New specimens of Alternantheraphiloxeroides were collected in September 2023 from Pabna, Bangladesh, and were taxonomically identified at the Salar Khan Herbarium, University of Dhaka (voucher no. DUSH-10819). The entire plant of Alternantheraphiloxeroides was air-dried and then oven-dried at 40˚C for 24 h. The dried material (100 g) was pulverized into a coarse powder and extracted by maceration with 250 mL of 100% methanol (1:2.5, w/v) at room temperature for 12 days, with occasional shaking. The extract was filtered and concentrated under reduced pressure to obtain 9.40 g of dried crude methanolic extract, corresponding to an extraction yield of 9.4% [12][13].

2.2.1. FTIR (Fourier Transform Infrared Spectroscopy) Analysis of the Sample

FTIR analysis of the methanolic crude extract was performed using a BRUKER ALPHA II FTIR spectrophotometer over the spectral range of 4000 - 500 cm⁻¹ to identify the functional groups present [14][15]. The dried extract was analyzed using the diffuse reflectance technique, and the resulting spectra were compared with reference data to assign functional groups [16][17].

2.2.2. ¹H NMR (600 MHz, CD₃OD) and ¹³C NMR (150 MHz, CD₃OD) Spectra of the Crude Extracts

Nuclear Magnetic Resonance (NMR) spectra were obtained on a Bruker Avance II 600 MHz spectrometer with CD₃OD (deuterated methanol) used as the solvent. Chemical shifts (δ) were expressed in parts per million (ppm) relative to tetramethylsilane (TMS) as the internal standard. The coupling constant (J) was reported in hertz (Hz). Spin multiplicities were labeled as singlet (s), broad singlet (brs), doublet (d), double doublet (dd), doublet of doublets (ddd), triplet (t), quintet (quint), doublet of triplets (dt), triplet of doublet doublets (tdd), and others [18].

2.3.1. Antimicrobial Assay

Preliminary antimicrobial activity was determined using the disc diffusion method. The crude extract (5 mg) was dissolved in 100 μL of methanol, and 10 μL (500 μg per disc) was loaded onto sterile discs. Antibacterial activity was examined against Staphylococcusaureus (ATCC 9144), Bacilluscereus (clinical isolate), Escherichiacoli (ATCC 11303), Salmonellatyphi (ATCC 13311), and Pseudomonasaeruginosa (ATCC 27833). Antifungal activity was evaluated against Aspergillusniger (ATCC 1004) and Aspergillusflavus (UCFT 02). After incubation (18 h at 37˚C for bacteria and 48 h at 28˚C for fungi), inhibition zones were recorded and compared with kanamycin (30 μg per disc) and ketoconazole (30 μg per disc) as positive controls, whereas methanol (10 μL per disc) was used as the negative control [19][20].

2.3.2. Antioxidant Activity

Antioxidant activity was assessed using the DPPH (2,2-diphenyl-1-picrylhydrazyl) free-radical scavenging assay, according to the procedure described by Brand-Williams etal. (1995). We performed the DPPH assay in three independent experiments, with three replicate measurements at each concentration. The crude extract (1.6 mg) was dissolved in methanol (400 μL) and subjected to serial dilution (0.78 - 200 μg/mL). For each dilution, 2 mL was combined with an equal volume of DPPH solution (20 μg/mL), incubated in the dark at room temperature for 30 min, and the absorbance was recorded at 517 nm. Ascorbic acid and butylated hydroxyanisole (BHA) were used as positive controls, while the DPPH solution mixed with solvent acted as the negative control. Methanol alone served as the solvent blank to zero the spectrophotometer before absorbance measurements at 517 nm. The IC₅₀ (concentration needed to scavenge 50% of DPPH radicals, expressed in µg/mL) was calculated by fitting a logarithmic regression to the plot of percent inhibition against concentration in Microsoft Excel software. These figures were subsequently compared with the corresponding values for the two standard antioxidants, ascorbic acid and butylated hydroxyanisole (BHA) [21]. Radical scavenging activity was calculated as

where A Blank is the absorbance of the blank and A Sample is the absorbance of the sample.

