Computational Study of Natural Therapeutic Alternatives against CBX4 Associated with the Development of Presbycusis

Abstract

Presbycusis is a clinical condition related to hearing impairment caused by chronic noise exposure, senescence, or genes that exacerbate the loss of the inner ear. Polycomb Chromobox (CBX) proteins regulate gene expression by targeting the Polycomb repressor complex 1 (PRC1) to histone H3K27me3 sites via their chromodomains, playing a key role in developing presbycusis. This study aims to search for new agents to decrease the progression of the hearing damage pathway CBX4. Initially, a review that identified 17 molecules with reported activity in auditory conditions was done. Molecular dynamics (MD) simulation was performed for the native protein and principal systems employing AMBER20 software. Root mean square deviation (RMSD), RMSF (root mean square fluctuation), solvent accessible surface area (SASA), radius of gyration (RoG) analyses, free energy molecular mechanics generalized born surface area (MMGBSA), and principal components analysis (PCA) calculations were obtained. The main results are that oridonin and curcumin have shown binding energy for the CBX4 protein, with higher affinity than UNC3688. In MDs, the CBX4-oridinin complex showed a more stable profile regarding RMSD and SASA, while the CBX4-curcumin compound evidenced better conformational and energetic stability. In conclusion, oridonin and curcumin could be potential inhibitors for CBX4.

Keywords:
molecular-docking; molecular-dynamics; presbycusis; oridonin; curcumin

Introduction

Presbycusis is a clinical condition with an underestimated prevalence at a general level. Even in countries such as the United States and Europe, which exhibit more significant research and epidemiological development, estimates are poorly characterized. Approximately 25-30% of individuals over 65 may have a hearing impairment, which intensifies in adults over 75; however, incidence reports may increase to 50% or higher, particularly in cases of significant hearing loss.¹ Entities such as the World Health Organization (WHO) have estimated that by 2025, approximately 1.2 billion older adults worldwide may experience a potential deterioration in quality of life due to presbycusis.² Therefore, early diagnosis and searching for comprehensive intervention based on mechanisms such as preventing potential risk factors and identifying sequelae affecting the patient’s condition are entirely relevant.³

In the clinical diagnosis of patients with presbycusis, emphasis has been placed on examinations utilizing audiological techniques that allow quantifying the degree of hearing loss with suspicion of the alteration. Therefore, tonal audiometry is available based on determinations of air and bone conduction in frequencies ranging from 250 to 8,000 Hz, with auditory intensities varying from 0 to 120 dB.⁴ Thus, patients with potential presbycusis impairment will be identified by fluctuating and descending slopes, exhibiting patterns of severe frequencies and anomalies in the high frequencies.⁵ On the other hand, early diagnosis associated with presbycusis is relevant to modifying risk factors that could increase the progression of hearing loss. Multiple factors can contribute to the progression of presbycusis, including acoustic trauma, viral infections, smoking, hypertension, diabetes mellitus, exposure to ototoxic agents (such as antibiotics, chemotherapeutics, and heavy metals), immunological disorders, hormonal factors, and dietary changes (high-fat foods).⁶,⁷

On the other hand, it has been shown that genetic phenomena predispose to age-related hearing loss.⁸ These have been linked to the decrease in high frequencies in patients under 50 years of age with variable genetic etiology, which may be associated with the presence of genetic polymorphisms in genes for oxidative stress regulatory enzymes such as glutathione S-transferase (GSTM1 and GSTT1) and N-acetyltransferase 2 (NAT2*6A). Similarly, the development of pathologies such as multiple familial age-related hearing loss (mARHL) and single/sporadic age-related hearing loss (sARHL), which are linked to variants such as MYO6, MYO7A, PTPRQ, and TECTA, expressed in 8.9% of ARHL cases, is associated with the development of atherosclerosis.⁹ Likewise, the development of afforestation in these populations may be related to atrial presbycusis, characterized by vascular stenosis and decreased irrigation, or mechanical presbycusis with morphological modifications at the level of the cochlear basilar membrane. At younger ages, phenomena masking central counterbalance mechanisms may develop. Some reports¹⁰ indicate that in individuals between 36 and 80, hearing loss associated with high-frequency deterioration can be attributed to both genetic and environmental factors, with genetic causes accounting for 35% of cases due to inner ear senescence.

Chromobox 4 (CBX4) is the only protein with enzymatic activity and can act as an E3-SUMO ligase for SUMO modification.¹¹ Increased expression of CBX4 caused by oxidative stress resulted in significantly decreased cell viability and elevated CASPase-3 activity, suggesting that SUMOylation inhibits proliferation and promotes apoptosis. This mechanism can occur in presbycusis and noise-induced deafness; the latter has been experimentally reported by Wang et al.¹² in a murine model, where it has been associated with hearing loss and some polymorphic variations of the CBX4 protein.

