Results and discussion

Inhibitory effects of H. brasiliense hexane extract and purified compounds on the planktonic growth of S. aureus

In this experimental set, a panel of eight S. aureus strains representing both MSSA and MRSA phenotypes was evaluated. In this context, S. aureus ATCC 29213 was included as the reference MSSA strain, while the MRSA group comprised two Brazilian epidemic clones (BECs HU25 and BMB9393), the community-associated USA300 lineage, and four clinical isolates (CR14-005, CR14-021, CR14-026 and CD16-016). All bacterial strains were tested against heHb and its purified components using the broth microdilution method in accordance with CLSI guidelines. Both heHb and uliB exhibited comparable inhibitory and bactericidal activities across the diverse genetic backgrounds analysed, with MIC values ranging from 3.125 to 6.25 μg mL-1 and MBC values from 6.25 to 12.5 μg mL-1 (Table 1). This consistency across MSSA and MRSA strains highlights the broad-spectrum and robust antibacterial potential of these compounds. Additionally, the close correspondence between MIC and MBC values indicates that heHb and uliB act predominantly through a bactericidal mechanism. Among the remaining isolated metabolites, isouliginosin B showed only moderate antibacterial activity, with MIC values of 6.25-12.5 μg mL-1 and MBC values of 12.5-50 μg mL-1, while japonicin A exhibited no detectable activity (MIC >100 μg mL-1) (Table 1). Geometric mean values were calculated for all strains and are presented in Table 1. Collectively, these findings further emphasize uliB as a bioactive component responsible for the potent anti-S. aureus effects observed in heHb. Given their pronounced efficacy, both heHb and uliB were selected for subsequent experimental analyses to elucidate their mechanisms of action and biological effects.

S. aureus strains | Concentration, μg mL-1
heHb | uliB | Isouliginosin B | Japonicin A
MIC | MBC | MIC | MBC | MIC | MBC | MIC | MBC
ATCC 29213 | 6.25 | 12.5 | 3.125 | 6.25 | 12.5 | 25 | >100 | >100
HU 25 | 3.125 | 6.25 | 3.125 | 3.125 | 12.5 | 25 | >100 | >100
USA 300 | 6.25 | 12.5 | 3.125 | 6.25 | 12.5 | 25 | >100 | >100
BMB 9393 | 6.25 | 6.25 | 3.125 | 6.25 | 12.5 | 25 | >100 | >100
CR14-005 | 3.125 | 6.25 | 3.125 | 6.25 | 12.5 | 50 | >100 | >100
CR14-021 | 3.125 | 6.25 | 6.25 | 12.5 | 6.25 | 25 | >100 | >100
CR14-026 | 3.125 | 6.25 | 3.125 | 6.25 | 6.25 | 12.5 | >100 | >100
CD16-016 | 6.25 | 12.5 | 3.125 | 12.5 | 12.5 | 25 | >100 | >100
Geometric mean | 4.42 | 8.11 | 3.41 | 6.82 | 10.51 | 25.00 | >100 | >100

Influence of inoculum size on the activity of H. brasiliense hexane extract and uliginosin B against S. aureus

Given the promising inhibitory effects of heHb and uliB on the proliferation of S. aureus, we next investigated whether these compounds were subject to an inoculum-dependent response. To this end, heHb and uliB were tested at their respective MICs against a range of initial bacterial densities (104, 10⁵, 10⁶, 107 and 108 CFU mL-1) and incubated for 24 h. Both compounds maintained strong activity at lower inoculum levels (104 to 10⁶ CFU mL-1), reducing metabolic activity by more than 80 %. Although a diminished response was observed at the highest bacterial loads (107 to 108 CFU mL-1), metabolic activity was still reduced by approximately 50 % (Figure 1). These results indicate that heHb and uliB retain substantial inhibitory activity even under elevated inoculum conditions, supporting their robustness and potential applicability against S. aureus populations of varying densities.