The endophytic fungus was isolated using a surface sterilization technique and subsequently identified based on macroscopic and microscopic morphological characteristics, as described in our previous study [22], followed by molecular analysis. Molecular identification was carried out through amplification, sequencing, and subsequent analysis of the internal transcribed spacer (ITS) region [23]. A 7- to 10-day-old fungal culture was used for genomic DNA extraction. The fungal biomass was scraped aseptically using a sterile surgical blade and homogenized with a pestle and mortar in liquid nitrogen. Genomic DNA was subsequently extracted using the Maxwell^® 16 LEV Plant DNA Kit (AS1420, Promega, USA) according to the manufacturer’s protocol [24]. Residual RNA contamination was removed by treatment with DNase-free RNase at 37˚C for 30 min. The ITS region was amplified by polymerase chain reaction (PCR) using the universal fungal primers ITS1F and ITS4. The amplified products were subjected to Sanger dideoxy sequencing [25][26]. The resulting sequences were manually edited using BioEdit v7.2 and compared with reference sequences deposited in the NCBI GenBank rRNA/ITS database using the BLASTn search tool for molecular identification.

For large-scale production of secondary metabolites, the fungal isolate was cultivated on potato dextrose agar (PDA). Each Petri plate contained 1.1 g of PDA, and a total of 540 culture plates containing 569.4 g of PDA were prepared. The cultures were incubated at 28˚C for 21 - 27 days. Following incubation, the fungal cultures were pooled and extracted three times with ethyl acetate (5 L total volume). The combined ethyl acetate extracts were concentrated under reduced pressure to yield 13.25 g of crude extract. The crude extract was subjected to silica gel column chromatography using silica gel (70 - 230 mesh, 283.75 g) packed in a glass column (bed height: 33 cm, internal diameter: 4.3 cm). Elution was carried out using a step gradient of n-hexane, n -hexane-dichloromethane (DCM), DCM, and DCM-methanol with increasing polarity. A total of 57 fractions (100 mL each) were collected and pooled into 15 combined fractions based on their TLC profiles. Fraction 9 (1.2 g) was re-chromatographed on a smaller silica gel column (bed height: 16 cm, internal diameter: 2 cm), yielding 23 fractions (50 mL each) that were grouped into 14 major fractions according to TLC monitoring. Fraction 5 was further purified by preparative thin-layer chromatography (PTLC; stationary phase: silica gel F₂₅₄, plate thickness: 0.2 mm) using toluene-20% ethyl acetate containing 0.5% acetic acid as the mobile phase. The desired target band (R_f = 0.53) was scraped from the plate and eluted sequentially with DCM, DCM-ethyl acetate (1:1, v/v, applied twice), ethyl acetate, and ethyl acetate containing 0.25% - 0.5% methanol. Evaporation of the solvent yielded compound 1 (3 mg) as a light brown to colorless oil. The structure of the isolated pure compound was elucidated through extensive 1D and 2D NMR spectroscopic analysis (¹H, ¹³C, COSY, HSQC, and HMBC) [27][28].

