Screening of microbial bioactives as potential breast cancer inhibitors: A structure-based multitargeted molecular docking analysis

Abstract

Breast cancer still ranks among the top causes of death globally; therefore, new therapeutic agents are urgently needed. Microbe-derived bioactive compounds present great promise for novel drug discovery. Considering this, the study aims to evaluate their binding affinity, efficacy, bioavailability, and toxicity against key breast cancer target proteins. Molecular docking was performed on 80 bacterial compounds against six breast cancer targets (Human Epidermal Growth Factor Receptor 2 (HER2), Estrogen Receptor Alpha (ERα), B-Cell Lymphoma 2 (BCL2), Breast Cancer Gene 1 (BRCA1), Breast Cancer Gene 2 (BRCA2), and Peroxisome Proliferator–Activated Receptor Gamma (PPARG)). Lead compounds were identified through molecular docking, the Lipinski rule of five, and toxicity profiling. Molecular dynamics simulations assessed the stability and dynamics of top ligand–protein complexes. In addition, a comparative analysis against known phytochemicals or drugs for breast cancer was performed. Eighteen microbial compounds showed strong binding affinities with multiple proteins. Marfuraquinocin A and Marfuraquinocin D were identified as promising candidates due to their favorable pharmacokinetic and safety profiles. Molecular dynamics simulations confirmed the structural stability of the top docked complexes, with minimal root mean square deviation fluctuations. Besides, comparative evaluation showed significant differences in binding affinities between microbial bioactives and known phytochemicals or drugs. The results suggest that microbial bioactives, particularly Marfuraquinocin A and Marfuraquinocin D, can potentially be effective breast cancer inhibitors. Future in vitro and in vivo experiments are recommended to confirm their therapeutic effectiveness and safety.

Keywords

microbial bioactive breast cancer inhibitor molecular docking in silico therapeutics

Introduction

Despite remarkable medical advances, breast cancer remains one of the leading causes of mortality and morbidity worldwide. With a projected 2.3 million new breast cancer cases in 2020, accounting for 11.7% of all new malignancies, there were 674,996 deaths.¹ Globally, rates of breast cancer are conspicuously higher in women than in men. The major risk factors of breast cancer include reproductive and hormonal risk factors, lifestyle risk factors, family history, and genetic predisposition. Current conventional treatments, such as surgery, chemotherapy (CT), radiation therapy (RT), and endocrine therapy (ET), effectively combat breast cancer but frequently have several complications such as fatigue, nausea, lymphedema, heart problems, and neurological issues.^2,3 Immunotherapy, macrophage-mediated therapy, and nanoparticle-based therapy have been developed and have shown encouraging findings, although their clinical usefulness is restricted.^4,5 Furthermore, the development of a multidrug-resistant phenotype in metastatic breast cancer is the primary cause of the failure of existing treatment regimens.^6,7 Hence, the development of therapeutic techniques that target tumor cells has gained significance as promising alternatives to conventional medicines that lack tumor selectivity and clinical efficacy and produce greater adverse effects. The limits of conventional cancer medicines have sparked interest in alternative therapeutic approaches, particularly those derived from natural sources, such as phytochemicals and microbial bioactives. A total of 77% of antibiotics approved by the Food and Drug Administration (FDA) since 2000 are derived from various natural compounds and bioactive microorganisms.⁸ Phytochemicals, which are naturally occurring molecules found in plants, mushrooms, and seaweeds, have long been known as anti-cancer medications, with approximately 2/3 of the potential therapeutic advantages.^9,10 They can enhance the action of other popular drugs and benefit therapeutics by reversing drug resistance and addressing new targets. Furthermore, they work as a therapy system on their own because they can function in a variety of ways without causing unexpected side effects, allowing them to affect several pathways.⁹ These plant-derived compounds, such as flavonoids, polyketides, polyphenols, metalloids, and alkaloids, have shown promise in various studies for their ability to inhibit cancer cell growth, lower migratory capacity, activate autophagy, arrest the cell cycle in the G2 phase, and induce reactive oxygen species (ROS)-dependent apoptosis.^9,11,12 Despite their potential, phytochemicals face challenges such as low bioavailability and stability, which may limit their therapeutic efficacy. Although natural products are promising anti-breast cancer drugs, cancer treatment is a complex process and we still rely on existing anti-cancer therapies.

Microorganisms create a wide range of natural chemicals with medicinal potential, including pigments, alkaloids, poisons, antibiotics, gibberellins, carotenoids, and biosurfactants.^13,14 Furthermore, tumor therapy based on bacteria/bacterial components may be a great way to produce systemic anti-tumor immune responses, boosting immunity and reducing tumor progression, which is an important element influencing cancer treatment.¹⁵ Secondary metabolites target endothelial cells and can impede angiogenesis. As they are able to break down carcinogens, prevent the transformation of procarcinogens into carcinogens, increase the activity of anti-cancer enzymes, and trigger apoptosis, similar to phytochemicals. According to recent studies, cancer can be prevented by upregulating pro-apoptotic or downregulating anti-apoptotic changes, which have anti-cancer effects on cancer cells. Surprisingly, certain microorganisms do not have cytotoxic effects on healthy cells but only on malignant cells. However, research on microbial bioactives is still in progress, with more studies needed to completely understand their efficacy and safety profiles in cancer treatment. Furthermore, the interactions between microbes and host cells, as well as the risk of microbial infections in immunocompromised patients, need to be properly understood.

Given the potential features of phytochemical and microbial compounds, a comparative investigation is needed to establish their strengths and limitations as cancer treatments. Unlike phytochemicals, microbial bioactives frequently demonstrate novel modes of action, making them promising candidates for drug development. For example, the mechanisms of action of the microbiota include modifying anti-inflammatory cytokines, adjusting prostaglandin release, and activating phagocytes, which in turn kill cancer cells in the early stages of the disease.¹³ Moreover, microbial metabolites are favorable in terms of production of constant quality and wide scalability, which is more trustworthy than extracting plant-based chemicals, which might vary depending on the plant source and environmental conditions. The diversity and potency of microbial bioactive compounds may make them more effective replacements or complements of phytochemicals in oncology. In recent years, computational methods such as molecular docking have been increasingly useful in drug discovery and development, particularly for assessing the interaction of natural chemicals with cancer-related targets. The integration of in silico and in vitro evaluations has become a cornerstone in identifying potent inhibitors for various malignancies. The primary goal of molecular docking is to contribute to predicting the ligand-receptor complex structure via computational techniques. Molecular dynamics (MD) simulations are a powerful computational method for studying the behavior of biological systems at the atomic level.¹⁶ By applying Newton’s laws to evaluate molecular motions, MD simulations provide insights into protein–ligand interactions, structural dynamics, and drug design.¹⁷ In order to be time and cost-effective, it is practicable to research medication compounds and their properties via commercial/industrial software.¹⁸

The management of breast cancer is increasingly hindered by the emergence of multidrug resistance, which often results from the compensatory activation of alternative signaling pathways. This necessitates the discovery of chemical agents capable of multitargeted inhibition to achieve a more robust therapeutic effect. While phytochemicals have been the traditional focus of natural product screening, microbial secondary metabolites remain largely unexplored for breast cancer drug discovery, justifying the need for identifying novel drug candidates. This study was conducted with the aim of screening microbial-derived compounds with anti-cancer potential to find prospective inhibitors of six breast cancer receptors (Human Epidermal Growth Factor Receptor 2 (HER2), Estrogen Receptor Alpha (ERα), B-Cell Lymphoma 2 (BCL2), Breast Cancer Gene 1 (BRCA1), Breast Cancer Gene 2 (BRCA2), and Peroxisome Proliferator–Activated Receptor Gamma (PPARG)) through a comprehensive in silico approach. Molecular docking simulations were used to determine the binding affinities of these compounds with the target proteins, providing insight into their potential inhibitory effects. The compounds were further evaluated for their pharmacological properties, including drug-likeness and oral bioavailability, through ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) predictions. Also, MD simulations were performed to evaluate the structural integrity and dynamic interactions of the top ligand-protein complexes. In addition, a statistical analysis was performed to compare the efficacy of microbial-derived compounds with phytochemicals and existing drugs, highlighting the superior potential of microbial bioactives.

Materials and methods

Selection of target proteins

On the basis of a literature search, six target proteins (HER2, ERα, BCL2, BRCA1, BRCA2, and PPARG) were selected as potential targets for breast cancer treatment. The three-dimensional (3D) crystal structures of HER2 (Protein Data Bank (PDB) ID: 3PP0), ERα (PDB ID: 3ERT), BCL2 (PDB ID: 4MAN), BRCA1 (PDB ID: 4Y2G), BRCA2 (PDB ID: 3EU7), and PPARG (PDB ID: 1nyx) were retrieved from the RCSB Protein Data Bank (https://www.rcsb.org). Descriptions of the selected proteins are given in Table 1.

Table 1.