Figure 1.: Inoculum effect and metabolic activity of S. aureus treated with heHb and uliB. Bacterial suspensions ranging from 104 to 108 cells were incubated for 24 h at 37 °C with MIC concentrations of heHb (6.25 μg mL-1) (A) or uliB (3.125 μg mL-1) (B). Untreated cells were used as the control. Metabolic activity was assessed using the XTT assay based on tetrazolium salt reduction by metabolically active cells. Statistical analyses were performed using one-way ANOVA followed by Dunnett’s post hoc test, with a significance level of 95 % (p <0.05) and Dunnett's post-test. ***p <0.0001

Alterations in the ultrastructure of S. aureus induced by H. brasiliense hexane extract and uliginosin B

To investigate the effects of heHb and uliB on the ultrastructure of S. aureus cells, SEM and TEM analyses were conducted. Two representative strains were selected: USA300 (MRSA), due to its global dissemination and clinical relevance, and ATCC 29213 (MSSA), which served as a standard reference strain. Bacterial cells were exposed to heHb and uliB at their respective MIC concentrations for 24 h prior to imaging.

SEM analyses revealed clear ultrastructural differences between untreated and treated S. aureus cells. Untreated MSSA ATCC 29213 and MRSA USA300 cells (Figs. 2A and 2B) displayed smooth surfaces, intact cell walls and well-preserved spherical morphology, with no signs of distortion or damage. In contrast, exposure to heHb (Figures 2C and 2D) induced marked morphological alterations, including pronounced surface depressions, membrane collapse, leakage of intracellular material, increased surface roughness, overproduction of extracellular vesicles and extensive cell lysis. Treatment with uliB (Figures 2E and 2F) produced a distinct ultrastructural profile compared with the crude extract. Bacterial cells exhibited severe membrane disruption, abundant vesicle formation, accumulation of cellular debris and multiple nodular or blister-like protrusions, particularly evident in the USA300 strain, indicating substantial membrane destabilization. Overall, the SEM findings demonstrate that both heHb and uliB compromise S. aureus cell integrity, with uliB inducing more pronounced membrane-focused damage.

Figure 2.: Scanning electron microscopy of S. aureus ATCC 29213 and USA 300 treated with heHb and uliB for 24 h. A and B - untreated S. aureus cells. C and D - Cells treated with heHb at the MIC concentration (6.25 or 3.13 μg mL-1). E and F - cells treated with uliB at the MIC concentration (6.25 or 3.13 μg/mL). White arrows indicate intact, well-defined cocci; yellow arrows indicate cell surface depression; orange arrows indicate cell lysis and extravasation of intracellular contents. Blue arrows indicate vesicles in the cell wall.

TEM analysis revealed that untreated cells of S. aureus ATCC 29213 and USA300 displayed well-preserved ultrastructure, characterized by homogeneous cytoplasmic density, intact plasma membranes and regular cell wall thickness (Figure 3A, 3C, 3G and 3J).

Figure 3.: Transmission electron microscopy of S. aureus ATCC 29213 and USA 300 treated with heHb and uliB for 24 h. A /D and G/J - untreated ATCC 29213/ USA 300 cells; B/E and H/K - ATCC 29213/ USA 300 cells treated with heHb at MIC concentration (6.25 or 3.13 μg mL-1); C/F and I/L - ATCC 29213/ USA 300 cells treated with uliB at MIC concentration (6.25 or 3.13 μg mL-1). The black arrows indicate the cell wall with teichoic acids (TA); peptidoglycan (PG); periplasm (PP). White arrows indicate the plasma membrane (PM). Yellow arrows indicate the septum (SP). Blue arrows indicate the genetic material (GM). Orange arrows indicate mesosome-like structures (ME).

In contrast, bacterial cells exposed to heHb exhibited extensive structural disruption. In ATCC 29213, treated cells assumed amorphous shapes and showed clear defects in cell division, including impaired formation of the characteristic asymmetrical septum (Figures 3B and 3E). Similar abnormalities were observed in USA300, where heHb caused pronounced cytoplasmic disorganization, areas of reduced electron density indicative of intracellular degradation, thinning of the cell wall and severe membrane distortion (Figures 3H and 3K). Septum formation was irregular and incomplete, highlighting the extract’s disruptive impact on cell division machinery (Figure 3K). Treatment with uliB induced ultrastructural alterations broadly comparable to those caused by heHb; however, several distinctive features were observed.