The crystal structures of rat neuronal nitric oxide synthase (PDB ID: 6NGJ; 1.76 Å), glutathione reductase (PDB ID: 3GRS; 1.54 Å), DNA gyrase (PDB ID: 1KZN; 2.30 Å) and penicillin-binding protein 2X (PDB ID: 5OJ0; 2.66 Å) were retrieved from the Protein Data Bank based on their reported relevance to antioxidant and antimicrobial activities [29]-[32]. The proteins were prepared and energy-minimized using Swiss-PdbViewer. The isolated Dendryphiellin I and reference compounds were prepared as ligands, and molecular docking was performed using PyRx with AutoDock Vina. The grid-box dimensions and center coordinates were specified along the X, Y, and Z axes. For 6NGJ, the grid-box dimensions were 69.341 × 79.512 × 59.232 Å, with center coordinates of X = 11.464, Y = 4.904, and Z = 22.600 Å. For 3GRS, the dimensions were 95.182 × 48.242 × 69.517 Å, with center coordinates of X = 70.381, Y = 51.349, and Z = 18.712 Å. For 1 KZN, the dimensions were 46.482 × 46.891 × 41.213 Å, with center coordinates of X = 19.519, Y = 19.178, and Z = 43.213 Å. For 5OJ0, the dimensions were 64.440 × 63.051 × 108.451 Å, with center coordinates of X = 47.963, Y = 14.717, and Z = 31.4596 Å. The resulting ligand-receptor complexes were analyzed using BIOVIA Discovery Studio Visualizer to identify and characterize intermolecular interactions, including conventional hydrogen bonds, carbon-hydrogen interactions, alkyl interactions, π-alkyl interactions, and other non-covalent interactions. The binding affinities obtained from AutoDock Vina were used to compare the predicted binding potential of Dendryphiellin I with that of the respective reference compounds [33]. To assess the reliability and reproducibility of the molecular docking protocol, a validation procedure was performed for the selected protein targets. For 6NGJ, 1KZN, and 5OJ0, conventional ligand redocking was performed using the experimentally determined co-crystallized ligands present in the respective crystal structures. The crystallographic ligands were removed from the corresponding protein structures and independently redocked into their original binding sites using the same docking software, search parameters, and docking protocol applied to the test compounds. The predicted binding poses were compared with the experimentally observed crystallographic poses by superimposition, and the root-mean-square deviation (RMSD) between the corresponding ligand atoms was calculated [34]. For 3GRS, validation was performed using a literature-guided binding-site and reference-ligand approach because conventional crystallographic ligand-redocking was not employed. The functionally characterized FAD/redox-active region of glutathione reductase was selected as the docking site based on previously reported structural and molecular docking studies [34][35]. The binding-site region comprised VAL49, GLU50, SER51, HIS52, GLY56, THR57, CYS58, VAL61, GLY62, CYS63, LYS66, GLY128, HIS129, ALA130, ALA155, THR156, GLY157, TYR197, ARG291, GLY330, ASP331, LEU337, LEU338, THR339, and PRO340 [36]. The docking grid was centered at X = 59.698, Y = 52.917, and Z = 18.674 Å, with the grid dimensions set to 25 × 25 × 25 Å. Butylated hydroxytoluene (BHT), a previously reported reference antioxidant, was used as a reference ligand and docked into the selected binding region using the same docking protocol. The resulting binding mode and interactions of BHT were compared with the previously reported literature findings to support the appropriateness of the selected binding site and docking procedure [36].

In silico ADME and toxicity profiles of the isolated compound and reference compounds were evaluated using SwissADME and pkCSM web-based platforms, respectively. Swiss ADME was used to predict physicochemical properties, lipophilicity, gastrointestinal absorption, blood-brain barrier permeability, and drug-likeness based on Lipinski’s rule of five [37][38]. The toxicity-related pharmacokinetic parameters were further estimated using the pkCSM platform [39].

The phytochemical profiling shows a better response for different groups of compounds in methanolic extracts. The FT-IR spectrum showed different characteristic absorption bands (Figure 1 and Table 1). The ¹H NMR (600 MHz, CD3OD) and 13C NMR (150 MHz, CD3OD) spectra of the plant crude extracts also displayed multiple chemical shifts (Figure 2 and Figure 3 and Table 2).

Figure 1. FTIR spectra of Alternantheraphiloxeroides.

Table 1. FT-IR spectrum peaks of plant crude extracts and their possible assigned functional groups [40].

Table 2.Compounds characterized via ¹H NMR spectroscopy.

Figure 2.¹H nuclear magnetic resonance (NMR) analysis of the plant’s crude extract.

Figure 3.¹³C NMR spectrum of plant crude extract.