Target proteins and their functions.

Protein	Description	Function	PDB ID	Reference
1. HER2	Human Epidermal Growth Factor Receptor 2	Function involves tightly binding to the cell surface receptor and plays a crucial role in the development, multiplication, and specialization of various types of cells	3PP0	Sohrab and Kamal¹⁹
2. ER Alpha	Estrogen Receptor Alpha	Regulates transcription of specific target genes by binding to associated DNA regulatory sequences	3ERT	Acharya et al.²⁰
3. BCL2	B-Cell Lymphoma 2	Anti-apoptotic protein that supports prolonged cell survival and growth	4MAN	Taghizadeh et al.²¹
4. BRCA1	Breast Cancer Gene 1	Repairs damaged DNA, regulates the cell cycle, and acts as a transcriptional activator	4Y2G	Hossain et al.⁶
5. BRCA2	Breast Cancer Gene 2	Maintains genome integrity, protects replication fork, and is involved in DNA repair	3EU7	Hossain et al.⁶
6. PPARG	Peroxisome Proliferator–Activated Receptor Gamma	Controls fatty acid storage and adipocyte differentiation and regulates glucose and lipid metabolism	1NYX	Basavarajappa et al.²²

Selection of microbial bioactive compounds

After a comprehensive literature search from January 2024 to April 2024, 130 bioactive bacterial compounds were obtained. The microbial extracts were searched and listed using the Google Scholar search engine and PubMed. The search keywords included “microbial bioactive compounds” OR “microbial metabolites” OR “bacterial bioactive compounds” AND “microbial derivatives” OR “bacterial bioactive extracts.” Among 130 compounds, the two-dimensional (2D) structures of 80 compounds were available on PubChem. The 2D structures of these 80 bacterial bioactive compounds were subsequently obtained in structured data file (SDF) format from the PubChem library (https://pubchem.ncbi.nlm.nih.gov) to evaluate their potency as inhibitors of HER2, ERα, BCL2, BRCA1, BRCA2, and PPARG.

Preparation of protein and ligand structures

The energy of the selected proteins was minimized. Then, co-crystallized water molecules, small molecules, nonpolar hydrogens, heteroatoms, and nonstandard residues were deleted, and the missing atoms were repaired. Hydrogens and Kollman charges were added and merged. All these steps were completed via MGL Tools 1.5.7. Finally, the protein files were converted into AutoDock4 compatible mode and saved in Protein Data Bank, Partial Charge (Q), & Atom Type (T) (PDBQT) format.²³ The structures of the 80 ligand molecules downloaded from PubChem were exported to Open Babel software 2.3.2 and converted to PDB format. Then, using MGL Tools 1.5.7, Gasteiger charges were added to these molecules, and all the rotatable bonds were set as non-rotatable. Furthermore, these compounds were saved in PDBQT format as output files and used for docking.

Molecular docking and visualization

Molecular docking was carried out in AutoDock 1.5.7 using the previously prepared PDBQT files of protein and ligand compounds to calculate the interactions between the protein residues and ligands. A rigid blind docking was performed by applying the Lamarckian genetic algorithm while covering the whole protein inside a grid box. The grid box parameters of each of the proteins are shown in Table 2. The docking run number for each of the compounds was 10, and the run with the lowest free binding energy was selected as the best binding pose. The docking files of the protein and ligand complexes were further analyzed, and the 3D interactions between the target–ligand complex and binding site residues were visualized in Biovia Discovery Studio version 20.1.0. Finally, LigPlot + v.2.2 was used to generate the 2D interaction plot, where hydrogen bonds and hydrophobic interactions between the ligand–protein complex along with the names of the interacting residues were visualized.

Table 2.

Grid box parameters of the target proteins with their active site residues.

Protein	Grid box parameters
Protein	Size points (x × y × z)	Spacing center (Å)
1. HER2	76 × 78 × 104	1
2. ER Alpha	126 × 126 × 126	0.502
3. BCL2	126 × 126 × 96	0.586
4. BRCA1	126 × 126 × 112	0.375
5. BRCA2	126 × 126 × 126	0.375
6. PPARG	126 × 126 × 126	0.375

Pharmacological properties and toxicity prediction

The prediction of ADMET properties is crucial for assessing the hazardous effects and efficacy of compounds inside the human body. After the completion of molecular docking, the compounds that showed good binding energy with each of the six proteins were further analyzed for their ADMET properties. These ligands were screened for their molecular properties via the web-based SwissADME tool that allows us to predict the physicochemical properties and drug-likeness based on the Lipinski rule of five (Molecular Weight (MW) ⩽ 500, Moriguchi octanol–water partition coefficient (MLogP) ⩽ 4.15, N or O ⩽ 10, NH or OH ⩽ 5).²⁴ Compounds violating two or more rules were excluded from further analysis. The ProTox 3.0 web tool was also used to predict the safety and toxicity of the compounds.

Lead compound selection

The binding energies of the protein–ligand complexes were evaluated, and on the basis of these binding scores, the list was narrowed down to prioritize lead compounds with high affinity for the target proteins. The compounds with the best binding affinity for all six proteins were further filtered out by analyzing their pharmacological properties, which eventually yielded potential lead compounds.

Statistical analysis

This study utilized statistical analysis to identify patterns, test hypotheses, draw conclusions, and determine the strength of correlations between variables, making it crucial for effective data analysis. Initially, Microsoft Excel served basic functions, but as the analysis grew, specialized software GraphPad Prism was used for advanced statistical analyses and customizable graph generation. Prior to statistical analysis, discrete data from docking scores were generated to characterize the central tendencies and variability, which included the mean, range, variance, and standard deviation. Multiple statistical tests have been employed to determine whether microbial bioactives outperform phytochemicals in anti-cancer therapeutic effectiveness. Specifically, the unpaired Student’s t-test and one-way analysis of variance (ANOVA) were used to compare the averages among three independent groups (microbial bioactives, phytochemical, and drug), whereas nonparametric tests were applied when the data deviated from the assumptions of a normal distribution. Finally, statistical significance was assessed when p-values were less than the alpha level of 0.05 and the confidence interval exceeded zero. Throughout the research, the confidence level averaged 95%. These statistical methods enable the evaluation of the strength of relationships and the derivation of meaningful conclusions from the data.

MD simulations

MD simulations of BRCA1–Marfuraquinocin D and ERα–Marfuraquinocin D complexes were performed using Gromacs version 2023.3. Preliminary protein–ligand structures for the simulations were derived directly from the molecular docking results.²⁵ Protein topologies were generated using the CHARMM27 force field with the TIP3P water model, while ligand parameters were derived from the SwissParam server, utilizing the Merck Molecular force field that is compatible with CHARMM (Chemistry at HARvard Molecular Mechanics).²⁶

Each complex was encapsulated in a rhombic dodecahedral box, with a distance of 1.2 nm and solvated with SPC216 type of water molecules.²⁵ The system was neutralized by adding water molecules and ions (Na⁺, Cl⁻), respectively, ensuring that the environment was both physiologically relevant and electrostatically stable. Through energy minimization, steric clashes were resolved and the system’s potential energy was optimized, while long-range electrostatic interactions were calculated. In addition, the x, y, and z directions were applied under periodic boundary conditions using the Particle Mesh Ewald (PME) approach.²⁷

Equilibration was carried out in two phases: 100 ps NVT at 323K using the V-rescale thermostat, followed by 100 ps NPT at 1 bar using the Parrinello–Rahman barostat. The LINCS (LINear Constraint Solver) algorithm was used to fix the hydrogen-involving bonds, while the van der Waals interactions were evaluated using a cutoff distance of 1.2 nm.^25
–27

After equilibration, the position constraints were removed, and a 500,000-step production simulation of 1 ns time cycle was performed. Post-MD simulation analyses, including fluctuations (root mean square fluctuation (RMSF)), deviations (root mean square deviation (RMSD)), compactness (radius of gyration (Rg)), hydrogen bonding, solvent exposure (solvent-accessible surface area (SASA)), solvation energy (ΔGsolv), residue surface area (RESAREA), and system volume, were performed using a specific set of GROMACS commands and visualized with QTgrace software.²⁷

Results

Molecular docking analysis

The main goal of this study was to calculate the binding affinity of microbe-based bioactive compounds for common breast cancer target proteins. Eighty bacterial compounds available on PubChem were docked against six common targets of breast cancer, and the binding energies of these docked compounds are depicted in a heat map (see Supplementary File 1). For a comparative evaluation of the binding affinity of microbe-based compounds, six control compounds were included and subjected to each of the six target proteins. The control compounds selected for HER2, ERα, BCL2, BRCA1, BRCA2, and PPARG were ZINC43069427,¹⁹ XAN (Xanthotoxol),²⁰ Theaflavin,²¹ Sclareol,⁶ α-Hederin,⁶ and Kaempferol,²² with binding energies of −11.0, −13.5, −8.3, −9.8, −11.0, and −16.7 kcal/mol, respectively.

