Experimental

Plant material

Hypericum brasiliense was collected in the municipality of Trajano de Moraes, Rio de Janeiro (22°12′17″ S, 43°11′35″ W), under authorization from SISBIO/ICMBio (no. 13659-21) and SisGen (A491A56). Botanical identification was performed by Dr. Marcelo Guerra (State University of Rio de Janeiro, UERJ), and voucher specimens were deposited in the Herbarium of the Faculty of Teacher Training (FFP/UERJ) to ensure traceability of the plant material. The whole plant was ground in a hammer mill, and the resulting powder was subjected to static maceration with hexane, following previously described procedures [11]. After extraction, the solvent was filtered and evaporated under reduced pressure using a rotary evaporator to obtain the crude extract. The extract was subsequently lyophilized and designated as the hexane extract of H. brasiliense (heHb), and the isolation of japonicin A, isouliginosin B and uliginosin B (uliB) was performed according to França et al. [9].

Bacterial strain lineages

The bacterial panel used in this study included the methicillin-susceptible S. aureus (MSSA) reference strain ATCC 29213, obtained from the American Type Culture Collection (USA). Methicillin-resistant S. aureus (MRSA) strains, including USA300 and the Brazilian epidemic clone (BEC) HU25, were generously provided by Dr. Fabio Aguiar Alves (Fluminense Federal University, Brazil, and Palm Beach Atlantic University, USA). Additional clinical isolates (CR14-005, CR14-021, CR14-026 and CD16-016), as well as the BEC strain BMB 9393, were kindly supplied by Dr. Agnes Marie Sá Figueiredo/Dr. Bernadete Teixeira Ferreira Carvalho (Federal University of Rio de Janeiro, Brazil). These clinical strains were originally recovered from patients at the Clementino Fraga Filho University Hospital of UFRJ, Brazil [14]. Among the clinical isolates, CR14-005 exhibited resistance to clindamycin, erythromycin, chloramphenicol, ciprofloxacin and cefoxitin; CR14-021 showed resistance only to cefoxitin; CR14-026 was resistant to erythromycin, gentamicin and cefoxitin, with intermediate susceptibility to chloramphenicol; and CD16-016 presented resistance to clindamycin, erythromycin, chloramphenicol, ciprofloxacin and cefoxitin. All strains were stored in brain heart infusion (BHI) broth with 1 % glycerol at -20 °C.

Evaluation of the antibacterial activity of H. brasiliense hexane extract and its purified compounds

For the antimicrobial susceptibility assays, serial dilutions of heHb and the isolated compounds uliB, isouliginosin B and japonicin A were prepared following CLSI M100-Ed33 guidelines in 96-well microtiter plates containing Mueller-Hinton broth, yielding final concentrations ranging from 100 to 0.78 μg mL-1. The bacterial inoculum was adjusted to a 0.5 McFarland standard (≈108 colony-forming units (CFUs) mL mL-1) and subsequently diluted to obtain a final concentration of 104 CFU mL-1 in each well. Plates were incubated for 24 h at 37 °C. Bacterial viability was assessed visually by inspecting the turbidity of each well, and the last well showing no visible growth was considered the minimum inhibitory concentration (MIC). To determine the minimum bactericidal concentration (MBC), 10 μL aliquots from wells of the MIC assay were plated onto Mueller-Hinton agar and incubated for 24 h at 37 °C. The MBC was defined as the lowest concentration that resulted in no visible bacterial growth [15].

Assessment of the antibacterial activity of H. brasiliense hexane extract and uliginosin B under varying bacterial densities

To assess the inoculum effect on metabolic activity, bacterial suspensions with initial concentrations of 104, 10⁵, 10⁶, 107 or 108 CFU mL-1 were cultured in 96-well plates with or without heHb and uliB at MIC values for 24 h at 37 °C. Following incubation, the 2,3-bis (2-methoxy-4-nitro-5-sulphenyl)-(2H)-tetrazolium-5-carboxanilide (XTT) and menadione solution were added to quantitatively assess cellular metabolic activity via tetrazolium reduction. The plates were further incubated for 3 h at 37 °C. Metabolic activity was finally quantified by measuring absorbance at 492 nm using a microplate reader Multiskan SkyHigh spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA).