UliB-treated cells exhibited increased and heterogeneous peptidoglycan thickness, highly irregular septum architecture and pronounced asymmetry during cell division. Additionally, signs of uneven chromosomal segregation and abnormal nucleoid distribution were evident (Figures 3C, 3F, 3I and 3L).

Elucidating the mechanisms of action of H. brasiliense hexane extract and uliginosin B in S. aureus

To capture early cellular responses prior to extensive growth-related effects, planktonic cultures of S. aureus treated with heHb and uliB for 3 h were evaluated for plasma membrane integrity and metabolic activity. Membrane permeability was assessed using PI staining, whereas metabolic activity was quantified using the XTT assay. Treatment with heHb or uliB did not significantly compromise membrane integrity, as PI uptake remained comparable to that of untreated cells and markedly lower than the high PI incorporation observed in boiled cells, which served as the positive control for membrane disruption (Figure 4A). In contrast, both compounds induced a marked reduction in metabolic activity, decreasing XTT reduction by approximately 80 % across all tested concentrations. These findings indicate that heHb and uliB primarily impair cellular metabolism rather than causing overt membrane rupture (Figure 4B). To further explore their mechanisms of action, intracellular ROS levels were quantified. After 30 min of exposure to 4×MIC of heHb or uliB, a substantial increase in ROS production was detected with the H₂DCF-DA probe (Figure 4C). Statistical comparisons showed no significant differences between the treated groups and the positive control (H₂O₂), while untreated cells exhibited minimal fluorescence (Figure 4C). Collectively, these results suggest that oxidative stress plays a key role in the antibacterial activity of both heHb and uliB compounds.

Figure 4.: Mechanisms of action of heHb and uliB in S. aureus. (A) Plasma membrane disruption in S. aureus treated with heHb and uliB for 3 h at 4× to 1/2× MIC concentrations (25 to 1.56 μg mL-1). Heat-killed cells were used as the positive control, and untreated cells as the negative control. Membrane integrity was assessed using propidium iodide, a membrane-impermeable fluorescent probe that penetrates cells with compromised membranes and binds to nucleic acids. (B) Metabolic activity of S. aureus during treatment with heHb and uliB. Bacteria were cultured for 3 h in the presence or absence of heHb and uliB at 4× to 1/2× MIC concentrations (25 to 1.56 μg mL-1), with untreated cells used as the control. Metabolic activity was evaluated using the XTT assay, based on tetrazolium salt reduction by metabolically active cells. (C) Induction of reactive oxygen species (ROS) in S. aureus treated with heHb and uliB at 4× MIC concentrations (25 and 12.5 μg mL-1) for 30 min. ROS levels were detected using the H₂DCF-DA probe, which is converted intracellularly into a fluorescent compound upon oxidation by ROS. Cells treated with H₂O₂ were used as the positive control, and untreated cells as the negative control. (D) Effect of the antioxidant NAC on bacterial growth at the MIC of heHb and uliB. S. aureus (ATCC 29213) cells were exposed to the MIC values of each compound in the presence or absence of NAC, and bacterial growth was assessed by measuring OD600 after incubation. Statistical analyses were performed using one-way ANOVA followed by Dunnett’s post hoc test, with a significance level of 95% (p <0.05). *** < 0.0001; **p < 0.001.

Treatment of S. aureus cells with heHb and uliB at their respective MIC values resulted in a marked reduction in bacterial growth, as expected (Figure 4D). However, co-treatment with the antioxidant NAC partially restored bacterial growth, as evidenced by increased OD600 relative to treatments without NAC. This finding suggests that the antimicrobial activity of both compounds is at least in part associated with oxidative stress induction. Notably, the incomplete recovery of growth in the presence of NAC indicates that additional ROS-independent mechanisms also contribute to the bactericidal effects of heHb and uliB.