Thus, at the physiological level, the development of hearing impairment can lead to the loss of primary auditory neurons and neurosensory hair cells due to various factors such as oxidative stress or inner ear deterioration. In this regard, the lack of studies dedicated to the bioinformatic exploration of gene datasets is a starting point for further research, leading to the initial exploration of potential biomarkers. Therefore, with limited pharmacological approaches, the aim of the present work is based on molecular screening based on molecular docking and molecular dynamics studies to search for promising molecules capable of interacting with potential biological targets such as CBX4, which will allow us to predict, prevent and offer proposals for possible therapeutic agents focused on presbycusis.

Methodology

Selection and preparation of ligands

An investigation was conducted to explore various ligands related to presbycusis and hearing¹³ (Table S1, Supplementary Information (SI) section) using the PubChem database.¹⁴ Seventeen molecules were downloaded in SDF format and subjected to an optimization process using Hartree-Fock 6-31G quantum methods through Gaussian software.¹⁵ Finally, the ligands were converted to a PDBQT-compatible format using AutoDock4.2.6 (Molecular Graphics Laboratory, MGL).¹⁶

Selection and preparation of the receptor

According to the genes identified in the core nodes in the preliminary studies, CBX4 was selected as promissory target for presbycusis.¹² The crystallized structure search was developed using the Protein Data Bank (PDB) databases.¹⁷ The selection was performed based on crystal resolution and protein reliability parameters. After protein selection, the structure (5EPL) was downloaded in PDB format for further refinement and optimization. Next, the protein was prepared with Discovery Studio v.2017 software (Dassault Systèmes),¹⁸ incorporating hydrogen atoms, eliminating solvent molecules, and excluding the B chain. The resulting UNC3866 ligand was designated as the reference ligand for subsequent study phases. The proteins were then subjected to the pdb4amber tool and minimized through the Amber20 package (University of California).¹⁹

Molecular docking

To carry out the docking, molecular dynamics, and other data analysis, a gaming PC with an Nvidia RTX 4090 graphics card from MSI, featuring 12 GB of memory, was used. Molecular docking was performed using AutoDock Vina 1.1.2. (Molecular Graphics Laboratory, MGL) software.²⁰ The CBX4 protein and selected ligands were run on a grid spacing of 1.00 Å, with a center grid spacing of x = -38.931, y = 39.961, z = 8.833 Å and shift values of x = 16, y = 20, z = 20 Å. Each molecular docking simulation was performed with a different seed in each run, resulting in a completeness of 8. Nine conformations were determined based on the efficiency value, binding free energy, and root mean square deviation (RMSD). The binding mode, orientation, and conformation of ligands at the binding site were identified. The results of the best affinity conformation were displayed in PDBQT format, visualized using PyMOL version 2.3.2 (Schrödinger) software, and then converted to PDB format.²¹ The results were expressed as the energy value in kcal mol^-1. The 3D ligand-protein complexes, interactions, and binding types were visualized using Molecular Operating Environment (MOE, Chemical Computing Group ULC) software. Molecular docking was performed in triplicate, and the results were expressed as the mean ± standard deviation.

The molecular coupling model was validated by constructing the ROC (Receiver Operating Characteristic) curve, for which 69 decoy molecules that bound little to the protein were used.

Molecular dynamics simulations

The complexes between oridonin, curcumin, and UNC3866, which showed the highest binding affinity against CBX4, were subjected to molecular dynamics (MD) simulations. The AMBER20 software was used to carry out minimization, equilibration, and production, as described by Alvíz-Amador et al.²² The preparation was based on applying the ff14SB force field for proteins and GAFF2 for ligands. Solvation was performed using the TIP3P water model, combined with minimization of 1000 steps of the steep descent method, followed by 1000 steps of the conjugate gradient method. Consequently, it was subjected to a 5000-step heating process from 100 to 300 K every two fs, at constant conditions and with hydrogen bond restriction via the SHAKE algorithm, using a tolerance of 0.00001.

Initial minimization employed a 1000-step steep descent with conjugate gradient minimization of 500 steps using two fs intervals. Subsequently, a molecular dynamics simulation was performed from 50 ps to 300 K, maintaining constant pressure and temperature using a Berendsen constant of 0.2 ps. The simulation time corresponded to 1 µs. The results were developed using CPPTRAJ,²³ expressed as root mean square deviation (RMSD), mobility analysis through root mean square fluctuation calculations (RMSF), solvent accessible surface area (SASA), complex compactness through radius of gyration (RoG), and principal component analysis (PCA) of the systems studied. The free energy of binding with different ligands was calculated using the Molecular Mechanics Generalized Born Surface Area (MMGBSA) method through MMPBSA.py, with a time frame of 1 µs for each molecular dynamics simulation (Table 1).