Effects of H. brasiliense hexane extract and uliginosin B on bacterial ultrastructure

Microscopy analyses were performed following the standardized procedures of the Microscopy Unit (UniMicro, UFRJ). S. aureus strains ATCC 29213 and USA300, previously cultured on tryptic soy agar (TSA), were inoculated into BHI broth, adjusted to 108 CFU mL-1, and incubated in the absence or presence of heHb and uliB at their respective MIC values (6.25-3.125 μg mL-1) for 24 h at 37 °C. After incubation, supernatants were removed, and the cells were washed once with phosphate-buffered saline (PBS, pH 7.2). For scanning electron microscopy (SEM), bacterial samples were fixed for 1 h at room temperature in a solution containing 25 % glutaraldehyde, 0.2 M sodium cacodylate buffer, and Milli-Q water. Following three washes with 0.1 M sodium cacodylate buffer, cells were post-fixed with 1 % osmium tetroxide for 1 h and washed again with the same buffer. Dehydration was performed through a graded ethanol series (30, 50, 70, 90 and 100 %), followed by critical point drying with CO₂. Samples were then sputter-coated and examined using a Thermo Fisher Quattro S scanning electron microscope. For transmission electron microscopy (TEM), the protocol was followed up to the post-fixation step. Samples were subsequently dehydrated through a graded acetone series (30, 50, 70, 90 and 100 %), infiltrated with Spurr resin at room temperature, and polymerized at 68 °C for 72 h. Ultrathin sections were stained with uranyl acetate and alkaline lead citrate (5 to 10 min each) and visualized using a FEI Tecnai Spirit Bio-Twin transmission electron microscope.

Effects of H. brasiliense hexane extract and uliginosin B on bacterial plasma membrane integrity

Bacteria grown overnight on TSA were adjusted to a final concentration of 107 CFU mL mL-1 [16]. Cell suspensions were then incubated for 3 h in the presence or absence of heHb and uliB at concentrations ranging from 4×MIC to ½×MIC (25 to 1.56 μg mL mL-1). Boiled cells were included as a positive control for membrane disruption, untreated cells incubated with the probe served as a negative control, and unstained cells were used to assess autofluorescence. Following incubation, samples were centrifuged at 4,000 rpm for 10 min, washed once with PBS, and stained with propidium iodide (PI, 1 μg mL-1) for 10 min at 37 °C in the dark [17]. PI fluorescence was measured in the FL3 channel using a BD LSRFortessa™ flow cytometer, and at least 10,000 events were recorded per sample to ensure statistical robustness. Data were analyzed using Flowing Software version 2.5.1, and fluorescence intensity was visualized in side-scatter (SSC) versus PI plots to quantify the proportion of PI-positive cells, which indicate membrane-compromised populations.

Effects of H. brasiliense hexane extract and uliginosin B on bacterial metabolic activity

Bacterial cells grown overnight on TSA were inoculated into BHI broth, adjusted to 107 CFU mL mL-1, and exposed to heHb and uliB at concentrations ranging from 4×MIC to ½×MIC (25-1.56 μg mL-1) for 3 h at 37 °C. Untreated cells were used as the positive control [18]. Following incubation, supernatants were discarded, cells were washed once with PBS, and an XTT solution (200 μg mL-1) supplemented with menadione (0.4 mM) was added. Samples were then incubated for an additional 3 h at 37 °C. Metabolic activity was quantified by measuring absorbance at 492 nm using a Multiskan SkyHigh spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA).

Effects of H. brasiliense hexane extract and uliginosin B on bacterial induction of reactive oxygen species