Evaluation of the antibiofilm activity of H. brasiliense hexane extract and uliginosin B

The effects of heHb and uliB on S. aureus biofilms were also investigated. For the biofilm formation assay, bacterial cultures were incubated for 24 h in the presence of increasing concentrations of heHb and uliB. Three classical biofilm parameters were quantified: total biomass, ECM and metabolic activity. The inhibition of biofilm formation was clearly concentration-dependent. At the highest concentration tested (2×MIC), both heHb and uliB reduced total biomass, ECM production and metabolic activity by more than 50 % (Figures 5A, 5C and 5E). Treatment with heHb at MIC produced inhibitory effects comparable to those observed at 2×MIC. In contrast, uliB at MIC exhibited strain-dependent behaviour, with some isolates still forming biofilms at levels similar to the untreated controls (Figs. 5A and 5B). Interestingly, exposure to sub-inhibitory concentrations (½×MIC and ¼×MIC) of both compounds led to a paradoxical increase in biofilm development, enhancing biomass, ECM content and metabolic activity beyond the levels observed in untreated controls (Figures 5A, 5C and 5E). This stimulatory effect at low concentrations suggests a stress-induced compensatory response commonly reported for S. aureus under sublethal antimicrobial pressure. The ability of heHb and uliB to disrupt mature S. aureus biofilms was also evaluated.

Figure 5.: Antibiofilm activity of heHb and uliB against different strains of S. aureus. Biofilm formation was evaluated during treatment with heHb and uliB at concentrations ranging from 2× to 1/4× MIC (12.5 to 0.78 μg mL-1) by analysing total biomass using crystal violet (A), extracellular matrix production using safranin (C), and metabolic activity using the XTT assay (E). The effects on preformed mature biofilms were assessed at concentrations ranging from 8× - 1× MIC (50 to 3.125 μg mL-1) by measuring total biomass with crystal violet (B), extracellular matrix production with safranin (D) and metabolic activity with XTT (F). Statistical analyses were performed using one-way ANOVA followed by Dunnett’s post hoc test, with a significance level of 95 % (p <0.05). ***p 0,0001; **p < 0.0009; * p < 0.01

Across all tested concentrations, both the extract and the isolated compound significantly reduced total biomass, ECM production and metabolic activity compared with untreated controls (Figures 5B, 5D and 5F). At the highest concentrations (8×MIC, 4×MIC and 2×MIC), biofilm disruption reached approximately 50 to 90 % across several strains, demonstrating strong antibiofilm activity. Treatments at MIC produced strain-dependent responses but still achieved notable reductions, ranging from 25 to 80 %. Overall, these findings indicate that heHb and uliB are effective not only in preventing biofilm formation but also in impairing established biofilms, underscoring their broad antibiofilm potential against S. aureus.

In silico prediction of uliginosin B molecular targets in S. aureus

Based on the mechanistic assays, both heHb and uliB primarily target the metabolic machinery of S. aureus, resulting in a marked reduction in metabolic activity, loss of cell viability and profound, irreversible damage to cellular morphology. Notably, this metabolic impairment was also evident in cells embedded within biofilms, indicating that the compounds effectively compromise bacterial activity in both planktonic and sessile states. These findings reinforce the idea that heHb and uliB act through metabolism-cantered mechanisms rather than through direct membrane disruption. To gain further insight into the molecular basis of this inhibitory effect, in silico analyses were performed to evaluate the binding potential of uliB to key enzymes of the S. aureus electron transport chain, particularly NADH:quinone oxidoreductase type II (NDH-2) and cytochrome bd oxidase. These enzymes are essential for maintaining cellular respiration and redox homeostasis, and their inhibition could plausibly explain the pronounced metabolic collapse observed in vitro.

Molecular docking analyses showed that uliB binds within the quinone-binding pocket of NDH-2, adopting an orientation that overlaps with that of the natural substrate, menadione (Figure 6). In the reference complex, menadione interacts with key residues of the quinone-binding site, including Gln320, Arg350, and Arg385, as well as Met323, Glu327, Ile382, and Lys389, through van der Waals contacts. UliB occupies the same core region, with one of its aromatic rings forming hydrogen bonds with Gln320 and Arg350, in addition to engaging in van der Waals interactions with Ala319, Met323, Ile382, Arg385, Ala386 and Lys389.