The disc diffusion method was used to evaluate antimicrobial activity. The crude plant extract showed no activity against the bacterial strain S.typhi or against all fungal strains (Table 3). The antibacterial and antifungal reference compounds, kanamycin and ketoconazole, demonstrated the greatest potency against the tested bacterial and fungal strains. As anticipated, the solvent methanol produced no activity. The plant extract produced the strongest antimicrobial activity against the bacterial strains B.cereus (11.5 mm) and P.aeruginosa (12 mm), and it showed only mild activity against the bacterial strains S.aureus (9 mm) and E.coli (9 mm).

Table 3. Antimicrobial activity of crude extract of Alternantheraphiloxeroides (Mart.) Griseb. at a concentration of 500 µg/disc.

“–” indicates no activity detected and “nd” indicates not done.

The antioxidant activity of the methanolic crude extract of Alternantheraphiloxeroides was evaluated using the DPPH radical-scavenging assay. The extract exhibited concentration-dependent radical-scavenging activity, with increasing inhibition observed at higher concentrations. The IC₅₀ value of the crude extract was 127.48 ± 0.53 µg/mL, whereas the reference antioxidants ascorbic acid and butylated hydroxyanisole (BHA) showed substantially stronger activity, with IC₅₀ values of 4.31 ± 0.07 and 3.40 ± 0.08 µg/mL, respectively (Figure 4). Thus, the crude extract demonstrated moderate DPPH radical-scavenging activity compared with the reference standards.

Figure 4. Free radical scavenging activity of crude extract and standards.

The assembled sequence was obtained from the forward-primer sequence and cleaned by removing the first 18 base pairs based on the chromatogram inspection. The cleaned sequence was then subjected to BLASTn analysis against the NCBI database, followed by phylogenetic analysis using MEGA 12, and the top-scoring hit was selected as the putative identification of the fungal strain. The fungal isolate’s sequence was successfully amplified and compared against the NCBI database using the BLASTn algorithm. The sequence produced significant alignments with several species within the genus Curvularia, a diverse group known for its endophytic and pathogenic lifestyles in plants. The ITS rDNA sequence obtained from isolate APLE-2 was subjected to BLASTn analysis against the NCBI GenBank database. The sequence exhibited its highest similarity (96.91% identity; 99% query coverage; E-value = 0.0) to the ITS sequence of Curvularialunata CBS 730.96 (type material; GenBank accession NR_138223.1). Phylogenetic analysis based on ITS sequences further confirmed the placement of APLE-2 within the genus Curvularia. However, APLE-2 did not form a well-supported sister relationship or cluster directly with C.lunata CBS 730.96, but instead occurred on a distinct lineage in the reconstructed tree. Therefore, considering both the BLASTn results and the phylogenetic placement, APLE-2 was designated Curvularia sp. Further multilocus sequence analysis incorporating additional phylogenetically informative loci would be necessary to achieve robust species-level identification. Following taxonomic confirmation, the validated ITS sequence will be submitted to the NCBI GenBank database for accession number assignment (Figure 5 and Figure 6).

ITS rDNA Sequence of the fungal isolate APLE-2:

The isolated compound (3 mg), with the molecular formula C₂₅H₃₂O₅, an exact mass of 412.22 Da, and an R_f value of 0.53, was characterized as Dendryphiellin I based on its ¹H and ¹³C NMR spectroscopic data (Figure 7 and Figure 8), which were consistent with the reported spectroscopic data for Dendryphiellin I. The ¹³C NMR spectrum exhibited 25 carbon signals. In the ¹H NMR spectrum, an aldehydic proton was observed at δH 9.52 (1H, s), together with characteristic olefinic signals at δH 7.30 (1H, d, J = 15.6 Hz), 5.66 (1H, d, J = 9.6 Hz), 6.85 (1H, s), 6.28 (1H, dd, J = 9.9, 5.1 Hz), 6.41 (1H, d, J = 9.6 Hz), 6.01 (1H, s) and 5.77 (1H, d, J = 15.6 Hz). The ¹³C NMR spectrum showed characteristic signals at δC 197.2, 192.8, 167.2, 162.0, 154.6, 149.2, and 74.9 ppm, corresponding to carbonyl, aldehyde, olefinic, oxygenated, methine, methylene, and methyl carbons (Table 4).