Against the HER2 and BRCA1 proteins, Tylosin (CID 5280440) had the maximum binding affinity with binding energy values of −16.0 kcal/mol and −15.3 kcal/mol, respectively. Gutingimycin (CID 136835719), a marine streptomycete metabolite, had the best inhibitory effect on both the ERα (−14.3 kcal/mol) and BCL2 (−15.8 kcal/mol) proteins. Streptomycin showed the highest docking score (−18.8 kcal/mol) with BRCA2, while Rapamycin had the maximum binding affinity (−18.9 kcal/mol) for the PPARG protein. Several compounds had binding energy values lower than those of the controls; as the study is multitarget-based, it focused on the compounds that showed greater binding affinity with all target proteins, as shown in Figure 1.

Figure 1.

Heat map of selected 18 ligands with six controls highlighting the binding energy against the six target proteins. The y-axis contains the names of the ligand compounds, and proteins are included on the x-axis. A lighter color indicates better binding affinity.

A total of 18 compounds out of the 80 compounds had the lowest binding energy values with each of the proteins. Some of the compounds showed greater affinity than almost all the control compounds. For example, Lobophorin F (CID 139584897) had a lower binding energy value than 5 of the control compounds. All 18 compounds had lower binding energies and hence greater binding affinities than Theaflavin (−8.3 kcal/mol). Further pharmacological evaluation was performed to determine whether these compounds fit into the Lipinski rule of five.

Comparative evaluation with phytochemicals

Following an extensive literature search, a list of existing phytochemicals and drugs along with their binding energies was compiled for comparison with microbial bioactive compounds. In Figure 2, a set of vertical bar charts compares the binding energy of the microbial bioactives (black) to phytochemicals (light green) and available therapeutic drugs (dark green) against six proteins associated with breast cancer: HER2, ERα, BCL-2, BRCA1, BRCA2, and PPARG. The investigation included various sample sizes for six proteins, consistently involving microbial bioactives, while the numbers for phytochemicals and drugs varied.

Figure 2.

Vertical bar graphs compare the binding energy of microbial bioactives (black), phytochemicals (light green), and drugs (dark green) to HER2, ER Alpha, BCL2, BRCA1, BRCA2, and PPARG. The value for the binding energy is shown on the vertical axis, and higher negative values in each case denote stronger binding. Each graph contains the sample sizes (n) for each group and reports statistical significance as follows: **** (p < 0.0001), *** (p < 0.001), ** (p < 0.01), * (p < 0.05), and ns (not significant).

On HER2 binding energy comparison, it is seen that there is no significant difference (ns) among the three categories, indicating similar affinities. However, ERα analysis revealed a significant difference (**), showing that microbial bioactives bind stronger than phytochemicals, with no marked difference (ns) existing with drugs. BCL2 results showed a highly significant difference (***), related outcomes that favored microbial bioactives over phytochemicals and showed no significant disparity (ns) concerning drugs. For BRCA1 and BRCA2, the data indicated the highest binding affinities by microbial bioactives, followed by phytochemicals and then drugs, with very highly significant differences (****) between microbial bioactives and phytochemicals, and very significant differences (**) between microbial bioactives and drugs. In the PPARG context, a very highly significant difference (****) was also observed with no significant difference (ns) over drugs, showing a stronger binding affinity of microbial bioactives over phytochemicals. In contrast, all the six proteins did not show any significant difference between phytochemicals and drugs; hence, the affinities across this category are similar.

In all, microbial bioactives with higher or comparable binding affinity with phytochemicals and drugs for all proteins under study have been identified. Such consistency of results may further support microbial bioactives as promising therapeutic candidates in the management of breast cancer based on their high inhibitory potential against key proteins.

ADMET prediction

Eighteen compounds were selected from 80 microbial bioactive compounds on the basis of their binding affinity with the target protein. For further screening, the pharmacological properties and toxicities of these 18 compounds were studied. All the pharmacological properties of these compounds are mentioned in Supplementary File 2. To determine the drug-likeness of the compounds that have a relatively high binding affinity with all six target proteins, the Lipinski rule of five was used as a standard. Drug-likeness parameters such as molecular weight, number of hydrogen bond acceptors (HBAs), number of hydrogen bond donors (HBDs), MLogP, and gastrointestinal absorption (GIA) of the 18 selected compounds are given in Table 3. HBA and HBD are fundamental parameters for indicating the oral bioavailability, mainly solubility and permeability, of a compound. An increased number of HBA indicates good solubility of a compound; however, too much HBA can also affect the permeability of the compound,^28,29 so HBA not more than 10 is considered the ideal value for both good solubility and permeability.²⁴ In addition to Marfuraquinocin A (CID 139584225) and Marfuraquinocin D (CID 73213688), all the compounds have HBA values greater than 10, which portray the poor permeability of these compounds.

Table 3.

Drug-likeness and bioavailability of the selected compounds.

Compound name	Molecular weight	Number of HBA	Number of HBD	MLogP	GI absorption	Lipinski violations
Lobophorin F	1027.2	17	5	0.72	Low	2
Marfuraquinocin A	438.56	5	1	2.84	High	0
Marfuraquinocin D	440.53	6	2	1.83	High	0
Dapiramicin A	511.48	13	7	–3.46	Low	3
Dapiramicin B	527.48	14	8	–4.2	Low	3
Komodoquinone A	529.54	11	6	–1.27	Low	3
Neoantimycin F	684.73	12	4	1.76	Low	2
Daunomycin	527.52	11	5	–1.35	Low	2
Doxorubicin	543.52	12	6	–2.1	Low	3
Rapamycin	914.17	13	3	1	Low	2
IB-00208	658.65	12	1	0.52	Low	2
Nonactin	736.93	12	0	2.73	Low	2
Chartreusin	640.59	14	5	–0.42	Low	2
Gutingimycin	1027.98	23	8	–3.35	Low	3
Himalomycins	828.9	16	6	–0.81	Low	3
Tylosin	916.1	18	5	–1.99	Low	2
Buanmycin	593.53	12	7	–0.54	Low	3
Rifampicin	822.94	14	6	0.12	Low	3

Like HBA, a high number of HBDs can also hamper the permeability of a certain compound. MLogP is a modified version of LogP that takes molecular size into account when estimating the lipophilicity of a compound.³⁰ Compounds with a LogP value of less than 5 are considered ideal for drug development. Surprisingly, all the compounds had an MLogP < 5; however, compounds with negative LogP values are considered less lipophilic.²⁴ For example, Dapiramicin B (CID 5487882) has the most negative MLogP value out of the 18 compounds and is thus less likely to cross the lipid bilayer. Marfuraquinocin D and Neoantimycin F (CID 73350712) have shown ideal MLogP values (1.35–1.8), which makes them good for oral and intestinal absorption. In addition, GIA and brain access to the compounds were depicted in the boiled egg model.³¹ Only two compounds, Marfuraquinocin A and Marfuraquinocin D, lie inside the white ellipse of the plot, which means that they are well absorbed in the intestine. However, these two compounds are PGP+ (P-glycoprotein), so they are predicted to be refluxed from the central nervous system (CNS) to the bloodstream or from the GI tract to the GI lumen, reducing their availability in the blood (shown in Supplementary File 1).

The bioavailability radar was generated on the basis of the following properties of the very bioactive compound: size, flexibility, solubility, saturation, and lipophilicity (see Supplementary File 1). A compound with two or more violations of the Lipinski rule was excluded from the study. In addition to their pharmacological properties, the toxicities of the 18 compounds were also analyzed. The toxicity profiles of these compounds are listed in Table 4. Blood–brain barrier (BBB) permeability, hepatotoxicity, carcinogenicity, and immunotoxicity were used as parameters to evaluate the toxicity of the compounds. Among the 18 compounds shown in Table 4, Dapiramicin A, Dapiramicin B, Daunomycin, Doxorubicin, Gutingimycin, and Himalomycins were toxicity classes II and III, indicating their harmfulness if swallowed. Lobophorin F, which had the best binding affinity for almost every target protein, and Rapamycin, which had the highest binding affinity for the PPARG protein (from Figure 1), did not result in satisfactory carcinogenicity. Other than that, no compounds showed any toxicity to the liver except Rifampicin (CID 135398735). However, all the compounds, except Nonactin (CID 72519), showed active immunotoxicity. Out of the 18 compounds, only two compounds, Marfuraquinocin A and Marfuraquinocin D, met the drug-likeness criteria with no Lipinski rule violations and exhibited high bioavailability scores (⩾ 0.55), indicating a high GIA rate (shown in Table 3 and Supplementary File 2).

Table 4.

Toxicity profiles.

