that the fragments T44-V55, E61-G66, V70-A78, T81-E83, and I88-K96 adopt the β-sheet in αS44–−96 protofibril [R] where the innermost β-sheet of the core including residues 71–82 was reported to be necessary for α-syn fibril formation. These regions are involved in the interactions of ligands across β-sheets observed in our docking interactions, which is supposed to go against α-syn fibrillation.
Overall, Curc–gluc derivatives molecules can serve as a β-sheet breaker to disrupt the β-sheet structures of α-syn43–83 protofibrils and can disturb the structural stability of α-syn tetramer. Thus, these molecules can disrupt the β-sheet structure of α-syn tetramer by reducing the β-sheet content of residues. Curcumin–glucoside derivatives molecules can also reduce the inter-chain H-bond number and destroy the E46-K80 salt bridges. The destabilization of α-syn protofibril resulting from Curc–gluc derivatives binding is supposed to prevent the peptide–peptide association and inhibit the subsequent fibrillation. Three binding sites were identified for Curc–gluc derivatives molecules interacting with α-syn tetramer with GLU61 20, THR72, GLY73, GLU57, THR59, and VAL74. The binding of Curc–gluc derivatives molecules to α-syn tetramer is dominantly driven by the electrostatic and H-bonding interactions. Overall, our work provides the molecular details of the disruptive effects of Curc–gluc derivatives molecules on α-syn proto-fibrillar oligomer, which helps develop new treatments (drug design or exercise therapy) against PD.
Prediction of physicochemical and ADMET properties
We predicted the physicochemical properties of Curc and its derivatives. 42 For five compounds, the water solubility was 0.017 to 4.3 mg/mL, brain/blood partition coefficient (Log BB) was less than 0.3, molecular weight lees than 500 g/mol for Curc and more than 500 g/mol for its derivatives, acceptor bonds no more than 10 for curcumin, a partition coefficient (LogP) were between −0.1 to 3 and donor bonds no more than 5 for Curc and curcumin mono-glucoside tetra acetate. The results are shown in Table 16.
Prediction of physicochemical properties of curcumin and its glucoside derivatives analyzed using ADMET Predictor®. Curcumin (1), Curcumin mono-glucoside (2), Curcumin di-glucoside (3), Curcumin mono-glucoside tetra acetate (4) and Curcumin di-glucoside tetra acetate (5).
Prediction of ADMET properties We predicted the ADMET properties 44 of Curc and its derivatives. The results are shown in Table 17. Compound 1's absorbed fraction (Fa%) is like that of compound 4; however, compounds 2, 3, and 5 are low in the absorbed fraction (Fa%). The bioavailable fraction (Fb%) is high for compound 4. All compounds (1–5) are P-glycoprotein substrates. Compounds 1, 4, and 5 are P-glycoprotein inhibitors. All derivatives (2–5) have a lower potential to cross the blood–brain lower potential to cross the blood-brain barrier than Curc. Compounds 4 and 5 could be CYP3A4 inhibitors, CYP1A2 r inhibitors, and CYP1A2 inhibitors. Oral rat acute toxicity (LD 50) and Oral rat chronic toxicity are also predicted, and results are shown in Table 17. According to the AMES test, compounds 2, 4, and 5 are expected not to introduce AMES toxicity. ADMET physicochemical results indicate that curcumin derivatives (Curcumin (1), Curcumin mono-glucoside (2), Curcumin di-glucoside (3), Curcumin mono-glucoside tetra acetate (4) and Curcumin di-glucoside tetra acetate (5)) do not obey the Lipinskís rules because more than two values are not complied, that means these compounds could have low absorption and permeability trough passive diffusion and suggest that mediated active transport is necessary. We predicted the physicochemical properties of curcumin and its derivatives. The results are shown in Table 17. Compound 1's absorbed fraction (Fa%) was like compound 4's; however, compounds 2, 3, and 5 were low the absorbed fraction (Fa%). Bioavailable fractions (Fb%) were high for compound 4. All compounds (1–5) were P-glycoprotein substrates. Compounds 1, 4, and 5 were P-glycoprotein inhibitors. All derivatives (2–5) have a lower potential to cross the blood–brain barrier than curcumin. Additionally, compounds 4 and 5 could be CYP3A4 inhibitors and CYP1A2 inhibitors. Oral rat acute toxicity (LD 50) and Oral rat chronic toxicity were also predicted, and results are shown