Table 1
Summary of the MD simulations performed in this study

Initial structure	System	Simulation time / µs
CBX4 native	PDB: 5EPL (non-ligand)	1
CBX4-oridonin	PDB: 5EPL-oridonin	1
CBX4-curcumin	PDB: 5EPL-curcumin	1
CBX4-UNC3866	PDB: 5EPL	1

Prediction of pharmacokinetic properties and toxicity

The pharmacokinetic prediction was developed using the SwissADME online tool of the Swiss Bioinformatics Institute²⁴ and admetSAR, a web server of the East China University of Science and Technology.²⁵ Parameters such as hydrogen bond acceptors and donors, molecular weight, logP, cytochrome P450 (CYP450) inhibitory isoforms, gastrointestinal absorption, P-glycoprotein binding, blood-brain barrier permeability, plasma protein binding, and Caco2 were considered. Additionally, compliance with Lipinski’s rule and ensuring bioavailability are essential. In silico toxicity prediction was also considered using the GUSAR Online server based on the lethal dose 50 (LD₅₀) for rats by oral route and presenting the classification based on the OECD Project.²⁶

Results and Discussion

Molecular docking

The simulation results of the 18 ligands evaluated against the CBX4 protein are detailed in Table S2 (SI section), where the joint energy scores for the most critical interactions are shown in Figure 1. It was identified that the oridonin ligand exhibited the best affinity energy with a value of -8.1 ± 0.0 kcal mol^-1. On the other hand, the ligands curcumin, galagin, ginkgolide B, and ursolic acid completed the top 5 affinity energy, with values of -8.0 ± 0.0, -7.5 ± 0.0, -7.4 ± 0.1, and -7.4 ± 0.1 kcal mol^-1, respectively. In contrast, the ligand UNC3866 (reference molecule) evidenced an energetic affinity of -4.5 ± 0.0 kcal mol^-1. The main interactions for all docked molecules are presented in Figures S1 and S2 (SI section). Regarding the validation of the molecular coupling model, the ROC (receiver operating characteristic) curve showed an area under the curve of 0.9983 ± 0.0019 at a 95% confidence level (Figure S3). The designed ROC curve indicates that the coupling shows a good classification of true positive results over false positives.

Figure 1
Main interactions between oridonin (a), curcumin (b), and UNC3866 peptide (c), reference molecule against CBX4 by molecular docking.

Molecular dynamics simulations

The results of molecular dynamics are shown in Figure 2, as indicated by RMSD, RMSF, RoG, and SASA. In Figure 2a, the native CBX4 receptor presented RMSD values between 2 and 4 Å during the simulation time, while the oridinin molecule showed more excellent stability with RMSD between 1 and 4 Å, only presenting a more significant fluctuation between 750 and 850 ns. The peptide UNC3866 and curcumin also exhibited trajectories between 1.5 and 5 Å about the native protein, with evidence in UNC3866 of an average stability of approximately 2.5 Å. In Figure 2b, the CBX4 protein and the CBX4-curcumin system exhibited RoG with minor variations and constants, indicating 11 and 12 Å values, which represent higher stability during the simulation trajectory.

Figure 2
Molecular dynamics analysis among natural metabolites and UNC3866 with CBX4. (a) RMSD; (b) RoG; (c) SASA; (d) RMSF.

However, the ligand UNC3866 exhibits moderate fluctuations for CBX4, with changes evident between 12 and 13.5 Å, and a slight variation between 850 and 900 ns. Oridonin presents the most pronounced fluctuations, with intermittent peaks, highlighting a significant increase between 600 and 850 ns. In Figure 2c, it was observed that the SASA values for CBX4 remained in the range of 4000 to 4500 Ų throughout the entire simulation. Similarly, the curcumin ligand behaved consistently within this same range throughout the simulation. However, the oridonin ligand showed relative stability in the 4000 to 4500 Ų range, except for a notable peak exceeding 5000 Ų between 750 and 800 ns of simulation.

The most considerable fluctuations in SASA values were observed in the UNC3866 ligand, which on average recorded values between 4500 and 5000 Ų and achieved a maximum of 5700 Ų between 250 and 300 ns, associated with lower stability to the compared systems. In Figure 2d, fluctuations of the residues were evidenced in a range of 8 to 7 Å and then decreased to values between 1 and 3 Å, remaining relatively stable in most cases, except in the case of curcumin. This ligand consistently showed more significant fluctuations, especially in the residue range 40 to 59, where elevations greater than 4 Å were recorded.

The histogram in Figure 3 displays the results of principal component analysis (PCA) for the molecular dynamics of molecules with the strongest interactions with the CBX4 protein. Projection 1 of the coordinates along the eigenvectors with the highest eigenvalues for each trajectory of native CBX4 in red, CBX4-oridonin in blue, CBX4-curcumin (violet), and CBX4-UNC3866 (yellow).

Figure 3
Histogram of the principal component analysis (PCA) between native CBX4 (red), oridonin (blue), curcumin (violet), and CBX4 UNC3866 (yellow).

By comparing the curves, you can infer how different ligands (oridonin, curcumin, UNC3866) influence the conformational dynamics of CBX4. Oridonin appears to restrict the movement of the protein with a curve that is significantly blue and PC1 approximately 0, leading to a very stable and well-defined conformation. Curcumin shows a broader distribution, with a significant peak around PC1 = 10-15 and a smaller peak around PC1 = -5. This indicates that the CBX4 protein exhibits more conformational flexibility or occupies a broader range of structural states when bound to curcumin compared to oridonin. The curve for CBX4 bound to UNC3866 exhibits a broader and somewhat bimodal distribution, with peaks at PC1 = -10 and PC1 = 5. This suggests that UNC3866 also allows for a range of conformations, possibly two main groups of states.