Ligands	Toxicity class	LD50 (mg/kg)	BBB	Hepatotoxicity	Carcinogenicity	Immunotoxicity
Lobophorin F	4	560	Inactive	Inactive	Active	Active
Marfuraquinocin A	4	500	Active	Inactive	Inactive	Active
Marfuraquinocin D	4	500	Active	Inactive	Inactive	Active
Dapiramicin A	2	26	Inactive	Inactive	Inactive	Active
Dapiramicin B	2	26	Inactive	Inactive	Inactive	Active
Komodoquinone A	4	607	Inactive	Inactive	Inactive	Active
Neoantimycin F	4	1000	Inactive	Inactive	Inactive	Active
Daunomycin	3	205	Inactive	Inactive	Inactive	Active
Doxorubicin	3	205	Inactive	Inactive	Inactive	Active
Rapamycin	5	2500	Inactive	Inactive	Active	Active
IB-00208	4	600	Active	Inactive	Inactive	Active
Nonactin	5	2500	Active	Inactive	Inactive	Inactive
Chartreusin	5	4000	Inactive	Inactive	Inactive	Active
Gutingimycin	3	300	Inactive	Inactive	Inactive	Active
Himalomycins	2	7	Active	Inactive	Inactive	Active
Tylosin	5	5000	Inactive	Inactive	Inactive	Active
Buanmycin	4	1550	Inactive	Inactive	Inactive	Active
Rifampicin	4	500	Inactive	Active	Inactive	Active

Lead compound selection

Identifying lead compounds served as an important step for shortlisting the compounds with the most potential for the treatment of breast cancer. The primary selection of lead compounds was assessed by the binding energies of protein–ligand complexes. The binding scores were analyzed systematically to rank the compounds with high affinities toward six target proteins involved in breast cancer. Among these compounds, 18 displayed binding energies less than −9 kcal/mol at the initial level of molecular docking studies, indicating good interaction with the target proteins. Consequently, additional selection was performed according to its pharmacological properties as a drug by applying key drug-likeness parameters. The screening of the selected criteria was carried out via the Lipinski rule of five, which eliminates compounds that are less likely to be orally active drugs in humans.

After this rigorous selection process, Marfuraquinocin A and Marfuraquinocin D were the most promising compounds. These two single compounds, which indeed displayed sufficient binding affinities (from Figure 1) to all six target proteins, also emphasized pharmacologically favorable profiles (from Table 3) and, therefore, are very good leads that could be followed through for further development. The preliminary efficacy and signs of the safety profile suggest the possible use of these compounds as potential inhibitors of breast cancer.

Visualization and analysis of lead compound binding

The detailed 3D interaction analysis between Marfuraquinocin A and six various target proteins is presented in Figure 3. Panels (a)–(f) show the pattern of binding of Marfuraquinocin A with HER2 (a), ERα (b), BCL2 (c), BRCA1 (d), BRCA2 (e), and PPARG (f).

Figure 3.

Distance of ligand–receptor complex and 3D interaction of Marfuraquinocin A with the six target proteins. (a) 3D interaction of Marfuraquinocin A and HER2. (b) 3D interaction of Marfuraquinocin A and ER Alpha. (c) 3D interaction of Marfuraquinocin A and BCL2. (d) 3D interaction of Marfuraquinocin A and BRCA1. (e) 3D interaction of Marfuraquinocin A and BRCA2. (f) 3D interaction of Marfuraquinocin A and PPARG.

In each panel, attention was focused on molecular interactions between the ligand and the receptor, and all the molecular orientations were focused near the binding sites. The different binding modes of Marfuraquinocin A toward proteins suggest its potential multifaceted role as a therapeutic agent for targeting different proteins related to breast cancer. Similarly, the Interactome3D of the molecule Marfuraquinocin D with six different target proteins set out in six panels (a)–(f) are shown in Figure 4. Visualizations of the active sites within these proteins revealed that the ligand bound to them interacted with the protein through hydrogen bonds and hydrophobic contacts. The view shows the binding sites and the 3D arrangement of the ligand within those sites.

Figure 4.

Distance of ligand–receptor complex and 3D interaction of Marfuraquinocin D with the six target proteins. (a) 3D interaction of Marfuraquinocin D and HER2. (b) 3D interaction of Marfuraquinocin D and ER Alpha. (c) 3D interaction of Marfuraquinocin D and BCL2. (d) 3D interaction of Marfuraquinocin D and BRCA1. (e) 3D interaction of Marfuraquinocin D and BRCA2. (f) 3D interaction of Marfuraquinocin D and PPARG.

Structural analysis of the protein–ligand complexes revealed the intricate binding interactions of Marfuraquinocin A and Marfuraquinocin D with six key breast cancer target proteins, including hydrogen bonding and hydrophobic contacts. The results (see Table 5 and Supplementary File 1) indicated that Marfuraquinocin A interacted with Asp863 residues with a calculated bond length of 2.87 Å, whereas Marfuraquinocin D could interact with Lys753, Asp863, Ser783, and Thr798 residues with calculated bond lengths of 3.30, 2.49, 3.13, and 2.95 Å, respectively, in the HER2 binding pocket. Strongly visible conventional hydrogen bonds among these residues were very well noticed.

Table 5.

Predicted active site residues involved in interactions of target proteins.

Protein	Ligand	Binding energy (kcal/mol)	Inhibition constant (nM)	Ligand efficiency	Residue involved
Protein	Ligand	Binding energy (kcal/mol)	Inhibition constant (nM)	Ligand efficiency	H-bond	Hydrophobic
HER2	Marfuraquinocin A	–9.8	63.37	–0.31	Asp863 (2.87 Å)	Leu796, Lys753, Val734, Thr798, Thr862, Leu785, Glu770, Ser783, Arg784, Met774, Ala771
HER2	Marfuraquinocin D	–9.9	55.7	–0.31	Lys753 (3.30 Å), Asp863 (2.49 Å), Ser783 (3.13 Å), Thr798 (2.95 Å)	Met774, Leu796, Glu770, Ala771, Leu785, Thr862, Val734
ER Alpha	Marfuraquinocin A	–10.4	21.29	–0.33	-	Met343, Leu346, Leu349, Ala350, Arg394, Glu353, Leu391, Phe404, His524, Leu525, Gly521, Gly420, Ile424
ER Alpha	Marfuraquinocin D	–11.1	6.67	–0.35	Pro324 (3.21 Å), Pro325 (2.89 Å)	Met357, Glu353, Ile386, Leu387, Lys449, Gly390, Glu323, Trp393, Ile326, Phe445
BCL2	Marfuraquinocin A	–9.1	193.75	–0.29	Arg124 (3.10 Å), Trp173 (2.83 Å)	Leu172, Asn169, Pro120, Phe121
BCL2	Marfuraquinocin D	–10.4	20.83	–0.33	Asn169 (2.34 Å)	Leu166, Arg124, Trp173, Pro120, Pro165, Phe121, Ala123
BRCA1	Marfuraquinocin A	–10.0	41.72	–0.31	Lys1702 (3.23 Å) Leu1701 (2.84 Å, 3.05 Å)	Ser1655, Gly1656, Val1654, Leu1657, Thr1700, Ser404, Arg405, Pro407
BRCA1	Marfuraquinocin D	–10.9	9.22	–0.34	Gly1656 (2.86 Å), Leu1657 (3.08 Å)	Phe1662, Thr1658, Pro1659, Val1654, Ser1655, Thr1700, Pro407, Arg405
BRCA2	Marfuraquinocin A	–15.0	0.009	–0.47	Leu931 (2.72 Å), Ser873 (2.67 Å), Tyr929 (3.05 Å), Val925 (3.05 Å)	Val923, Val932, Cys933, Ile922, Pro924, Met875, Ala874, Gly1166, Asp927, Pro926, Val928
BRCA2	Marfuraquinocin D	–15.3	0.00522	–0.48	Asp927 (2.94 Å), Val928 (2.60 Å), Val925 (2.44 Å), Ser873 (3.20 Å), Ala874 (2.89 Å)	Pro926, Tyr929, Gly1166, Phe876, Met875, Val932, Pro924, Val923, Leu931, Cys933, Ile922
PPARG	Marfuraquinocin A	–11	8.72	–0.34	-	Leu330, Leu333, Ile341, Glu343, Arg288, Ser342, Ile281, Cys285, Leu353, Leu340, Met364, Val339
PPARG	Marfuraquinocin D	–11.8	2.19	–0.37	Asp475 (2.97 Å, 3.11 Å)	Lys319, Lys474, Val450, Thr447, Val446, Tyr320, Tyr473, Ile472

Moving to ERα, Marfuraquinocin A forms hydrophobic interactions with a variety of residues, such as Met343, Leu346, Leu349, Ala350, Arg394, Glu353, Leu391, Phe404, His524, Leu525, Gly521, Gly420, and Ile424. Interestingly, it does not form any traditional hydrogen bonds with ERα. Marfuraquinocin D forms hydrogen bonds with the Pro324 and Pro325 amino acids, with bond lengths of 3.21 Å and 2.89 Å, respectively. In the context of BCL2, Marfuraquinocin A forms hydrogen bonds with Arg124 at a distance of 3.10 Å and with Trp173 at a distance of 2.83 Å, demonstrating its specific binding interactions with this protein (Table 6). On the contrary, Marfuraquinocin D creates a hydrogen bond with Asn169 at 2.34 Å inside the BCL2 binding pocket, highlighting its unique method of interaction with this target.