Figure 6.: Molecular docking of uliB and the substrate menadione in the NDH-2 enzyme from S. aureus. (A) overlay of the predicted poses for uliB (green) and menadione (purple) complexed with the enzyme from S. aureus and ubiquinone (white) complexed with the homologous enzyme Ndi1 from S. cerevisiae (PDB code 4G73). Intermolecular interactions of (B) uliB and (C) menadione with the S. aureus enzyme. Hydrogen bonds are shown as black dashed lines. The FAD cofactor is shown in orange

Due to its larger and more complex structure, uliB extends further toward the entrance of the binding pocket, where its second aromatic ring establishes additional stabilizing contacts with Gln324, Phe366, Met378 and Val381 (Figure 6). The predicted binding affinity of uliB (-25.104 kJ mol-1 (1 kJ = 0.239 kcal) was similar to that of menadione (-27.614 kJ mol-1), suggesting that uliB may competitively interfere with quinone binding and consequently disrupt electron transport, providing a plausible molecular explanation for the observed reduction in bacterial metabolic activity.

In subunit I of S. aureus cytochrome bd oxidase (CydA), uliB adopted a binding pose that overlapped substantially with that of the reference inhibitor quinestrol, resulting in a comparable interaction pattern (Figure 7). Owing to the predominantly hydrophobic characteristics of the binding pocket, van der Waals forces were the primary contributors to ligand stabilization. Both uliB and quinestrol engaged residues Met235, Leu270, Leu287, Leu294 and Ile331, which form the hydrophobic core of the pocket. In addition to these contacts, quinestrol established a hydrogen bond with Asp239, providing an extra stabilizing interaction absent in the uliB complex (Figure 7). Despite the similarity in binding orientation, quinestrol displayed a more favourable predicted binding energy (-29.29 kJ mol-1) relative to uliB (-20.92 kJ mol-1), indicating a higher theoretical affinity for CydA. This suggests that, although uliB can interact with cytochrome bd oxidase, its inhibitory potential for this enzyme may be lower than that inferred for quinestrol.

Figure 7.: Molecular docking of uliB with cytochrome bd oxidase subunit I (CydA) from S. aureus and comparison with the steroidal inhibitor quinestrol. (A) overlap of the predicted binding mode for quinestrol (cyan) and uliginosin B (green). Intermolecular interactions between the protein and (B) uliB (green) or (C) quinestrol (cyan) are shown. Hydrogen bonds are shown as black dashed lines. The closest heme B group is shown in yellow

Evaluation of the in vitro and in vivo toxicity profiles of H. brasiliense hexane extract and uliginosin B

Toxicity analyses were conducted to assess the safety of heHb and uliB across multiple experimental models. Both compounds exhibited low haemolytic activity, with ≤20 % haemolysis observed at the highest concentrations tested (156.2 μg mL-1 for heHb and 78.12 μg mL-1 for uliB), which correspond to 10 to 20× their MIC values. Because haemolysis did not reach 50 % at any tested concentration, the CC₅₀ for red blood cells could not be determined and is therefore inferred to be greater than these maximum values. Cytotoxicity toward mammalian cells was evaluated using Vero (kidney epithelial) and HaCaT (keratinocyte) cell lines. In Vero cells, the CC₅₀ values were 414.75 μg mL-1 for heHb and 97.6 μg mL-1 for uliB, whereas in HaCaT cells, the CC₅₀ values were 393.57 and 212.26 μg mL-1, respectively. These findings demonstrate that both compounds exhibit markedly greater selectivity for S. aureus than for mammalian cells. In vivo toxicity was assessed using G. mellonella larvae exposed to 100× MIC concentrations (625 μg mL-1 for heHb and 312.5 μg mL-1 for uliB). Larval mortality remained below 20 % throughout the 168 h observation period, with only one death recorded for heHb within 24 h and one for uliB within 72 h. No statistically significant differences were observed compared with the PBS control group, indicating good tolerability in the invertebrate model.