Figure 5. BLAST result of the fungal isolate APLE-2.

Figure 6. Phylogenetic analysis of fungal isolate APLE-2 based on ITS rDNA sequences.

Figure 7.¹H NMR spectrum (600 MHz, CDCl3) of compound 1.

Figure 8.¹³C NMR spectrum (150 MHz, CDCl3) of compound 1.

Table 4.¹H (600 MHz, CDCl3) and ¹³C (150 MHz, CDCl3) NMR data for compound 1 with reference data [41].

Molecular docking analysis was performed to investigate the binding potential of Dendryphiellin I against the selected antioxidant target proteins, 6NGJ and 3GRS. Dendryphiellin I exhibited binding affinities of −8.5 kcal/mol and −8.1 kcal/mol against 6NGJ and 3GRS, respectively, which were considerably higher in magnitude than those of the reference compound ascorbic acid (−6.5 and −5.6 kcal/mol, respectively) (Table 5). These results suggest that Dendryphiellin I showed a comparatively stronger predicted binding affinity toward both antioxidant targets than the reference compound. The interaction analysis showed that Dendryphiellin I established multiple non-covalent interactions within the binding sites of both target proteins (Table 6). In the 6NGJ complex, the ligand was involved in conventional hydrogen bonding with TRP409, TRP587, and TYR706, with bond lengths of 2.57 - 2.96 Å. Additional alkyl, π-alkyl, and carbon–hydrogen interactions were observed with several neighboring amino acid residues, further contributing to ligand stabilization within the binding pocket. In comparison, ascorbic acid formed conventional hydrogen bonds with SER413, ARG414, ASN411, TYR698, and GLU705, with bond distances of 2.18 - 2.87 Å, together with additional carbon-hydrogen interactions.

Table 5. Binding affinities of Dendryphiellin I and reference compounds against selected antioxidant and antimicrobial target proteins.

Table 6. Non-bonding interactions of Dendryphiellin I and reference compound ascorbic acid with antioxidant target proteins 6NGJ and 3GRS.

A = Alkyl, PA = Pi-Alkyl, CHB = Conventional Hydrogen Bond, CH = Carbon Hydrogen Bond.

For the 3GRS complex, Dendryphiellin I showed a conventional hydrogen-bond interaction with ASN294 (2.21 Å), along with additional alkyl and π-alkyl interactions involving residues including VAL61, LYS53, VAL107, and HIS52. Ascorbic acid, on the other hand, exhibited conventional hydrogen-bond interactions with GLY158, GLU50, and ALA155, together with carbon-hydrogen interactions involving GLY157, THR156, and GLU50. The overall interaction profiles and binding poses of Dendryphiellin I and ascorbic acid within the respective target proteins are presented in Figure 9.

Figure 9. Molecular docking interactions and binding poses of Dendryphiellin I and the standard drug against the selected antioxidant target proteins, PDB ID: 6NGJ and PDB ID: 3GRS.

Molecular docking analysis was further performed to evaluate the binding potential of Dendryphiellin I against the selected antimicrobial target proteins, 1 KZN and 5OJ0. As shown in Table 5, Dendryphiellin I exhibited binding affinities of −7.3 kcal/mol against 1 KZN and −8.5 kcal/mol against 5OJ0. The corresponding binding affinities of the reference drug kanamycin were −7.0 and −7.4 kcal/mol, respectively. Thus, Dendryphiellin I showed slightly stronger predicted binding toward 1 KZN and a comparatively stronger binding affinity toward 5OJ0 than kanamycin.

