To investigate the antibacterial and antifungal properties of 04 selected molecules, molecular docking calculations were performed using Glide module [21] of the Schrödinger software suite (Schrödinger, LLC, New York, NY, 2023) using the extra precision (xp) mode [22]. The structure of compounds 1 - 4 were drawn using the 2D sketch panel of a Schrodinger and prepared using the Ligprep module (Schrödinger Release 2024-3, LigPrep, Schrödinger, LLC, New York, NY, 2024a) with OPLS force field [23], generating lower energy conformers and ionization states at pH 7.0 ± 2.0. The co-crystallized structures of the chosen receptors were retrieved from the Protein Data Bank (PDB) website (https://www.rcsb.org/pdb/) Lanosterol 14-α-demethylase (CYP51) from candida albicans (pdb id: 5TZ1, 2.0 Å), dihydrofolate reductase (DHFR) from staphylococcus aureus (pdb id: 3SRW, resolution 1.7 Å), and DNA gyrase subunit B (GyrB) from Escherichia coli (pdb id: 4DUH, resolution: 1.5 Å). These targets were selected for their established roles in fungal and bacterial viability and for alignment with the reported in vitro activity of Xeroderris stuhlmannii extracts. CYP51 is a key enzyme in ergosterol biosynthesis making it appropriate for antifungal evaluation [24][25]. DHFR catalyzes the reduction of dihydrofolate to tetrahydrofolate and is a validated antibacterial target, notably for gram positive pathogens [26]. DNA gyrase mediates DNA supercoiling and is a canonical target for antibacterial agents active against gram negative organisms [25]. The protein structures were prepared using the protein preparation wizard of Schrödinger suite (Schrödinger Release 2024-4, Protein Preparation Wizard; Epik, Schrödinger, LLC, New York, NY, 2024; Impact, Schrödinger, LLC, New York, NY; Prime, Schrödinger, LLC, New York, NY, 2024b). The preparation protocol involved: assigning of bond orders and addition of hydrogen, removal of water molecules beyond 5 Å from the native ligand, missing side chains and loops were modelled, protonation states were assigned using Epik at pH 7.0 ± 2.0, and structures were subjected to restrained minimization with OPLS4 force field. Receptor grids were generated centered on the coordinates of the co-crystallized ligands. Docking simulation was validated by redocking the native ligands into their respective active site. Validation success was determined by RMSD values between the docked and crystallographic pose, with RMSD < 2.0 Å considered acceptable [27].

To refine the accuracy of the ligand, pose ranking and evaluate the thermodynamic stability and selectivity of the docked complexes, the binding free energies ΔGbind were computed by applying the prime Molecular Mechanics generalized Born Surface Area (MMGBSA) approach (Schrödinger Suite 2024 User Manual; Prime MM-GBSA Calculation Methodology and Solvation Models, Schrödinger, New York, NY). This method provides a more rigorous estimation of the binding energy compared to standard scoring functions by incorporating solvation effects. The binding free energies ΔGbind was determined from the output poses obtained from docking simulation using the following equations:

ΔGbind = ΔG (complex) − (ΔG (Protein) + ΔG (Ligand) (1)

The free energy, (for each component is estimated by the molecular mechanics energy (EMM), the solvation free energy (Gsolv) and the product of the absolute temperature (T) and entropy change (ΔS):

ΔGbind = ΔEMM + ΔGsol − TΔS (2)

QikProp was used to determine the toxicity, bioavailability, and pharmacodynamic characteristics of 4 selected hits FAXD, WF₅₄, NKL₁₉ and NKL₈ isolated from X. stuhlmannii. The compounds were chosen based on factors such as Lipinski’s rule of five (Ro5), gastrointestinal absorption, inhibition of CYP450 isoenzymes, hepatotoxicity, ocular irritation, corrosion, biodegradation, and others. Compounds that passed the drug-likeness screening were docked using the gliding docking techniques.

To prevent adverse ADMET features (absorption, distribution, metabolism, excretion, and toxicity), pharmacokinetics and toxicity should be considered throughout the initial stages of drug development. The compounds potential druggability was estimated using QikProp. In this study, we first estimated the database’s ADMET attributes. Table I shows the results for the 04 selected hits.