Table 6.

Binding profiles of representative compounds of different classes.

Protein target (PDB ID)	Lead scaffold	Binding energy (kcal/mol)	Key residues (H-bond/hydrophobic)	MD stability (RMSD)	ADMET highlights	Reference
HER2	Lapatinib (Drug)	–7.65	H-bond: Lys736, Pro802; Hydrophobic: Thr798, Leu852, Val734, Gly804	Stable (0.5 nm)	FDA-approved, good oral bioavailability	Verma et al.³²
HER2	ZINC43069427	–11.0	H-bond: Asp80, Hydrophobic: Thr798, Val734, Thr862, Leu852; Pi-alkyl: Ala751, Leu726, Leu796; Pi-sigma: Val734	0.32 nm	HIA: 92.48%, Caco-2: 1.024 No AMES/hepatotoxicity	Sohrab et al.¹⁹
ER Alpha	Lomitapide (Drug)	–13.1	H-bond: His524	2.5 nm	FDA-approved Anti-cancer effects reported	Düzgün and Korkmaz³³
ER Alpha	Luteoxanthin (Phytochemical)	–10.9	H-bond: Glu419; Hydrophobic: Leu391, Leu387, Ala350, Leu384, Pro324	0.25-0.29 nm (~2.5-2.9 Å)	GI absorption 94.6%; Caco-2 permeable; Lipinski compliant; no toxicity	Robin et al.³⁴
BCL2	Saquayamycin F (Phytochemical)	−12.0	H-bond: ARG109, ARG146, GLU136, TYR108, ASN143, TRP144	~2.0 Å (0.2 nm)	Lipinski compliant, low toxicity	Almansour et al.³⁵
BCL2	Venetoclax (Drug)	–10.6	H-bond: Gly101, Asp103; Hydrophobic: Val141	~3.1 Å (0.31 nm)	FDA-approved, optimal pharmacokinetics	³⁵
BRCA1	ZINC000070454071	–11.6	H-bond: Arg1762, Ile1760, Glu1849, Cys1847	1.66 Å	Zero Lipinski violation, can cross the blood–brain barrier, least toxicity, absence of mutagenicity, and carcinogenicity	Singh et al.³⁶
BRCA2	Narirutin	–9.2	-	2.45 Å	Narirutin is unlikely to penetrate the BBB due to its size and polarity, which may reduce neurotoxicity risk.	Turini et al.³⁷
PPARG	Resolvin E1	−9.9	Hie449 (H11) Tyr327 Glu291	3.5 Ǻ	-	Muralikumar et al.³⁸

In the case of BRCA1, Marfuraquinocin A creates connections with Lys1702 at 3.23 Å, Leu1701 at 2.84 Å, and Leu1701 at 3.05 Å through hydrogen bonds. Similarly, Marfuraquinocin D interacts with BRCA1 through hydrogen bonds with Gly1656 at 2.86 Å and Leu1657 at 3.08 Å, which emphasizes its specificity and affinity for this crucial protein associated with breast cancer. For BRCA2, Marfuraquinocin A forms hydrogen bonds with Leu931 at 2.72 Å, Ser873 at 2.67 Å, Tyr929 at 3.05 Å, and Val925 at 3.05 Å. These findings indicate that it can bind to the protein through multiple binding modes. Marfuraquinocin D complements this by forming hydrogen bonds with Asp927 at 2.94 Å, Val928 at 2.60 Å, Val925 at 2.44 Å, Ser873 at 3.20 Å, and Ala874 at 2.89 Å in the BRCA2 binding site, indicating its strong binding affinity to the protein. With respect to PPARG, Marfuraquinocin A engages in hydrophobic interactions with various residues, including Leu330, Leu333, Ile341, Glu343, Arg288, Ser342, Ile281, Cys285, Leu353, Leu340, Met364, and Val339. Notably, it does not form conventional hydrogen bonds with PPARG. Marfuraquinocin D, however, forms hydrogen bonds with Asp475 at bond lengths of 2.97 Å and 3.11 Å, in addition to other binding interactions, highlighting its diverse and complex binding mechanisms with this protein target.

Moreover, the inhibition constant (Ki) and ligand efficiency (LE) values strongly indicate the efficacy of these compounds. The inhibition constant (Ki) is directly proportional to the binding energy and reflects the drug concentration required for 50% inhibition.³⁹ A lower Ki value means that a smaller amount of the drug is required for achieving inhibition. For Marfuraquinocin A, the Ki values ranged from 63.37 nM for HER2 to 0.009 nM for BRCA2, indicating a broad spectrum of strong inhibition potency across different targets (from Supplementary File 1). Other Ki values for Marfuraquinocin A include 21.29 nM for ERα, 193.75 nM for BCL2, 41.72 nM for BRCA1, and 8.72 nM for PPARG. Similarly, Marfuraquinocin D exhibits Ki values ranging from 55.7 nM in HER2 to 0.00522 nM for BRCA2, which correlates with its high binding affinity. Other Ki values for Marfuraquinocin D include 6.67 nM for ERα, 20.83 nM for BCL2, 9.22 nM in BRCA1, and 2.19 nM in PPARG. The LE values, which are indicative of the binding energy per atom, show that Marfuraquinocin A and Marfuraquinocin D are not only effective binders but also efficient in their interactions. For Marfuraquinocin A, the LE values are −0.31 with HER2, −0.33 for ERα, −0.29 for BCL2, −0.31 for BRCA1, −0.47 for BRCA2, and finally −0.34 for PPARG. While for Marfuraquinocin D, the LE values are −0.31 for HER2, −0.35 with ERα, −0.33 with BCL2, −0.34 with BRCA1, −0.48 with BRCA2, and finally −0.37 with PPARG. These values underline the fact that both compounds maintain high efficiency in binding across multiple protein targets, making them very promising candidates for further development as inhibitors of breast cancer.

MD simulation analysis

To assess the dynamic stability and conformational flexibility of Marfuraquinocin D in complex with BRCA1 and ERα proteins, 1000 ps MD simulations were performed. The simulation trajectories were evaluated using key structural and energetic parameters, including RMSD, RMSF, Rg, SASA, hydrogen bonding, RESAREA, system volume and density, and free energy of solvation (ΔGsolv).

RMSD and RMSF analysis

The RMSD profile (Figure 5) demonstrated that both protein–ligand complexes equilibrated early in the simulation. The BRCA1–Marfuraquinocin D complex exhibited minimal structural deviation, stabilizing between 0.09 and 0.13 nm for the remainder of the trajectory. The ERα complex displayed slightly greater deviation, plateauing at approximately 0.15 and 0.2 nm, peaking around 0.45 ns, indicating higher conformational adaptability. These stable RMSD trends suggest that both complexes retained their structural integrity throughout the simulation.

Figure 5.

The RMSD plot (left) of the promising drug candidates Marfuraquinocin D with BRCA1 (green highlight) and ERα (red highlight) over 1 ns time trajectory. The RMSF plot (right) of the promising drug candidates Marfuraquinocin D with BRCA1 (black highlight) and ERα (red highlight).

The RMSF plot (Figure 5) indicates that ERα exhibits greater atomic fluctuations compared to BRCA1, reflecting higher structural flexibility throughout the trajectory. ERα shows prominent peaks around atom indices ~500, ~1800, ~2500, ~3400, and especially near 4000, where the fluctuation exceeds 0.55 nm, suggesting highly flexible regions, likely at the termini or loop areas. In contrast, BRCA1 maintains lower and more stable fluctuations, generally below 0.15 nm, with minor peaks near 1300, 2500, and 2700, where the RMSF slightly rises above 0.2 nm, showing localized but limited flexibility. Overall, the data suggest ERα is more dynamic, while BRCA1 retains a more rigid and stable conformation.

Rg, SASA, and hydrogen bond profile

In Figure 6, the Rg values indicated a relatively compact structure for both complexes. The BRCA1–Marfuraquinocin D complex maintained Rg values between 1.88 and 1.93 nm with minor fluctuations, suggesting a consistently folded and compact state. In contrast, the ERα complex had a slightly expanded structure with an Rg fluctuating around 1.86–1.93 nm, indicative of increased flexibility or transient unfolding events.

Figure 6.

(a) The plot of the radius of gyration (Rg). (b) The SASA curves during MD simulation. (c) The pattern of H-bonding observed during MD simulation.

The SASA analysis in Figure 6 revealed a lower solvent-exposed surface area in the BRCA1–Marfuraquinocin D complex, with values ranging from approximately 118 to 125 nm², compared to 140 to 148 nm² in ERα, supporting the notion of a more tightly packed structure in BRCA1. Both systems had moderate SASA fluctuations, implying minimal global conformational rearrangement over the simulation time.