in Table 17. According to the AMES test, compounds 2, 4, and 5 were expected not to introduce AMES toxicity. The prediction of pharmacokinetic profiles after simulation of oral administration of all compounds (1–5) presented a gradual decrease in plasma concentration (Figure 8). Compounds 2 and 3 presented lower volume of distribution and shorter half-life than compounds (1, 4, 5). Additionally, compound 2 presented lower tmax, higher Cmax, and AUC (Table 18). ADMET properties indicate that curcumin mono-glucoside tetra acetate showed a higher bioavailable fraction (Fb%). Some compounds could obstruct the efflux transporters that allow medications to be pumped out of cells and could be able to detect and transport them. According to the AMES test, compounds like curcumin mono-glucoside tetra acetate, curcumin mono-glucoside, and Curcumin di-glucoside tetra acetate are expected to not introduce AMES toxicity, indicating that it is unlikely that they will result in cell mutations. Prediction of pharmacokinetic profiles of curcumin derivatives (2–5) presented lower volume of distribution than Curc, which means there is not too much drug in tissues. Additionally, compounds 2 and 3 presented a lower volume of distribution and a shorter half-life than other compounds (1, 4–5). According to the results of physicochemical and ADMET properties,42–44 we can propose our future in vitro and in vivo studies in the following order of priority: curcumin mono-glucoside, curcumin mono-glucoside tetra acetate, curcumin di-glucoside, and curcumin. From a point of view, biopharmaceutical analysis of curcumin mono-glucoside, curcumin mono-glucoside tetra acetate, and curcumin di-glucoside showed better biopharmaceutical properties than Curc.
Simulated human plasma concentration using 10 mg of each compound.
Prediction of ADMET properties of curcumin and its derivatives analyzed by ADMET Predictor® software. Curcumin (1), Curcumin mono-glucoside (2), Curcumin di-glucoside (3), Curcumin mono-glucoside tetra acetate (4) and Curcumin di-glucoside tetra acetate (5).
Predictions of pharmacokinetics parameters of curcumin and its glucoside derivatives analyzed using ADMET Predictor®. Curcumin (1), Curcumin mono-glucoside (2), Curcumin di-glucoside (3), Curcumin mono-glucoside tetra acetate (4) and Curcumin di-glucoside tetra acetate (5).
Discussion
PD is a progressive disorder with degenerative changes in dopaminergic neurons in the substantia nigra region. 45 α-syn aggregation as Lewy bodies is the hallmark pathology in PD neurodegeneration. The α-syn aggregation is the critical step in neuronal cell dysfunction. 46 The pathways involved in α-syn aggregation are many, like monomer to oligomer and oligomer to aggregation, but the still mechanisms of aggregation are not clear. 46 α-syn has self-aggregation propensity, but factors like aluminum Al), Fe, and Cu favor aggregation.47,48 The current therapeutic strategy in PD drug discovery is to inhibit α-syn - aggregation. Natural small molecules have become potential avenues in protecting against neurodegeneration by preventing α-syn aggregation and toxicity.49,50 Curc has become one such potential molecule, but the molecule has less blood-brain-barrier crossing ability, and the reaching of a good concentration of Curc into the brain is not achievable because of its soluble chemistry. To overcome this, earlier, we synthesized Curc-gluc and formed mono-glucose and di-glucose molecules, which have high solubility and good partition coefficient and potential bioavailability. 30 The Curc-gluc is also able to bind to α-syn in oligomer/fibril condition but not to the monomer, as evidenced by our earlier Micro-Cal studies. 30 In the present study, our focus is on Curc-gluc binding to α-syn through mathematical analysis and docking models. Our studies indicated that Curc-gluc is more effective than Curc in preventing α-syn aggregation by many folds. Curc-gluc also effectively prevented Cu and Fe-induced aggregation. The self-analysis data also highlighted the potential role of Curc-gluc in the prevention of self-assembly of α-syn and Cu and Fe induced α-syn aggregation pathways, and in the study, we indicated that Curc-gluc potentially favored α-helix and random coil over beta-sheet, and it is observed that Curc-gluc reduced the metal-induced beta-sheet formation in α-syn.