Moreover, this curve represents the native (unbound) CBX4 protein. It displays a comprehensive and somewhat multi-modal distribution, spanning a wide range of PC1 values (from approximately -30 to 20). This is typical for a native protein, as it tends to be more dynamic and explore a larger conformational space when not bound to a ligand. The peaks around PC1 = -20, 0, and 10-15 suggest that even in its native state, CBX4 might have preferred conformational substates.

Free energy calculation

A 1 µs simulation was performed using 100 ns of the trajectory for the stable binding energy analysis, employing MMGBSA calculations with polar and non-polar solvation parameters. Table 2 reveals that the average binding energy for CBX4-curcumin was -35.0283 kcal mol^-1. Similarly, CBX4-UNC3866 and CBX4-oridonin exhibited binding energies of -9.3128 and -3.3089 kcal mol^-1, respectively. Nonpolar contributions van der Waals (VDWAAL) and SASA to binding were more pronounced. This is evidenced by the significant negative VDWAAL energy values recorded for the complexes. The energy of the curcumin complex was the most significant contributor to the hydrophobic contributions with the protein (-38.1896 kJ mol^-1), as evidenced in Figure 1 by the interactions of amino acids such as Val10, Phe11, Val13, Trp32, Trp35, Ile48, Leu49, and Leu53. The hydrogen bonds formed by the reference peptide when interacting with the protein are characterized by the electrostatic energy presented in Table 2, which is -141.3494 kcal mol^-1, indicating a hydrophilic character for this molecule. The opposite is true for curcumin and oridonin, whose values are incredibly high compared to the reference compound.

Table 2
Calculate the free energy of higher-affinity metabolites and the reference peptide with CBX4

Energy components	Oridonin	Curcumin	UNC3866
VDWAALS / (kcal mol-1)	-19.8147	-38.1896	-6.4842
EEL / (kcal mol-1)	-5.7694	-7.5088	-141.3494
EGB / (kcal mol-1)	12.4097	15.0585	76.2423
ESURF / (kcal mol-1)	-12.7845	-4.3885	-1.3760
∆Ggas / (kcal mol-1)	-25.5841	-45.6983	-84.1791
∆Gsolv / (kcal mol-1)	22.2752	10.6700	74.8663
∆GTOTAL / (kcal mol-1)	-3.3089	-35.0283	-9.3128

Prediction of pharmacokinetic and toxicological properties

The pharmacokinetic and toxicological predictions established that the ligands oridonin and curcumin presented similar values to the ligand UNC3866, which violates three of the Lipinski rules (see Table 3). Where the models showed a good intestinal absorption capacity; it does not present permeation to the blood-brain barrier, it is characterized by inhibitors of CYP3A4 and 2D6, involved in the metabolism of xenobiotics that could influence its absorption, and finally, its bioavailability; however, the predictive models used indicated that it has a coefficient of 0.55, which characterizes a considerable bioavailability.

Table 3
ADME properties (absorption, distribution, metabolism and excretion) and in silico toxicity using ADMETSAR, SwissADME, and GUSAR online prediction servers

Property	Oridonin	Curcumin	UNC3866
Molecular weighta / (g mol-1)	364.44	368.38	795.02
No. H-bond acceptorsb	6	6	9
No. H-bond donorsc	4	2	6
Lipinski No. violationsd	0	0	3
TPSAe	107.22	93.06	195.27
GI absorptionf	high	high	low
BBB permeantg	no	no	no
Bioavailability score	0.55	0.55	0.55
Pgp substrate	yes	no	yes
CYP1A2 inhibitor	no	no	no
CYP2D6 inhibitor	no	no	no
CYP3A4 inhibitor	no	yes	no
log Kp / (cm s-1)	-8.76	-6.28	-7.23
Caco-2	0.3351	0.6485	0.7108
Rat oral LD50 / (mg kg-1)	3.0769 (III)	2.5468 (III)	2.5617 (III)

• ^a
  Molecular weight less than 500 g mol^-1;
• ^b
  less than or equal to 10 hydrogen bond acceptors; ^c less than or equal to 5 hydrogen bond donors; ^dLipinski’s rule of five violations less than or equal to 1; ^etopological polar surface area (TPSA) less than or equal to 140.2;
• ^f
  gastrointestinal absorption;
• ^g
  blood-brain barrier permeant.