Hydrogen bonding analysis (Figure 6) showed that the ERα–Marfuraquinocin D complex consistently formed two to five hydrogen bonds throughout the simulation, indicating strong and relatively stable interactions. In contrast, the BRCA1–Marfuraquinocin D complex exhibited one to three hydrogen bonds, suggesting moderately stable binding with more transient interactions. In addition, hydrogen bond pairs within 0.35 nm further supported these findings, showing greater and more sustained interactions for ERα compared to BRCA1.

Volume, density, and ΔGsolv

The ΔGsolv analysis (Figure 7) demonstrated stronger solvation stability for BRCA1, with values ranging between −10 and −30 kJ/mol. There are noticeable dips, particularly around 380 ps and 850 ps, reaching values close to −28 kJ/mol, suggesting it has a stronger interaction with the solvent and is more thermodynamically stable in the solvent environment. In contrast, the ERα complex showed less favorable ΔGsolv values fluctuating from 5 to −15 kJ/mol, indicating moderate solvation and stability in the solvent environment. However, this indicates less solvated or more hydrophobic compared to BRCA1.

Figure 7.

The free energy of solvation (left) and volume and density (right) plots curves during the MD simulation.

The volume (Figure 7) for both BRCA1 and ERα remains very stable throughout the simulation, staying close to 45 nm³. BRCA1’s density (orange line) in Figure 7 fluctuates slightly above 900 g/L, while ERα’s density (black line) remains slightly lower, just under 900 g/L. These results suggest that while both proteins are structurally stable, BRCA1 is marginally more compact than ERα.

RESAREA profile

The “Area per residue over the trajectory” plot (Figure 8) compares the SASA of ERα and BRCA1 proteins. For ERα, the average SASA per residue (top left, green) ranges up to 2.5nm² across residues ~300–550. The corresponding standard deviation (bottom left, blue) mostly stays below 0.25nm², with some spikes reaching close to 0.45nm². In contrast, BRCA1 (top right, black) shows an average SASA per residue up to 3.8 nm² across residues ~1650–18,500. Its standard deviation (bottom right, red) also remains mostly below 0.2nm², with a few regions peaking around 0.35nm². Overall, ERα appears to be structurally more flexible and surface-exposed, whereas BRCA1 maintains a more compact and stable conformation.

Figure 8.

The area per residue over the trajectory plot.

Biological properties of Marfuraquinocin A and Marfuraquinocin D

These two secondary metabolites (Figure 9) are a subclass of sesquiterpenoids generated by the marine bacteria Streptomyces niveus SCSIO 3406. These compounds were shown to have cytotoxic effects on a panel of human cancer cell lines, including the human glioblastoma cell line SF 268, the human breast adenocarcinoma cell line MCF-7, the human lung cancer cell line NCI-H460, and the human hepatocarcinoma cell line HepG2.^40
–42

Figure 9.

2D structures of Marfuraquinocin A (left) and Marfuraquinocin D (right).

Furthermore, their adherence to drug-likeness criteria indicated that they have characteristics similar to existing treatments. They specifically met the Lipinski rule, indicating good physicochemical properties for absorption, distribution, metabolism, and excretion. Likewise, bioavailability scores surpassed the threshold of ⩾0.55, which showed high GIA. In addition, high bioavailability is dangerous for cancer therapy because it enables effective delivery to tumor sites and desired effects.⁴³

These features improve their chances of successfully transitioning from preclinical research to clinical trials. However, knowledge regarding their biosynthetic processes and features has been limited to a few studies; so, more research is needed to examine these compounds and better understand their modes of action, pharmacokinetics, and efficacy in cancer models.

Discussion

Breast cancer remains one of the most prevalent cancers worldwide, accounting for a significant proportion of cancer-related deaths among women. The complexity of breast cancer, along with resistance to standard therapies, makes it challenging to treat.^44,45 Despite significant advances in breast cancer research, the underlying mechanisms driving its progression have not been fully elucidated to date. In addition, because of the side effects and poor clinical efficacy of conventional anti-cancer treatments, it is necessary to discover effective anti-breast cancer drugs that can overcome resistance and improve patient outcomes. Over the last few decades, hundreds of natural bioactive molecules from plants, fungi, bacteria, and marine organisms have been approved as anti-cancer drugs.^46
–48 Regardless of the success of phytochemicals, pharmaceutical development and large-scale isolation remain major challenges for novel drug discovery.⁴⁹ Owing to their selective cytotoxicity and ease of production on a large scale, microbial bioactive compounds have become a scalable solution for drug development. Several primary and secondary metabolites from bacteria, such as actinomycin D, bleomycin, antimycins, and chartreusin have shown impressive anti-tumor activity and are currently in clinical and preclinical trials.^50,51 The unique mechanism of action and the great therapeutic potential of microbial metabolites can be key factors in shifting the focus from chemically synthesized molecules and phytochemicals to microbial bioactives as effective anti-cancer drugs.

The most common key marker proteins and signaling proteins involved in breast cancer were selected as targets for this study on the basis of a thorough literature search and the available crystal structures of proteins. The initiation of breast cancer depends on several cellular pathways, and ERα and HER2 receptors trigger the activation of these signaling pathways.^52,53 The overexpression of HER2 receptors is commonly associated with the progression of breast cancer. The HER2 receptor is crucial for the growth and development of normal cells, and alterations in this process lead to aggressive disease behavior; thus, inhibition of the activity of this receptor can improve the prognosis of breast cancer patients.^54,55 To promote tumor growth, evasion of apoptosis is a crucial feature of cancer cells, BCL2, an anti-apoptotic protein, plays a central role in keeping pro-apoptotic proteins in check and protecting cells from apoptotic death and promoting tumorigenesis.^56,57 BCL2 overexpression is dependent mainly on ER-positive tumors and serves as a prognostic marker.^58,59 The BRCA1 and BRCA2 genes are involved in DNA repair, and aggressive forms of breast cancer are strongly associated with mutations in these genes.^60
–62 In the present study, microbe-derived metabolites were explored for their potential for multitarget inhibition of breast cancer markers and signaling proteins by molecular docking, ADMET analysis, and statistical analysis, which revealed that the bioactivity of microbial bioactives can differ significantly from that of phytochemicals.

Molecular docking has emerged as an effective tool for structure-based drug design. Computational docking is often used to visualize the interaction between ligand–receptor complexes and calculate their binding affinity.^63,64 The primary objective of this study was to evaluate the binding affinity of microbial bioactives against breast cancer target proteins and compare them with known phytochemicals. On the basis of their established efficacy against target proteins, the control compounds ZINC43069427, Xanthotoxol, Theaflavin, Sclareol, α-Hederin, and Kaempferol were selected against HER2, ERα, BCL2, BRCA1, BRCA2, and PPARG, respectively. Several bacterial bioactive compounds, such as Lobophorin F, Tylosin (HER2, −16.0 kcal/mol; BRCA1, −15.3 kcal/mol), Gutingimycin (ERα, −14.3 kcal/mol; BCL2, −15.8 kcal/mol), and Rapamycin (PPARG, −18.9 kcal/mol) demonstrated significantly higher binding affinities than most controls did. Because of the heterogeneity of breast cancer, screening compounds on the basis of their multitarget potential is advantageous. A total of 18 compounds were filtered out based on their high binding affinity against multiple target proteins. However, it is difficult to predict the safety of a potential drug without absorption, toxicity, and excretion-based screening. Thus, a combination of molecular docking and ADMET properties can play a significant role in the development of new drugs. The ADMET analysis of the 18 compounds revealed the bioavailability, pharmacological properties, and toxicity profiles of the compounds. Only two compounds, Marfuraquinocin A and Marfuraquinocin D, met all the criteria for drug-likeness and exhibited acceptable safety profiles. These compounds did not violate the Lipinski rule, had good MLogP values, and had an ideal bioavailability score for good absorption. Compared with Marfuraquinocin A and Marfuraquinocin D, compounds such as Lobophorin F and Rapamycin had better binding affinities for target proteins but presented multiorgan toxicity, poor absorbance, and violations of the Lipinski rule. Therefore, Marfuraquinocin A and Marfuraquinocin D were chosen as the lead compounds of the study and the most promising candidates for further analysis.