Further, we explore the understanding of mechanisms of the potential Curc–gluc derivatives to α-syn by using in silico docking techniques and the mechanisms of Curc–gluc derivatives with α Syn 43 to 83α Twister Polymorph (PDB ID- 6CU8) that confer affinity and specificity.51,52 Overall, Curc–gluc derivative molecules can serve as a β-sheet breaker to disrupt the β-sheet structures of α-syn43–83 protofibrils and can disturb the structural stability of α-syn tetramer. Thus, these molecules can disrupt the β-sheet structure of α-syn tetramer by reducing the β-sheet content of residues. Curc–gluc derivatives molecules can also reduce the inter-chain H-bond number and destroy the E46-K80 salt bridges. The destabilization of α-syn protofibril resulting from Curc–gluc derivatives binding is supposed to prevent the peptide–peptide association and inhibit the subsequent fibrillation. Three binding sites were identified for Curc–gluc derivatives molecules interacting with α-syn tetramer with GLU61 20, THR72, GLY73, GLU57, THR59, and VAL74. The binding of Curc–gluc derivatives molecules to α-syn tetramer is dominantly driven by the electrostatic and H-bonding interactions. Overall, our work provides the molecular details of the disruptive effects of Curc–gluc derivatives molecules on α-syn proto-fibrillar oligomer. Further, biopharmaceutical data analysis indicated that curcumin mono-glucoside, curcumin mono-glucoside tetra acetate, and curcumin di-glucoside showed better biopharmaceutical properties than Curc.
Earlier studies on Curc indicated that Curc can bind to α-synuclein in monomeric form and prevent aggregation of α-syn by causing the protein to be more diffusive. 53 Liu et al. 54 reported that phase separation has a potential role in regulating α-syn aggregation and found that Curc interferes with phase separation and prevents aggregation. Maji et al. 55 showed that Curc can bind to preformed oligomers and fibrils and inhibit the toxicity by modulating the hydrophobic surface exposure of α-syn. These findings indicate that Curc or derivatives have the potential to prevent syn pathology. Further, Zhang et al. 56 reported that 4-arylidene curcumin derivatives can inhibit the α-syn aggregation, favor the stabilization of α-syn proteostasis conformation, and also inhibit the β-sheets formation. Also, the compounds can disintegrate the preformed α-syn oligomers and fibrils. The studies by Gupta et al. 57 showed that Curc derivatives like Curc-pyrazole and Curc-isoxazole and their derivatives could be able to prevent α-syn aggregation, fibrillization, destabilization of preformed fibrils, and prevent α-syn induced cell toxicity. Also, Sierks et al. 58 showed that the overexpression of oligomeric α-syn causes cell toxicity through the accumulation of α-syn and enhanced oxidative stress and caspases. Nehru team 59 reported modulatory functions of Curc in inhibiting astrocytic activation. Also, Curc prevented the LPS-induced upregulation in NFκB and proinflammatory cytokines (TNF-α, IL-1β, and IL-1α). Curc also resulted in significant improvement in the glutathione system (GSH, GSSG, and redox ratio) and also prevented iron deposition in neurons. Curc also inhibited α-syn aggregates in the dopaminergic neurons. Taneja et al. 60 reported that Curc prevented glutamine-rich (Q-rich) and non-glutamine-rich (non-Q-rich) amyloid aggregates and found that Curc disrupted the preformed htt72Q-GFP aggregates. Our pharmacokinetics prediction data indicated the high bioavailability of Curc-gluc over Curc alone and also observed a differential distribution pattern. A previous study 61 indicated negligible distribution of the Curc to hepatic tissue or other tissues beyond the gastrointestinal tract. Vareed et al. 62 studied the pharmacokinetics of Curc conjugate metabolites in healthy human subjects and the significant bioavailability of Curcumin conjugate in blood. Also, Sharma et al. 61 highlighted the pharmacodynamics and pharmacokinetics of Curc in health and disease. Further, the recent review63–65 highlighted the bioavailability of nano-curcumin. Curc, being a pleiotropic molecule with lipophilic properties, possesses low bioavailability and is quickly metabolized and thus has limited clinical application. Hence, the application of Curc as nanostructures enhances bioavailability and pharmacokinetics. The metadata analysis indicated that different nanostructures have different pharmacokinetics in plasma, liver, tumor, lung, brain, kidney, and spleen.