Indeed, Table 1 illustrates that ligands FAXD, NKL₈, NKL₁₉, AND WF₅₄ (See Figure 1) exhibit favorable PHO, HOA, and QPPCaco permeability values, indicating good oral absorption and permeability. Notably, NKL₁₉ has the lowest PSA, which supports enhanced membrane permeability and distribution. Based on the overall ADMET profiles, which rank the ligands from most to least likely to be good drug candidates FAXD and WF₅₄ have high cell permeability, solubility, and oral absorption (PSA) contributing to good oral absorption; low PSA and high lipophilicity enable good distribution, low risk of cardiotoxicity; and show strong absorption and distribution properties. However, they show a slightly higher metabolism risk due to hydrophobic regions and low brain penetration. To summarize, both selected hits exhibit an optimal balance of drug likeness properties and ADME potential.

Figure 1. Chemical structures of the selected compounds for molecular docking studies.

Table 1. QikProp pharmacokinetic and toxicity predictions of selected ligands.

SASA = Solvent Accessible Surface Area, FOSA = Hydrophobic component of SASA, FISA = Hydrophilic component of SASA, PISA = Pi (flat) component of SASA, WPSA = Total SASA excluding PISA, R_O5/3 = Role of Five/Three, Volume = Molecular volume, Donor HB = Estimated number of hydrogen bonds donated by the molecule, AccptHB = Estimated number of hydrogen bonds accepted by the molecule, HOA = Predicted human oral absorption on 0 - 3 scale, PHO = Percentage of oral absorption, PSA = Polar surface area, QplogS = Predicted aqueous solubility, CIQPlogS = Consensus log S (average of two prediction methods), QPlogHERG = Predicted IC₅₀ value for blockage of HERG K+ channels, QPPCaco = Predicted apparent Caco-2 cell permeability, QPlogBB = Predicted brain/blood partition coefficient.

Molecular docking simulation was conducted to investigate the binding interactions of the isolated compounds with key microbial targets: fungal lanosterol 14-α-demethylase (PDB: 5TZ1), bacterial dihydrofolate reductase (DHFR, PDB: 3SRW), and bacterial DNA gyrase subunit B (GyrB, PDB: 4DUH). The binding affinities were evaluated using docking score, XP score, Glide score, with more robust binding energy estimates provided by the MMGBSA calculations (Table 2). To ensure the reliability of the docking procedure, redocking of each receptor’s native ligand was performed, and the resulting RMSD superimposition values were all below 2 Å (Figure 2).

Figure 2. Redocking superimposition of Cocrystal ligand of (A) 5TZ1, (B) 3RSW, and (C) 4DUH.

Table 2. Docking score results of compounds.

WF₅₄ = Stuhlmarotenoid A; NKL₁₉ = 7-O geranyl formononetin; NKL₈ = Conrauinone C and FAXD = Conrauinone A.

Figure 3. 3D pose view and 2D interaction map of fungal lanosterol 14-α-demethylase complexes (A) NKL₈-5TZ1, (B) NKL₁₉-5TZ1, (C) WF₅₄-5TZ1 and (D) FAXD-5TZ.

Figure 4. 3D pose view and 2D interaction map of (A) Nystatin-5TZ1, (B) Cipro-3SRW and (C) Cipro-4HUD.

3.2.1. Docking Interactions with Fungal Lanosterol 14-α-Demethylase

Docking against lanosterol 14-α-demethylase showed critical interactions with the Heme group and active cavity (Figure 3). The analysis of binding poses within the lanosterol 14-α-demethylase active site demonstrated distinct interaction profiles for each compound. Conrauinone C (NKL₈) exhibited a superior docking score of −10.392 Kcal/mol, forming a pi-pi T-shaped interaction and a pi-donor hydrogen bond with essential HEM601 group, alongside extensive pi-alkyl interactions with Met508, Phe380, His377, and Leu376 with the active site (Figure 3(A)). All those high interactions strongly support the fact that the geranyl moiety enhances the antibacterial activity of these compounds. Similarly, 7-O-geranylformononetin (NKL₁₉), showed a docking score of −10.131 kcal/mol, engaged in pi-pi T-shaped interactions with Tyr118, His377, and Phe233, a Van der Waals interaction with HEM601 group, and pi-alkyl interaction with Tyr132, Leu376, and Met508 (Figure 3(B)). Stuhlmarotenoid A (WF₅₄) exhibited a docking score of −9.482 kcal/mol. The compound directly engages the catalytic heme group through pi-pi T-shaped interactions and stabilizes its position in the active site via hydrogen bond with Tyr505 and a pi-pi T-shaped interaction with Phe228 (Figure 3(C)). In contrast, the binding mode of Conrauinone A (FAXD) with docking score −7.408 kcal/mol implies an alternative mechanism. Conrauinone A anchors itself within the active site through pi-cation interaction with Lys90, supported by extensive hydrophobic contacts (alkyl and pi-alkyl) with residues His377, Pro230, and Tyr64 (Figure 3(D)). The reduction in term of docking score in conrauinone A could be due to the replacement of the geranyl moiety by the 2,2-dimethylpyran substituents on ring A. The control drug Nystatin achieved binding through five hydrogen bonds with Asn187, Asp225, and Glu514 (Figure 4(A)).