Preliminary studies explored the CBX4 gene identified for their potential involvement in hearing loss.¹² The enrichment study revealed that genes such as CBX4 are linked to the regulation and metabolism of nucleosides, nucleotides, and nucleic acids. Moreover, the molecular functions of these genes can be associated with transcriptional regulatory activities by 20%, while ribonucleic acid (RNA) binding accounts for 60%.²⁷ Conversely, the Chromobox 4 (CBX4) system is identified as a protein linked with enzymatic activity, specifically with SUMO E3 ligase functions for SUMO modification.²⁸ Increased expression of CBX4 caused by oxidative stress has been linked to a significant decrease in cell viability and an elevation of CASPASE-3 activity, suggesting that SUMOylation inhibits prooxidative stress has been linked to a substantial decrease in cell viability and an elevation of CASPASE-3 activity, suggesting that SUMOylation inhibits proliferation and promotes apoptotic mechanisms. This mechanism may occur in presbycusis and noise-induced deafness, as reported experimentally by Wang et al.¹² in murine models, where it has been associated with hearing loss and some polymorphic variations of the CBX4 protein. Specifically, it was demonstrated that the transcriptional factor SP1 influences the promoter activity of the intron rs1285250 of CBX4. Likewise, the suppression of SP1 was associated with decreased CBX4 expression and reduced apoptosis in HEI-OC1 cells.

Regarding the molecular docking results, the ligands with the best interaction affinity with the CBX4 protein were oridonin and curcumin. These are related candidates for drugs that play a protective role in the ear against noise-induced injury. Oridonin has been studied in a murine model as a modulator of the NLRP3 receptor that could block the inflammasome.²⁹ It has also been shown in a murine model to prevent the recurrent impact of noise on the ear through an antioxidant effect.³⁰ Additionally, although oridonin has been linked to anti-inflammatory activity and curcumin has its antioxidant potential, it has been considered that the latter, being an antioxidant and playing a role in preventing oxidative stress, could be a pharmacological target of action the CBX4 receptor in presbycusis and noise-induced hearing loss.³¹

In the in silico study, the main interactions at the atomic level between oridonin and CBX4 were examined, demonstrating the electron-accepting capacity between the OH group at position 9 of the structure and the Thr41 residue. Additionally, the participation of acidic and basic residues, such as Glu43 and His9, was also observed. However, there is evidence of distribution in the pocket of hydrophobic interactions involving Phe11, Trp32, and Trp35, which may play a stabilizing role in the binding structure to the CBX4 binding site. On the other hand, upon binding to CBX4, the curcumin molecule exhibited a broader binding affinity to the target protein, indicating a preference for polar residues with acid-base properties, such as Glu43 and His9. However, apolar arene-arene interactions between one of the aromatic rings of curcumin and the Trp32 residue were notably significant. This contribution of arene-arene interactions seems relevant to the contribution of dispersion strength in biological systems, which tends to be more favorable with the more complex presence of substituents on the interacting arene groups, independent of their electronic capacity.³² Otherwise, the redocking of peptide (UNC3866) seems to indicate that methyl-lysine (Kme3) interacts with residues such as Phe11, Trp32, Trp35, and a polar residue such as Tyr39, which form the aromatic cage that guarantees stability in the pocket through the predominance of hydrophobic bonds, which can be considered relevant in the recognition between the CBX chromodomain and potential inhibitors. Furthermore, it was identified that Asn47 forms polar interactions with UNC3688, where it was shown that Asn47 formed hydrogen bonds with the C-terminal group of the peptide, establishing the role of the Glu43 residue in the increase of hydrogen bonding contacts, evidenced in the study with the formation of acidic polar residues.¹¹ Reports³⁵ have also shown that including aromatic groups favors stacking interactions and increases inhibitory activity in similar systems, such as CBX7 (see Table 4).

Table 4
Affinities and interrelationships between major ligands and CBX4

Ligand	Binding energy / (kcal mol-1)	Hydrophobic interaction	Polar interaction
Oridonin	-8.1 ± 0.0	Phe11, Trp35, Trp32	His9, Tyr39, Thr41, Glu43
Curcumin	-8.0 ± 0.0	Val10, Phe11, Val13, Trp32, Trp35, Ile48, Leu49, Leu53	His9, Tyr39, Thr41, Glu43, Asn47, Asp50
Galagin	-7.5 ± 0.0	Phe11, Trp32, Trp35	His9, Tyr39, Thr41, Glu43, Asn47
Ginkgolide B	-7.4 ± 0.1	Phe11, Trp32, Trp35	Glu8, His9, Tyr39, Thr41, Glu43
Ursolic acid	-7.4 ± 0.1	Phe11, Trp32, Trp35	His9, Tyr39, Thr41, Glu43, Asn47
UNC3866	-4.5 ± 0.4	Val10, Phe11, Trp32, Trp35, Trp42, Leu49, Pro44	Glu8, His9, Thr41, Glu43, Asn47