Based on our data derived from experimental, modeling and mathematical base provided a novel mechanism for Curc-gluc in preventing Syn pathology (Figure 9). Our hypothesis indicates that Lewy bodies are the key components in PD pathology. Lewy bodies disrupt axonal transport and cell-to-cell communication, leading to neuronal dysfunction in PD. α-syn aggregation is vital in the formation of Lewy bodies. α-syn aggregation kinetics is complex, and the prevention of aggregation is the key factor for therapeutic intervention. α-syn gets detached from the membrane and forms a random coil conformation as a monomer. During the PD pathology development, Cu and Fe levels are elevated in cells, and these metals are attracted to bind to random coil monomers of α-syn at the amino acids Lys, His, Asp, Glu, and Met, leading to the formation of α-syn in beta coil oligomers. Further, Cu and Fe favor the protofibril formation, leading to the formation of matured aggregates and then finally Lewy body formation. Our present and past studies indicated that Curc-gluc prevents the formation of oligomer and subsequent fibril formation, and Curc-gluc can also disintegrate fibrils. Curc-gluc, being both water and lipid soluble with a good partition coefficient, had good bioavailability and hope to be a future drug for PD through modulating α-syn pathology. This molecule has the above unique properties over the parent compound Curc. Our current hypothesis is to identify the biomolecule that can work holistically through α-syn folding integrity, favoring the healthy cell function, thus preventing neurodegeneration. Overall, the in-depth data from our paper showed new information that Curc-gluc derivatives will be helpful small molecules for developing new treatments through novel drug design, which modulates α-syn pathology and supports the treatment options for PD.
α-syn aggregates leading to Lewy bodies is the hall mark pathology in PD. α-syn after detaching from membrane adopts random coil conformation as monomer. PD cells are enriched with high concentrations of Cu and Fe and these metals bind to monomers of α-syn at the His, leading to the formation of α-syn oligomers which further follow the chain reaction to protofibril formation and matured aggregates and finally Lewy body formation. Lewy bodies are synuclein complexed with metals and other biomaterials affecting the cell function. We hypothesize based on our present and previous study that Curc-gluc prevents the formation of oligomer and subsequent fibril formation and very significantly Curc-gluc can disintegrate the fibrils. Curc-gluc being both water and lipid soluble with good partition coefficient had good bioavailability and hope to be a future drug for PD through modulating α-syn.
Footnotes
Acknowledgements
The authors wish to thank Dr P. Srinivas, Former Scientist, CFTRI, Mysore, for providing the sample of curc-gluc.
Author contributions
Lakshmi Sowmya Emani (Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Resources; Software; Supervision; Validation; Visualization; Writing - original draft; Writing - review & editing); Jayanth K Rao (Data curation; Investigation; Methodology; Software; Supervision; Validation; Visualization; Writing - original draft; Writing - review & editing); Jagadeesha Kumar Dasappa (Formal analysis; Investigation; Methodology; Resources; Software; Supervision; Validation; Visualization; Writing - original draft; Writing - review & editing); Marisín Pecchio (Data curation; Formal analysis; Investigation; Methodology; Software; Supervision; Validation; Writing - original draft; Writing - review & editing); Johant Lakey-Beitia (Investigation; Methodology; Software; Supervision; Validation; Writing - original draft; Writing - review & editing); Hansapani Rodriguez (Data curation; Formal analysis; Investigation; Methodology; Software; Supervision; Validation; Writing - original draft; Writing - review & editing); Jessica Cruz-Mora (Investigation; Methodology; Writing - original draft; Writing - review & editing), Priya Narayan (Data curation; Formal analysis; Investigation; Methodology; Writing - original draft; Writing - review & editing); Nikhilesh Anand (Investigation; Methodology; Software; Visualization; Writing - original draft; Writing - review & editing); Govindaraju Mullur (Investigation; Methodology; Supervision; Validation; Writing - original draft; Writing - review & editing); Rajanna Ajumeera (Data curation; Investigation; Methodology; Supervision; Writing - original draft; Writing - review & editing), Bharathi S Gadad (Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Resources; Software; Validation; Writing - original draft; Writing - review & editing); and Jagannatha Rao Kosagisharaf (Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Visualization; Writing - original draft; Writing - review & editing).
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: LSE is thankful to KLEF for its financial support. J.L.-B and M.P would like to acknowledge Secretaría Nacional de Ciencia, Tecnología e Innovación (SENACYT grant numbers FID23-003), and SENACYT Project FID23-003 for support. J.L.-B thanks the National System of Investigators (SNI) for financial support.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data availability statement
All data generated or analyzed during this study are included in this published article and available from the corresponding author's lab.