3.2.2. Docking Interactions of Compounds with Bacterial Dihydrofolate Reductase (Dhfr)

Docking against DHFR shows that high binding affinity involves extensive hydrophobic contact in the active site. The docking results against DHFR revealed Stuhlmarotenoid A (WF₅₄) as top binder with docking score and MMGBSA score of −7.63 kcal/mol and −61.39 kcal/mol respectively. Its high affinity is driven by a hydrogen bond with Gln20 that anchors the molecule in DHFR active site, and extensive hydrophobic contacts across the wide surface of the binding pocket, engaging Leu29, Leu21, Val32, Lys33, Val7, Ile15, Phe99, and Pro26 (Figure 5(A)). The highest values observe with stuhlmarotenoid A is due to the important presence of methoxy group on almost all the cycle as compared to other compounds. NKL₈ exhibited a docking score and MMGBSA scores of −6.890 kcal/mol and −57 kcal/mol. The compound was stabilized in DHFR active through a conventional hydrogen bond with Th122. It was also observed to engage pi-alkyl interactions with as well as alkyl interactions with Lys30, Lys33, and Phe99 (Figure 5(B)). On the other hand, FAXD achieved a docking score of −5.848 kcal/mol and MMGBSA score of −48.83 kcal/mol. The ligand was stabilized in the active site through pi-alkyl interactions with residues Leu29, Leu55, Ile15, Phe99, and Phe93, complemented alkyl interactions with Ile51, suggesting a predominantly hydrophobic stabilization with the active pocket (Figure 5(C)). Lastly, NKL₁₉ showed a relatively weal affinity for DHFR (−3.147 kcal/mol), stabilized mainly by pi-alkyl interactions with Leu29, Leu55, Leu21, and Val32 (Figure 5(D)). Ciprofloxacin exhibited a well-defined binding mode, forming hydrogen bonds with Gln96 and Thr122, a halogen bond with Gln20, and hydrophobic interactions with Leu21, Phe93, Phe99, and Ile15 (Figure 4(B)).

3.2.3. Docking Interactions of Compounds with DNA Gyrase B

Docking against DNA-gyrase demonstrates that the inhibition correlates with disruption of the ATP-binding site through key electrostatic interactions (Figure 6). NKL₈ was the most potent achieving a docking score of −6.668 kcal/mol, forming a hydrogen bond with Asp73, pi-cation contact with Glu50, and an amide-pi stacked interaction with Asn46 (Figure 6(A)). FAXD also demonstrated a strong binding pose driven by three hydrogen bonds to Arg136 with Arg76 and Lys103, a pi-anion interaction with Glu50, as well as van der Waals with Asn46 (Figure 6(B)). NKL₁₉ demonstrated a docking score of −5.780 kcal/mol. It exhibited a water mediated hydrogen bond, pi-cation contact with Arg76 and Lys103. It also displayed a pi-anion contact with Glu50, and Van der Waals contact with Asn46 (Figure 6(C)). WF₅₄ (docking score −3.920 kcal/mol) interacted was stabilized through hydrogen bond with Arg136, pi-cation interaction with Arg76 and Lys103, pi-anion contact with Glu50 and lastly an amide-pi stacked with Asn46 (Figure 6(D)). Notably, several compounds consistently engaged ASN46, a residue that coordinates the catalytic Mg²⁺ required for ATP hydrolysis, indicating that ligand binding at this site may directly disrupt enzymatic activity [28][29]. Ciprofloxacin reproduced the expected binding pattern, confirming docking reliability (Figure 4(C)).

Figure 5. 3D pose view and 2D interaction map of bacterial dihydrofolate reductase complexes (A) WF₅₄-3SRW, (B) NKL₈-3SRW, (C) FAXD-3SRW, (D) NKL₁₉.

Table 3. Structure-activity relationship between non and prenylated compounds from X. stuhlmannii.

Figure 6. 3D poses view and 2D interaction map of bacterial DNA gyrase complexes (A) NKL₈-4HUD, (B) FAXD-4HUD, (C) NKL₁₉-4HUD, (C) WF₅₄-4HUD.