Similarly, the molecular dynamics simulation results show that the CBX4 system simulated in this study behaves similarly to other CBX2 and CBX7 polycomb proteins, with RMSD values between 1.5 and 3 Å, both in the native and bound states to the natural inhibitor UNC3688, which is a promising peptide for the treatment of hepatocellular cancer. The RMSD results reported by Liu et al.,¹¹ as cited by Deng et al.³³ regarding RMSF, indicate that the CBX4-oridonin and CBX4-UNC3866 systems exhibit similar behavior, with minimal variations in residue mobility compared to the native system. Notably, oridonin is more stable throughout the simulation time. It is important to note that this stability of the complexes occurs in areas between the 30-40 aa residues, which may be supported by the studies of Lamb et al.,³⁴ who describe that the most significant contacts between CBX7 and deoxyribonucleic (DNA) systems occurred in the Trp32-Ser40 loop, generating a groove binding with greater depth. Furthermore, residues involved in the aromatic cage, such as Trp35 and Tyr39, appear to be stabilized by methyl-lysine mimetic groups of CBX7 inhibitory peptides, which have close contact with DNA. However, the CBX4-curcumin system shows a more significant fluctuation between the zone of residues above 40 aa, which may be linked to the mobility of some key residues such as Thr41, Glu43, Asn47, Ile48, Leu49, Asp50, and Leu53, relevant in the interaction with the chromodomain systems presented similarly in CBX7.³⁵

The SASA values obtained in the CBX4 systems show lower values for the CBX4-curcurmin complex, which was similar to that presented by the native protein, establishing that the system showed more compaction; however, the oridonin and UNC3688 systems showed variations in the majority of the simulation time, and even the reference system presented a more significant change in folding and diffusion in the interaction of the ligand with the protein (Rahimi et al.).³⁶ In parallel, the RoG results established that the CBX4-curcumin complex showed lower RoG values than the other systems and was similar to the behavior of the native protein, suggesting considerable compactness and stability. While the influence of the selective inhibitor UNC3688 appears to induce unfolding of the protein, as indicated by the increase in RoG values throughout the simulation time compared to the native protein, the CBX4-curcumin complex exhibits lower RoG values than the other systems. Like the native protein, this suggests considerable compactness and stability. For the CBX4-oridonin complex, increases of around 15 Å in 100 to 150 ns, 250 to 300 ns, and more abundant fluctuations from 500 ns onwards, which could be associated with the most flexibility, lower compactness of CBX4 in the presence of the ligand and lower structural stability of the system (Ali et al.).³⁷

In the analysis of the first principal component (P1:1), it is observed that it varies between approximately -40 and 20 on the X-axis, capturing most variability of the data. The frequency on the Y-axis ranges from 0 to 0.07, representing the frequency of conformations as a function of the principal component. Native CBX4 in red shows a variable distribution around P1:1 = -20 and P1:1 20 with a frequency of 0.035, while CBX4-oridonin in blue exhibits a sharp peak around P1:1 = -10, indicating specific conformational states. On the other hand, CBX4 curcumin in violet exhibits a broader distribution around P1:1 = 0, suggesting higher conformational adaptability. The inhibitor UNC3866 shows two prominent peaks at approximately P1:1 = 10 and P1:1 = 15, which could indicate two stable conformations or binding modes during simulation. In summary, curcumin induces fluctuations leading to more variable conformational changes in CBX4, but at specific sites, which is consistent with the RMSF analysis.

On the other hand, it is essential to highlight the capacity for conformational adaptation of the CBX4 protein with curcumin, which is reinforced by the calculation of free energy by the generalized Bohr method, where the affinity energy of curcumin for this receptor is -35.0283 kcal mol ¹ higher than that reported for oridonin and the natural inhibitor UNC3866. Additionally, curcumin presented excellent results in pharmacokinetic and toxicological predictions, which make it a promising ligand with an affinity for CBX4 superior to UNC3866 and oridonin, characterized by conformational stability, fluctuations, and compaction.

Conclusions

Potential hub genes for the prophylactic treatment of presbycusis have been identified, as in the case of CBX4, which has an available crystallographic structure and a known inhibitor (UNC3866), providing a basis for investigating potential pharmacological agents that could inhibit its activity and prevent hearing loss. Thus, molecular docking results revealed that oridonin and curcumin showed the best interaction affinity with CBX4, indicating representative interactions that suggest better behavior for UNC3866. Molecular dynamics simulation showed that simulated CBX4 behaves similarly to other polycomb proteins, with high affinity and conformational stability with curcumin and oridonin, making them promising ligands for future studies. In conclusion, curcumin and oridonin could be promising pharmacological agents for treating presbycusis and noise-induced hearing loss by interacting favorably with CBX4. However, further research is needed to validate these findings and develop effective therapies.

Supplementary Information

Supplementary data are available free of charge at http://jbcs.sbq.org.br as PDF file.

Data Availability Statement

All data for this manuscript are available in the text.

Acknowledgments

The authors thank the University of Cartagena and the Rafael Núñez University Corporation.