Both Marfuraquinocin A and Marfuraquinocin D have the highest affinities for the BRCA2 protein, with binding energies of −15.0 and −15.3 kcal/mol, respectively (shown in Figure 1). Compared with control phytochemicals, these two compounds have significantly greater binding affinities against BCL2, BRCA1, and BRCA2. Structural analysis of the target protein and Marfuraquinocin complex revealed that Marfuraquinocin A forms hydrogen bonds with HER2, BCL2, BRCA1, and BRCA2 residues and that Marfuraquinocin D forms hydrogen bonds against all the target proteins. Marfuraquinocin A displayed strong hydrophobic interactions but no hydrogen bonds with ERα and PPARG residues. Marfuraquinocin D exhibited lower inhibition constant (Ki) values than Marfuraquinocin A, signifying more potent inhibition. The Marfuraquinocin (A and D) exhibited interaction motifs that mirrored established drug and phytochemical scaffolds. For instance, hydrophobic bond formation with Thr798 residue can be seen in Marfuraquinocin A, ZINC43069427, and Lapatinib. In ERα, Marfuraquinocins engaged the hydrophobic cluster comprising Leu387 and Leu391, analogous to the binding mode of the Luteoxanthin. With the lowest inhibition constant of 0.00522 nM, Marfuraquinocin D showed the strongest overall inhibition of the BRCA2 protein. Marfuraquinocin D also produced better results in terms of LE. In addition to HER2, Marfuraquinocin D has better LE against all the target proteins. For all the above-mentioned parameters, Marfuraquinocin D had a slight edge over Marfuraquinocin A. Among the lead compounds, Marfuraquinocin D was forwarded to MDs study as it showed better affinity against the target proteins, especially ERα and BRCA1. Post-MD analysis highlighted that the ligand–protein complex was indeed stable and densely packed for 1 ns.

In conclusion, this study demonstrated that bacterial bioactive compounds displayed greater binding affinity against multiple breast cancer targets than did phytochemical controls. However, the safety profile of these microbe-based compounds needs further evaluation. Overall, Marfuraquinocin A and Marfuraquinocin D displayed increased binding affinities, hydrogen bonds, and hydrophobic interactions with multiple residues of the breast cancer proteins and improved pharmacological properties. On the basis of the results of this study, we suggest that natural compounds such as Marfuraquinocin A and Marfuraquinocin D can be unique and effective drugs against breast cancer. However, additional research to confirm and expand on the current findings is desirable for the selected compounds. Therefore, further validation with in vitro and in vivo studies is necessary.

Conclusion

In summary, the results are promising and support the use of microbial bioactives as novel sources of natural anti-tumor agents for breast cancer based on quantitative computational evidence. A structure-based in silico docking procedure was used to screen bacterial compounds against the active sites of the six target proteins including HER2, ERα, BCL2, BRCA1, BRCA2, and PPARG.

Several ligands demonstrated binding energies below −9 kcal/mol, with the strongest affinities reaching up to −18.9 kcal/mol. Among them, Marfuraquinocin A and Marfuraquinocin D showed consistent interactions across all targets, with binding energies between −15.0 kcal/mol and −15.3 kcal/mol, respectively. The inhibition constants (Ki) were in the nanomolar range, with values as low as 0.009 nM for Marfuraquinocin A and 0.00522 nM for Marfuraquinocin D toward BRCA2. These results highlighted their strong potency. Both compounds also satisfied the Lipinski rule with zero violations and exhibited high predicted bioavailability scores (⩾0.55).

MD simulations against ERα (3ERT) and BRCA1 (4Y2G), along with molecular docking, and ADME predictions demonstrated that Marfuraquinocin A and Marfuraquinocin D possess intriguing drug-likeness and the best safety profiles. The BRCA1–Marfuraquinocin D complex remained highly stable, with RMSD values between 0.09 and 0.13 nm, while the ERα complex plateaued around 0.15–0.2 nm. Hydrogen bond analysis further supported stable interactions, with two to five bonds maintained in ERα and one to three bonds in BRCA1 throughout the simulation. The advanced statistical evaluation also emphasized the superior binding affinity of microbial bioactives. Stable hydrogen bonding patterns were observed throughout the simulation, involving two to five bonds in ERα and one to three bonds in BRCA1, while statistical analysis further validated the comparatively higher binding affinity of the microbial bioactives.

Compared to phytochemicals, they showed very highly significant differences (****p < 0.0001) for BRCA1, BRCA2, and PPARG, highly significant differences (***p < 0.001) for BCL2, and significant differences (**p < 0.01) for ERα. When compared to existing drugs, microbial actives showed no significant difference (ns) for HER-2 and PPARG. This multifaceted approach—integrating computational modeling, structural bioinformatics, and biostatistics—holds great promise for enhancing knowledge and discovering new breast cancer medicines based on secondary metabolites obtained from microbes. Nevertheless, further research is essential for fully comprehending the potential of these native bacterial substances, supporting the development of more successful yet less harmful cures for breast cancer patients.

Supplemental Material

sj-docx-1-chl-10.1177_17475198261448924 – Supplemental material for Screening of microbial bioactives as potential breast cancer inhibitors: A structure-based multitargeted molecular docking analysis

Supplemental material, sj-docx-1-chl-10.1177_17475198261448924 for Screening of microbial bioactives as potential breast cancer inhibitors: A structure-based multitargeted molecular docking analysis by Al Nahian Joy, Marua Khan, S. M. Shahriar Alam, Nasiha Shehreen, Wahidul Hasan, Ritika Amin Oishe, S. M. Bakhtiar Ul Islam, Kazi Md. Mostafizur Rahman, Nayeema Bulbul, Jinath Sultana Jime, Md. Asaduzzaman Shishir, Ashrafus Safa and Md. Fakruddin in Journal of Chemical Research

Footnotes

Acknowledgements

Not applicable.

ORCID iDs

Al Nahian Joy

Marua Khan

S. M. Shahriar Alam

Md. Fakruddin

Ethical consideration

This article does not contain any studies with animal participants.

Consent to participate

There are no human participants in this article, and informed consent is not required.

Consent for publication

Not applicable.

Author contributions

M.F., S.M.B.U.I., and M.A.S. contributed to conceptualization.

A. N.J., S.M.S.A., and M.K. contributed to writing—original draft.

S. M.B.U.I., N.B., J.S.J., and A.S. contributed to writing—review and editing.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

All the data and materials supporting the conclusion of this study are included within the article and supplementary files. The data used during the current study are available from the corresponding author upon reasonable request.

Supplemental material

Supplemental material for this article is available online.

References

Hong

Breast cancer: an up-to-date review and future perspectives. Cancer Commun 2022; 42: 913–936.

Anand

Dey

Chandel

AKS

, et al. Cancer chemotherapy and beyond: current status, drug candidates, associated risks and progress in targeted therapeutics. Genes Dis 2022; 10: 1367–1401.

Burguin

Diorio

Durocher

Breast cancer treatments: updates and new challenges. J Pers Med 2021; 11: 808.

Sadri

Heidari

Faridzadeh

, et al. Potential applications of macrophages in cancer immunotherapy. Biomed Pharmacother 2024; 178: 117161.

Kou

Lou

, et al. Nanoparticles in tumor microenvironment remodeling and cancer immunotherapy. J Hematol Oncol 2024; 17: 16.

Hossain

Ray

Sarkar

, et al. Natural compounds or their derivatives against breast cancer: a computational study. Biomed Res Int 2022; 2022: 5886269.

Gote

Nookala

Bolla

, et al. Drug resistance in metastatic breast cancer: tumor targeted nanomedicine to the rescue. Int J Mol Sci 2021; 22: 4673.

Yuan

Long

Zhou

, et al. PI3K–AKT-targeting breast cancer treatments: natural products and synthetic compounds. Biomolecules 2023; 13: 93.

Islam

Nafady

, et al. Natural small molecules in breast cancer treatment: understandings from a therapeutic viewpoint. Molecules 2022; 27: 2165.

10.

El-Beltagi

Mohamed

, et al. Phytochemical and potential properties of seaweeds and their recent applications: a review. Mar Drugs 2022; 20: 342.

11.

Shrihastini

Muthuramalingam

Adarshan

, et al. Plant derived bioactive compounds, their anti-cancer effects and in silico approaches as an alternative target treatment strategy for breast cancer: an updated overview. Cancers 2021; 13: 6222.

12.

Kumar

Nirmal

Kumar

, et al. Major phytochemicals: recent advances in health benefits and extraction method. Molecules 2023; 28: 887.

13.

Abdelghani

Hourani

Zaidan

, et al. Therapeutic applications and biological activities of bacterial bioactive extracts. Arch Microbiol 2021; 203: 4755–4776.

14.

Rani

Saini

Bast

, et al. Microorganisms: a potential source of bioactive molecules for antioxidant applications. Molecules 2021; 26: 1142.

15.

Zong

Ruan

Liu

, et al. Bacteria and bacterial components as natural bio-nanocarriers for drug and gene delivery systems in cancer therapy. Pharmaceutics 2023; 15: 2490.

16.

Hernández-Rodríguez

Rosales-Hernández

Mendieta-Wejebe

, et al. Current tools and methods in molecular dynamics (MD) simulations for drug design. Curr Med Chem 2016; 23: 3909–3924.

17.

Nair

Miners

JO.

Molecular dynamics simulations: from structure–function relationships to drug discovery. In Silico Pharmacol 2014; 2: 4.