Based on these toxicity results, the SI for S. aureus was calculated using MIC and CC₅₀ values obtained from Vero and HaCaT cell lines (Table 2). All SI values exceeded 10, demonstrating a strong preferential activity of both compounds toward bacteria rather than mammalian cells. For heHb, SI values ranged from 66.4 to 132.9, whereas for uliB they ranged from 15.6 to 68. Because CC₅₀ values could not be determined for haemolysis or G. mellonella due to the minimal toxicity observed, the SI for these assays was summarized in Table 2, further reinforcing the high bacterial selectivity of heHb and uliB.

S. aureus strains | Selectivity index
heHb | uliB
Vero | HaCaT | Erythrocytes | G. mellonella | Vero | HaCaT | Erythrocytes | G. mellonella
ATCC 29213 | 66.4 | 62.98 | >100 | >100 | 31.25 | 68 | >100 | >100
HU 25 | 132.9 | 126.14 | >200 | >200 | 31.25 | 68 | >100 | >100
USA300 | 66.4 | 62.98 | >100 | >100 | 31.25 | 68 | >100 | >100
BMB 9393 | 66.4 | 62.98 | >100 | >100 | 31.25 | 68 | >100 | >100
CR14-005 | 132.9 | 126.14 | >200 | >200 | 31.25 | 68 | >100 | >100
CR14-021 | 132.9 | 126.14 | >200 | >200 | 15.6 | 34 | >50 | >50
CR14-026 | 132.9 | 126.14 | >200 | >200 | 31.25 | 68 | >100 | >100
CD16-016 | 66.4 | 62.98 | >100 | >100 | 31.25 | 68 | >100 | >100

Predicted pharmacokinetic, toxicological and drug-like properties of uliginosin B

To further characterize the pharmacokinetic and toxicological profile of uliB, in silico ADMET analyses were performed and compared with those of the reference antibiotic vancomycin. These assessments provide additional insight into the drug-likeness and potential safety of uliB as a therapeutic candidate. UliB demonstrated favourable pharmacokinetic attributes, including good human intestinal absorption (Table 3). It also complied with Lipinski’s “Rule of Five” [31], indicating a high likelihood of good oral bioavailability; unlike vancomycin, which, despite its clinical efficacy, is not orally bioavailable. Similar to vancomycin, uliB was not predicted to cross the blood-brain barrier under physiological conditions. With respect to P-glycoprotein (PgP), uliB was predicted to act as an inhibitor but not as a substrate (Table 3), a characteristic that may enhance its potential for synergistic interactions when combined with other antimicrobials. Toxicity predictions indicated that uliB is non-genotoxic, non-carcinogenic, and non-irritant to the skin and eyes, paralleling the profile of vancomycin. However, uliB showed positive predictions for hepatotoxicity and respiratory toxicity, whereas vancomycin was predicted to exhibit only respiratory toxicity.

Predicted endpoint | Uliginosin B | Vancomycin
HIA | Absorbed | Non-Absorbed
Lipinski rule of five | Approved | Rejected
BBB | Non-Penetrable | Non-Penetrable
PgP substrate | Non-Substrate | Non-Substrate
PgP inhibitor | Inhibitor | Non-Inhibitor
CYP1A2 substrate | Substrate | Non-Substrate
CYP2C9 substrate | Non-Substrate | Non-Substrate
CYP2C19 substrate | Substrate | Non-Substrate
CYP2D6 substrate | Non-Substrate | Non-Substrate
CYP3A4 substrate | Substrate | Non-Substrate
CYP1A2 inhibitor | Inhibitor | Non-Inhibitor
CYP2C9 inhibitor | Non-Inhibitor | Non-Inhibitor
CYP2C19 inhibitor | Non-Inhibitor | Non-Inhibitor
CYP2D6 inhibitor | Non-Inhibitor | Non-Inhibitor
CYP3A4 inhibitor | Inhibitor | Non-Inhibitor
Pfizer rule 3/75 | Approved | Approved
Genotoxicity | Low risk | Low risk
Carcinogenicity | Low risk | Low risk
Nephrotoxicity | Low risk | Low risk
Cardiotoxicity | Low risk | Low risk
Hepatotoxicity | High risk | Low risk
Skin irritation | Low risk | Low risk
Eye irritation | Low risk | Low risk
Respiratory toxicity | High risk | High risk