References

• ¹ Rodríguez-Valiente, A.; Álvarez-Montero, Ó.; Górriz-Gil, C.; García-Berrocal, J. R.; Acta Otorrinolaringologica 2020, 71, 175. [Crossref]
  » Crossref
• ² Labambe, N.; Tenri Sanna, A.; Maulani, D.; Baso Sulaiman, V.; Hamriani, I.; J. EduHealth 2024, 15, 154. [Crossref]
  » Crossref
• ³ Gutiérrez-Torres, A.; Baez, S. In The Palgrave Encyclopedia of Disability; Bennett, G.; Goodall, E., eds.; Springer Nature: Cham, Switzerland, 2024, p. 1-13. [Crossref]
  » Crossref
• ⁴ Yilmaz, N.; Soylemez, E.; Sanuc, M. B.; Bayrak, M. H.; Sener, V.; Eur. Arch. Oto-Rhino-Laryngol. 2023, 280, 2265. [Crossref]
  » Crossref
• ⁵ Fancello, V.; Fancello, G.; Palma, S.; Monzani, D.; Genovese, E.; Bianchini, C.; Ciorba, A.; Medicina 2023, 59, 608. [Crossref]
  » Crossref
• ⁶ Wang, J.; Puel, J. L.; J. Clin. Med. 2020, 9, 218. [Crossref]
  » Crossref
• ⁷ Jung, S. H.; Lee, Y. C.; Shivakumar, M.; Kim, J.; Yun, J. S.; Park, W. Y.; Won, H. H.; Kim, D.; Biobank, P. M.; BMC Med 2024, 22, 141. [Crossref]
  » Crossref
• ⁸ Ninoyu, Y.; Friedman, R. A.; Trends Genet 2024, 40, 228. [Crossref]
  » Crossref
• ⁹ Boucher, S.; Tai, F. W. J.; Delmaghani, S.; Lelli, A.; Singh Estivalet, A.; Dupont, T.; Niasme-Grare, M.; Michel, V.; Wolff, N.; Bahloul, A.; Bouyacoub, Y.; Bouccara, D.; Fraysse, B.; Deguine, O.; Collet, L.; Thai-Van, H.; Ionescu, E.; Kemeny, J. L.; Giraudet, F.; Lavieille, J.-P.; Devèze, A.; Roudevitch Pujol, A.L.; Vincent, C.; Renard, C.; Franco-Vidal, V.; Thibult-Apt, C.; Darrouzet, V.; Bizaguet, E.: Coez, A.; Aschard, H.; Michalski, N.; Lefevre, G. M.; Aubois, A.; Avan, P.; Bonnet, C.; Petit, C.; Proc. Natl. Acad. Sci. U. S. A. 2020, 117, 31278. [Crossref]
  » Crossref
• ¹⁰ Jayakody, D. M. P.; Friedland, P. L.; Martins, R. N.; Sohrabi, H. R.; Front Neurosci 2018, 12 [Crossref]
  » Crossref
• ¹¹ Liu, H.; Li, Z.; Li, L.; Comput. Biol. Chem 2018, 74, 339. [Crossref]
  » Crossref
• ¹² Wang, B.; Wan, L.; Sun, P.; Zhang, L.; Han, L.; Zhang, H.; Zhang, J.; Pu, Y.; Zhu, B.; Hum. Mol. Genet 2022, 31, 2109. [Crossref]
  » Crossref
• ¹³ Wu, Y.; Zhang, J.; Liu, Q.; Miao, Z.; Chai, R.; Chen, W.; Acta Pharm. Sin. B 2024, 14, 455. [Crossref]
  » Crossref
• ¹⁴ Kim, S.; Thiessen, P. A.; Bolton, E. E.; Chen, J.; Fu, G.; Gindulyte, A.; Han, L.; He, J.; He, S.; Shoemaker, B. A.; Wang, J.; Yu, B.; Zhang, J.; Bryant, S. H.; Nucleic Acids Res 2016, 44, D1202. [Crossref]
  » Crossref
• ¹⁵ Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J.; Gaussian 09, Revision E.01; Gaussian, Inc., Wallingford CT, 2009.
• ¹⁶ Morris, G. M.; Huey, R.; Lindstrom, W.; Sanner, M. F.; Belew, R. K.; Goodsell, D. S.; Olson, A. J.; J. Comput. Chem 2009, 30, 2785. [Crossref]
  » Crossref
• ¹⁷ Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T. N.; Weissig, H.; Shindyalov, I. N.; Bourne, P. E.; Nucleic Acids Res 2000, 28, 235. [Crossref]
  » Crossref
• ¹⁸ Dassault Systèmes BIOVIA, Discovery Studio Visualizer 4.5, 2016. [Crossref] accessed in June 2025
  » Crossref