Bacterial cells grown overnight in TSA were adjusted to a 0.5 McFarland standard (≈108 CFU mL-1) in BHI broth and exposed to the MIC concentrations of heHb and uliB (6.25 and 3.125 μg mL-1) for 30 min at 37 °C. The reactive oxygen species (ROS) assays used a 30 min incubation to capture early oxidative responses that would not be accurately represented at longer time points. After treatment, cells were incubated for an additional 30 min with 10 μM of the probe 2’,7’-dichlorodihydrofluorescein diacetate (H₂DCF-DA; Sigma-Aldrich, USA) [19]. Cells treated with hydrogen peroxide (50 mM) served as positive control, while untreated probe-incubated cells served as a negative control, and unlabelled cells were included to account for autofluorescence [20]. Fluorescence measurements were obtained using black 96-well plates (excitation 488 nm, emission 535 nm) on a SpectraMax M3 microplate reader (Molecular Devices). To normalize fluorescence signals for variations in bacterial density, cultures were subsequently transferred to transparent plates for absorbance measurements at 600 nm. In parallel, the effect of the antioxidant N-acetylcysteine (NAC) on bacterial growth (ATCC 29213) at the MIC was evaluated by measuring optical density at 600 nm (OD600). Bacterial suspensions prepared as described above were exposed to the previously determined MIC values of heHb and uliB in the presence or absence of NAC (final concentration of 5 mM) and incubated at 37 °C for 18-24 h. Following incubation, OD600 was measured using the Multiskan SkyHigh spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) to assess bacterial growth. Comparisons between conditions with and without NAC were used to determine the influence of antioxidant supplementation on bacterial growth inhibition at the MIC level.

Effects of H. brasiliense hexane extract and uliginosin B on bacterial biofilm formation and disruption

S. aureus was initially cultured on TSA for 24 h at 37 °C and subsequently transferred to TSB supplemented with 1 % glucose for an additional 24 h. The cultures were then adjusted to a final density of 108 CFU mL mL-1. For biofilm formation assays, standardized bacterial suspensions were inoculated into 96-well polystyrene microtiter plates and incubated with heHb and uliB at concentrations ranging from 2×MIC to ¼×MIC (12.5 to 0.78 μg mL-1). Untreated cells served as positive controls. To evaluate biofilm disruption, mature 24 h pre-formed biofilms were exposed to higher compound concentrations, ranging from 8×MIC to MIC (50 to 3.125 μg mL-1), for an additional 24 h. In both experimental setups (biofilm formation and disruption), the well contents were gently aspirated, the wells were washed once with PBS, and the remaining biofilms were subsequently analysed for three classical parameters: biomass, metabolic activity (viability), and extracellular matrix (ECM) content. Biomass was quantified in methanol-fixed biofilms using crystal violet staining, with absorbance measured at 590 nm on a Multiskan SkyHigh spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) [21,22]. ECM content was assessed in non-fixed biofilms at 530 nm following safranin incorporation [23,24]. Metabolic activity was evaluated in non-fixed biofilms using the XTT reduction assay (200 μg mL-1 XTT with 0.4 mM menadione), with absorbance measured at 492 nm [25].

Effects of uliginosin B on bacterial oxidative phosphorylation proteins: a molecular docking approach

The three-dimensional structure of S. aureus NADH:quinone oxidoreductase type II (NDH-2) was retrieved from the Protein Data Bank (PDB ID: 5NA1). As no experimental structure is available for S. aureus cytochrome bd oxidase, a structural model of subunit I (CydA) predicted by AlphaFold2 (A0A2S6D6T4) was employed. Heme cofactors were incorporated via structural alignment with the E. coli cytochrome bd oxidase (PDB ID: 7OSE), followed by energy minimization in Swiss-PdbViewer 4.1. The resulting CydA model displayed satisfactory stereochemical quality, with 94.6 % of residues located in favoured regions of the Ramachandran plot. Ligand structures (uliB, quinestrol and menadione) were obtained from PubChem, geometry-optimized using OpenBabel 3.1.1, and subsequently energy-minimized under the MMFF94 force field. Partial atomic charges were assigned using the EEM method at the DFT-B3LYP/6-311G/NPA level. Molecular docking simulations were performed using AutoDock Vina 1.1.2, treating protein structures as rigid and ligands as fully flexible. Docking grids were cantered on previously characterized quinone/quinol binding sites. Resulting protein-ligand complexes were visualized and analysed using PyMOL and Discovery Studio Visualizer 2021 (Dassault Systèmes BIOVIA, San Diego, CA).