The interaction analysis revealed that Dendryphiellin I engaged in multiple noncovalent interactions with residues within the binding sites of both antimicrobial targets (Table 7). Against 1 KZN, Dendryphiellin I formed conventional hydrogen-bond interactions with ALA96 at distances of 2.47 and 2.69 Å, together with carbon-hydrogen interactions involving HIS95 (2.71 - 2.97) Å and GLY119 (2.84 Å). Additional π-alkyl interactions were observed with TYR568 and PHE570, while alkyl interactions involved ALA47, ILE78, ILE90, VAL43, VAL120, and VAL167. In comparison, kanamycin formed several conventional hydrogen bonds with ASN46, GLY77, ALA96, VAL120, ASP73, THR165, and GLU50, with bond distances ranging from 2.09 to 2.98 Å, along with additional carbon-hydrogen interactions.

For 5OJ0, Dendryphiellin I showed a conventional hydrogen-bond interaction with HIS594 (2.85 Å). It also exhibited π-alkyl interactions with TRP374, TYR568, PHE570, and HIS594 within the binding pocket. Kanamycin displayed several conventional hydrogen-bond interactions involving ASN185, ALA312, ARG186, ASP313, ALA153, LEU130, and TYR129, with bond distances ranging from 2.28 to 2.90 Å, in addition to carbon–hydrogen interactions. The predicted binding modes and intermolecular interactions of Dendryphiellin I and kanamycin with the selected antimicrobial targets are presented in Figure 10.

Table 7. Non-bonding interactions of the isolated compound and kanamycin with antimicrobial target proteins 1 KZN and 5OJ0.

A = Alkyl, PA = Pi-Alkyl, CHB = Conventional Hydrogen Bond, CH = Carbon Hydrogen Bond.

Figure 10. Molecular docking interactions and binding poses of Dendryphiellin I and the standard drug against the selected antibacterial target proteins, PDB ID: 1 KZN and PDB ID: 5OJ0.

Docking protocol validation was assessed by redocking the experimentally determined co-crystallized ligands into their corresponding binding sites. The calculated RMSD values were 2.40 Å, 3.60 Å, and 2.03 Å for 6 NGJ, 1 KZN, and 5OJ0, respectively (Figure 11). Although these values were slightly above or, in the case of 1 KZN, clearly above the commonly applied 2.0 Å threshold, the redocking results provided partial structural support for the selected docking protocols. The higher RMSD observed for 1 KZN indicates limited reproduction of the crystallographic ligand orientation and was therefore not interpreted as definitive validation. For 3 GRS, conventional redocking validation was not performed because an experimentally resolved co-crystallized ligand suitable for this purpose was not available. Instead, the docking site was defined based on the reported FAD/redox-active region of glutathione reductase, and BHT was docked as a reference antioxidant ligand. The resulting pose was located within the reported functional region, providing structural support for the selected binding site rather than formal validation of the docking protocol. Accordingly, the docking results were interpreted as predictive/supportive evidence and were considered together with the experimental findings rather than as definitive proof of binding.

The predicted drug-likeness and ADMET profiles of Dendryphiellin I and the reference compounds are presented in Table 8. Dendryphiellin I showed high gastrointestinal absorption and no predicted blood-brain barrier permeability. It complied with Lipinski’s rule of five without violations and showed a bioavailability score of 0.55. However, it exhibited relatively high lipophilicity (XLOGP3 = 5.14) and was predicted to inhibit CYP3A4. In comparison, ascorbic acid also showed high gastrointestinal absorption and no Lipinski violations, whereas kanamycin showed low gastrointestinal absorption and two Lipinski violations. Dendryphiellin I showed a lower predicted aqueous solubility (ESOL logS = −5.11) than ascorbic acid (logS = 0.23) and kanamycin (logS = 1.93), indicating comparatively lower solubility. Its predicted skin permeability (log Kp = −5.17) was higher than that of ascorbic acid (−8.54) and kanamycin (−14.19). None of the tested compounds were predicted to cause hepatotoxicity or inhibit hERG I/II channels.

Overall, Dendryphiellin I exhibited drug-likeness comparable to ascorbic acid, while showing a high predicted gastrointestinal absorption profile and better compliance with Lipinski’s rule than kanamycin.

Figure 11. Docking validation of the selected protein targets.

Table 8. Drug-likeness and ADMET analysis of studied compounds.