18.

Ravichandran

Sundararajan

In silico-based virtual drug screening and molecular docking analysis of phytochemical derived compounds and FDA approved drugs against BRCA1 receptor. J Cancer Prev Curr Res; 8: 1–6.

19.

Sohrab

Kamal

MA.

Screening, docking, and molecular dynamics study of natural compounds as an anti-HER2 for the management of breast cancer. Life 2022; 12: 1729.

20.

Acharya

Chacko

Bose

, et al. Structure based multitargeted molecular docking analysis of selected furanocoumarins against breast cancer. Sci Rep 2019; 9: 15743.

21.

Taghizadeh

Niazi

Moghadam

, et al. Experimental, molecular docking and molecular dynamic studies of natural products targeting overexpressed receptors in breast cancer. PLoS ONE 2022; 17: e0267961.

22.

Basavarajappa

Rehman

Shiroorkar

, et al. Therapeutic effects of Crataegus monogyna inhibitors against breast cancer. Front Pharmacol 2023; 14: 1187079.

23.

Forli

Huey

Pique

, et al. Computational protein–ligand docking and virtual drug screening with the AutoDock suite. Nat Protoc 2016; 11: 905–919.

24.

Lipinski

Lombardo

Dominy

, et al. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv Drug Deliv Rev 2001; 46: 3–26.

25.

Ismail

Jusoh

SA.

Molecular docking and molecular dynamics simulation studies to predict flavonoid binding on the surface of DENV2 E protein. Interdiscip Sci Comput Life Sci 2016; 9: 499–511.

26.

Zoete

Cuendet

Grosdidier

, et al. SwissParam: a fast force field generation tool for small organic molecules. J Comput Chem 2011; 32: 2359–2368.

27.

Kushwaha

Maurya

Singh

, et al. Bulbine frutescens phytochemicals as novel ABC-transporter inhibitor: a molecular docking and molecular dynamics simulation study. J Cancer Metastasis Treat 2021; 7: 2.

28.

Coimbra

JTS

Feghali

Ribeiro

, et al. The importance of intramolecular hydrogen bonds on the translocation of the small drug piracetam through a lipid bilayer. RSC Adv 2021; 11: 899–908.

29.

Zeki

Mustafa

YF.

Digital alchemy: exploring the pharmacokinetic and toxicity profiles of selected coumarin-heterocycle hybrids. Results Chem 2024; 10: 101754.

30.

Kujawski

Popielarska

Myka

, et al. The log P parameter as a molecular descriptor in computer-aided drug design—an overview. Comput Methods Sci Technol 2012; 18: 81–88.

31.

Daina

Zoete

A BOILED-Egg to predict gastrointestinal absorption and brain penetration of small molecules. ChemMedChem 2016; 11: 1117–1121.

32.

Verma

Goyal

Kumari

, et al. Structural investigations on mechanism of lapatinib resistance caused by HER-2 mutants. PLoS ONE 2018; 13: e0190942.

33.

Düzgün

Korkmaz

FD.

Lomitapide as a potential estrogen receptor inhibitor: a computational drug repurposing study. Kocaeli Univ Saglik Bilim Derg 2024; 10: 8–14.

34.

Robin

Sarder

Ahmed

, et al. In-silico identification of phytochemical modulators targeting estrogen receptor alpha for endometrial carcinoma therapy. Int J Multidiscip Res 2025; 7: 1–19.

35.

Almansour

Allemailem

Abd

Aty

, et al. In silico mining of Natural Products Atlas (NPAtlas) database for identifying effective BCL-2 inhibitors: molecular docking, molecular dynamics, and pharmacokinetics characteristics. Molecules 2023; 28: 783.

36.

Singh

Sangwan

Chauhan

, et al. Screening and identification of phytochemical drug molecules against mutant BRCA1 receptor of breast cancer using computational approaches. Mol Cell Biochem 2022; 477: 885–896.

37.

Turini

Dunjic

Nejkovic

, et al. Comparative molecular docking analysis of natural and synthetic ligands targeting BRCA1, BRCA2, ER, and PR in breast cancer treatment. Asian Pac J Cancer Prev 2025; 26: 4349–4361.

38.

Muralikumar

Vetrivel

Narayanasamy

, et al. Probing intermolecular interactions of PPARγ-LBD with polyunsaturated fatty acids and their anti-inflammatory metabolites. Lipids Health Dis 2017; 16: 17.

39.

Muhammad

Saba

Khera

, et al. Virtual screening of potential inhibitor against breast cancer-causing estrogen receptor alpha (ERα): molecular docking and dynamic simulations. Mol Simul 2022; 48: 1163–1174.

40.

Song

Huang

Chen

, et al. Cytotoxic and antibacterial marfuraquinocins from deep South China Sea-derived Streptomyces niveus SCSIO 3406. J Nat Prod 2013; 76: 2263–2268.

41.

Saide

Lauritano

Ianora

A treasure of bioactive compounds from the deep sea. Biomedicines 2021; 9: 1556.

42.

Zhu

Cheng

Song

, et al. Complete genome sequence of the deep South China Sea-derived Streptomyces niveus SCSIO 3406. PLoS ONE 2021; 16: e0248404.

43.

El-Kattan

Varma

Oral absorption, intestinal metabolism and human oral bioavailability. InTech eBooks 2012; 22: 1–34.

44.

Yersal

Biological subtypes of breast cancer: prognostic and therapeutic implications. World J Clin Oncol 2014; 5: 412.

45.

Łukasiewicz

Czeczelewski

Forma

, et al. Breast cancer—epidemiology, risk factors, classification, prognostic markers, and current treatment strategies—an updated review. Cancers 2021; 13: 4287.

46.

Boopathy

Kathiresan

Anticancer drugs from marine flora: an overview. J Oncol 2010; 2010: 1–18.

47.

Asma

Acaroz

Imre

, et al. Natural products/bioactive compounds as a source of anticancer drugs. Cancers 2022; 14: 6203.

48.

Shakya

Naik

RR.

The chemotherapeutic potentials of compounds isolated from plant, marine, fungus, and microorganism. J Trop Med 2022; 2022: 1–17.

49.

Atanasov

Waltenberger

Pferschy-Wenzig

E-M

, et al. Discovery and resupply of pharmacologically active plant-derived natural products. Biotechnol Adv 2015; 33: 1582–1614.

50.

Mohan

Rangappa

Nayak

, et al. Bacteria as a treasure house of secondary metabolites with anticancer potential. Semin Cancer Biol 2021; 86: 998–1013.

51.

Yeh

C-T

C-L

Huang

C-YF

, et al. Preclinical evaluation of antimycin A as a potential anti-lung cancer stem cell agent. Evid Based Complement Alternat Med 2013; 2013: 1–13.

52.

Mehta

Tripathy

Co-targeting estrogen receptor and HER2 pathways in breast cancer. Breast 2013; 23: 2–9.

53.

Giuliano

Trivedi

Schiff

Bidirectional crosstalk between estrogen receptor and HER2 signaling pathways in breast cancer. Breast Care 2013; 8: 256–262.

54.

Yarden

Biology of HER2 and its importance in breast cancer. Oncology 2001; 61: 1–13.

55.

Tarantino

Morganti

Curigliano

Targeting HER2 in breast cancer: new drugs and paradigms on the horizon. Explor Target Antitumor Ther 2021; 23: 139–155.

56.

Merino

Lok

Visvader

, et al. Targeting BCL-2 to enhance vulnerability in estrogen receptor-positive breast cancer. Oncogene 2015; 35: 1877–1887.

57.

Kaloni

Diepstraten

Strasser

, et al. BCL-2 protein family: attractive targets for cancer therapy. Apoptosis 2022; 28: 20–38.

58.

Eom

Kim

Lee

, et al. BCL2 as a subtype-specific prognostic marker for breast cancer. J Breast Cancer 2016; 19: 252.

59.

Kawiak

Kostecka

Regulation of BCL-2 family proteins in estrogen receptor-positive breast cancer. Cancers 2022; 14: 279.

60.

Gorodetska

Kozeretska

Dubrovska

BRCA genes: role in genome stability and therapy resistance. J Cancer 2019; 10: 2109–2127.

61.

Narod

Salmena

BRCA1 and BRCA2 mutations and breast cancer. Discov Med 2011; 12: 445–453.

62.

Satyananda

Oshi

Endo

, et al. High BRCA2 gene expression and aggressive breast cancer. Ann Surg Oncol 2021; 28: 7356–7365.

63.

Morris

Huey

Lindstrom

, et al. AutoDock4 and AutoDockTools4: automated docking with selective receptor flexibility. J Comput Chem 2009; 30: 2785–2791.

64.

Syriopoulou

Markopoulos

Tzakos

, et al. Ligand–receptor interactions and drug design. Methods Mol Biol 2021; 24: 89–104.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

15.56 MB

0.00 MB