• ¹⁹ Case, D. A.; Aktulga, H. M.; Belfon, K.; Cerutti, D. S.; Cisneros, G. A.; Cruzeiro, V. W. D.; Forouzesh, N.; Giese, T. J.; Götz, A. W.; Gohlke, H.; Izadi, S.; Kasavajhala, K.; Kaymak, M. C.; King, E.; Kurtzman, T.; Lee, T. S.; Li, P.; Liu, J.; Luchko, T.; Luo, R.; Manathunga, M.; Machado, M. R.; Nguyen, H. M.; O’Hearn, K. A.; Onufriev, A. V.; Pan, F.; Pantano, S.; Qi, R.; Rahnamoun, A.; Risheh, A.; Schott-Verdugo, S.; Shajan, A.; Swails, J.; Wang, J.; Wei, H.; Wu, X.; Wu, Y.; Zhang, S.; Zhao, S.; Zhu, Q.; Cheatham III, T. E.; Roe, D. R.; Roitberg, A.; Simmerling, C.; York, D. M.; Nagan, M. C.; Merz Jr., K. M.; J. Chem. Inf. Model. 2023, 63, 6183. [Crossref]
  » Crossref
• ²⁰ Trott, O.; Olson, A. J.; J. Mol. Struct 2010, 31, 455. [Crossref]
  » Crossref
• ²¹ Yuan, S.; Chan, H. C. S.; Hu, Z.; Wiley Interdiscip. Rev.:Comput. Mol. Sci. 2017, 7, e1298. [Crossref]
  » Crossref
• ²² Alvíz-Amador, A.; Galindo-Murillo, R.; Perez-Gonzalez, H.; Rodríguez-Cavallo, E.; Vivas-Reyes, R.; Mendez-Cuadro, D.; J. Chem. Inf. Model. 2020, 60, 805. [Crossref]
  » Crossref
• ²³ Roe, D. R.; Cheatham, T. E.; J. Chem. Theory Comput. 2013, 9, 3084. [Crossref]
  » Crossref
• ²⁴ Daina, A.; Michielin, O.; Zoete, V.; Sci. Rep. 2017, 7, 42717. [Crossref]
  » Crossref
• ²⁵ Yang, H.; Lou, C.; Sun, L.; Li, J.; Cai, Y.; Wang, Z.; Li, W.; Liu, G.; Tang, Y.; Bioinformatics 2019, 35, 1067. [Crossref]
  » Crossref
• ²⁶ Lagunin, A.; Zakharov, A.; Filimonov, D.; Poroikov, V.; Mol. Inform 2011, 30, 241. [Crossref]
  » Crossref
• ²⁷ Zhao, R.; Guo, Y.; Zhang, L.; Huang, Z.; Li, X.; Lan, B.; Zhong, D.; Chen, H.; Xuan, C.; Cell Death Dis 2024, 15, 378. [Crossref]
  » Crossref
• ²⁸ Soria-Bretones, I.; Cepeda-García, C.; Checa-Rodriguez, C.; Heyer, V.; Reina-San-Martin, B.; Soutoglou, E.; Huertas, P.; Nat. Commun 2017, 8, 113. [Crossref]
  » Crossref
• ²⁹ He, H.; Jiang, H.; Chen, Y.; Ye, J.; Wang, A.; Wang, C.; Liu, Q.; Liang, G.; Deng, X.; Jiang, W.; Zhou, R.; Nat. Commun 2018, 9, 2550. [Crossref]
  » Crossref
• ³⁰ Li, N.; Yan, X.; Huang, W.; Chu, M.; Dong, Y.; Song, H.; Peng, Y.; Shi, J.; Liu, Q.; Biochem. Pharmacol 2023, 212, 115575. [Crossref]
  » Crossref
• ³¹ Doostkam, A.; Iravani, K.; Malekmakan, L.; Gholamabbas, G.; Roozbeh, J.; Soltaniesmaeili, A.; Sci. Rep 2024, 14, 17576. [Crossref]
  » Crossref
• ³² Andreasson, M.; Donzel, M.; Abrahamsson, A.; Berner, A.; Doimo, M.; Quiroga, A.; Eriksson, A.; Chao, Y. K.; Overman, J.; Pemberton, N.; Wanrooij, S.; Chorell, E.; J. Med. Chem. 2024, 67, 2202. [Crossref]
  » Crossref
• ³³ Deng, N.; Dai, W.; Levy, R. M.; Phys. Chem. B 2013, 117, 12787. [Crossref]
  » Crossref
• ³⁴ Lamb, K. N.; Bsteh, D.; Dishman, S. N.; Moussa, H. F.; Fan, H.; Stuckey, J. I.; Norris, J. L.; Cholensky, S. H.; Li, D.; Wang, J.; Sagum, C.; Stanton, B. Z.; Bedford, M. T.; Pearce, K. H.; Kenakin, T. P.; Kireev, D. B.; Wang, G. G.; James, L. I.; Bell, O.; Frye, S. V.; Cell Chem. Biol 2019, 26, 1365. [Crossref]
  » Crossref
• ³⁵ Dong, C.; Liu, Y.; Lyu, T.-J.; Beldar, S.; Lamb, K. N.; Tempel, W.; Li, Y.; Li, Z.; James, L. I.; Qin, S.; Wang, Y.; Min, J.; Cell Chem. Biol 2020, 27, 827. [Crossref]
  » Crossref
• ³⁶ Rahimi, M.; Taghdir, M.; Abasi, F.; Sci. Rep 2023, 13, 14179. [Crossref]
  » Crossref
• ³⁷ Ali, S.; Hassan, Md.; Islam, A.; Ahmad, F.; Curr. Protein Pept. Sci. 2014, 15, 456. [Crossref]
  » Crossref

• Editor handled this article:
  Paulo Augusto Netz (Associate)

Publication Dates

• Publication in this collection
  29 Sept 2025
• Date of issue
  2025

History

• Received
  22 Apr 2025
• Accepted
  21 Aug 2025