Effects of H. brasiliense hexane extract and uliginosin B on the toxicity of mammalian cell lineages

The HaCaT (human keratinocyte) and Vero (monkey kidney epithelial) cell lines were maintained in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10 % foetal bovine serum (FBS) at 37 °C in a 5 % CO₂ atmosphere. Mammalian cells (10⁵ cells per well) were first seeded in 96-well tissue culture plates and allowed to adhere for 4 h under standard culture conditions. Non-adherent cells were then removed by gentle washing with sterile DMEM, after which the wells were replenished with fresh DMEM containing 10 % FBS. Cells were exposed to increasing concentrations of heHb and uliB (7.81-500 μg mL-1) and incubated for an additional 24 h at 37 °C in a 5 % CO₂ atmosphere. Following treatment, the culture medium was discarded and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was added to each well (25 μg per well). Plates were incubated for 3 h in the dark at 37 °C, centrifuged at 500 × g for 8 min, and the supernatant removed. The resulting formazan crystals were dissolved in 200 μL of DMSO, and absorbance was measured at 570 nm using a SpectraMax M3 microplate reader (Molecular Devices). The 50 % cytotoxic concentration (CC₅₀) was calculated by nonlinear regression analysis [26].

Effects of H. brasiliense hexane extract and uliginosin B on erythrocyte lysis

Two-percent sheep erythrocytes (Cultilab, Rio de Janeiro, RJ) were incubated in 96-well plates with serial dilutions of heHb and uliB, starting at 100× MIC (625-2.44 μg mL-1), for 3 or 24 h at 37 °C. Following incubation, the plates were centrifuged at 2,100 rpm for 10 min, and the supernatants were collected for absorbance measurement at 415 nm using a Multiskan SkyHigh spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) [27]. Triton X-100 (0.1 %) was used as the positive control for complete haemolysis, whereas PBS served as the negative control.

Effects of H. brasiliense hexane extract and uliginosin B on Galleria mellonella larval survivability

Galleria mellonella larvae weighing approximately 0.2-0.3 g and displaying clear, uniform coloration were selected for the in vivo toxicity assay. Experimental groups consisted of 10 larvae each, including a negative control group injected with PBS. The compounds heHb (625 μg mL-1) and uliB (312.5 μg mL-1), corresponding to 100× MIC, were tested by injecting 10 μL of each solution into the last left proleg using an insulin syringe (BD Ultra-Fine, Franklin Lakes, NJ, USA). Following injection, the larvae were incubated at 37 °C for 168 h. Larval survival was monitored daily. Death was determined by the absence of movement in response to gentle physical stimuli applied to the head and body, as well as by the presence of extensive melanisation. Throughout the entire experimental period, all larvae were maintained under controlled conditions at 37 °C.

Determination of the selectivity index for H. brasiliense hexane extract and uliginosin B

The selectivity index (SI) of heHb and uliB was determined as the ratio between their MIC values against S. aureus and their CC₅₀ values in mammalian cells and in G. mellonella larvae. Compounds exhibiting SI values greater than 10 were considered promising candidates for further investigation [28].

In silico analysis of pharmacokinetic, toxicological and drug-like properties of uliginosin B

The pharmacokinetic, toxicological, and drug-likeness properties of uliB were predicted using computational approaches, with vancomycin as the reference compound. Human intestinal absorption, blood-brain barrier permeability and P-glycoprotein interactions were evaluated using the admetSAR 3 server [29]. The Deep-PK platform was used to predict potential substrate or inhibitory activity toward major CYP450 isoforms, including CYP1A2, CYP2C9, CYP2C19, CYP2D6 and CYP3A4 [30]. Drug-likeness was assessed according to Lipinski’s “Rule of Five” and the Pfizer 3/75 rule, which provide estimates of oral bioavailability and preclinical toxicity risk [31]. Predicted toxicological endpoints, including genotoxicity, carcinogenicity, hepatotoxicity, nephrotoxicity, cardiotoxicity (hERG inhibition), respiratory toxicity and skin or eye irritation, were also evaluated through admetSAR 3.

Statistical analysis

All experiments were performed at least three times in biological triplicate. Triplicate values for each sample were averaged, and results were analysed by one-way ANOVA with a 95 % confidence level (p < 0.05), using Dunnett's post hoc test to compare treated samples with untreated controls or Sidak for multiple comparisons. In vitro toxicity data were analysed using Student’s t-test with a 95% confidence level (p < 0.05). In vivo toxicity was assessed using the Mantel-Cox log-rank test. All statistical analyses were performed using GraphPad Prism 8.0.1 [25].