Discussion

The increasing prevalence of antimicrobial resistance continues to outpace the development of effective therapeutics, reinforcing the need for new compounds capable of overcoming resistance mechanisms in S. aureus [32]. In this study, the heHb exhibited potent anti-staphylococcal activity, with an MIC of 3 μg mL-1, while its constituent, uliB, showed comparable inhibitory effects. These findings align with previous reports describing the antibacterial potential of heHb and uliB, which demonstrated activity within the same concentration range [12,13]. Moreover, other Hypericum species, such as H. beanii, H. calycinum, H. foliosum, H. hircinum, H. lagarocladum, H. olympicum and H. revolutum, have been reported to inhibit S. aureus at similar or higher concentrations (64 to 256 μg mL-1), supporting the genus as a rich source of bioactive metabolites [33]. The results obtained here further confirm the promising antibacterial profile of heHb and provide novel insights into its mechanism of action. By linking inhibitory activity with alterations in bacterial metabolism, oxidative stress induction and morphological damage, this work expands the current understanding of how heHb and uliB constituents act against resistant S. aureus strains, highlighting their potential as prototypes for new antimicrobial agents.

The inoculum effect, defined as the increase in MIC values with higher bacterial densities [34], is an important factor influencing antimicrobial efficacy. During infection, bacterial loads vary considerably across tissues and high-density infections, such as biofilms, abscesses or bloodstream, and chronic infections, pose major therapeutic challenges [35]. This phenomenon may result from reduced drug availability or induced bacterial tolerance [36]. In this context, both heHb and uliB maintained a consistent inhibitory effect even at elevated inoculum levels, suggesting that their antimicrobial activity is less affected by cell density, an advantageous property for treating infections associated with high bacterial loads.

SEM revealed that untreated S. aureus ATCC 29213 and USA300 cells exhibited intact and well-defined morphologies, whereas treatment with heHb and uliB caused pronounced surface alterations, including depressions, vesicle formation and roughened cell walls. Similar structural damage has been associated with the action of plant-derived antimicrobials such as citral and trans-cinnamaldehyde, which induced deformation and irregular cocci formation in S. aureus [37], and with extracts from Phyllanthus emblica and Lycium shawii, which caused extensive cell lysis and leakage of intracellular contents [38]. The rough and nodular cell surface observed after treatment with heHb and uliB resembles the morphological effects reported for rhamnolipid-treated S. aureus [39], which may reflect alterations in cell wall maintenance similar to those described previously, where S. aureus mutants lacking autolysin (atl) exhibited division defects and increased surface roughness [40].

TEM complemented these findings, showing loss of the characteristic spherical shape, septal malformation and cytoplasmic disorganization in treated cells. The alterations resembled those described for S. aureus exposed to Hymenaea stigonocarpa extracts, which disrupted cell division and induced membrane rupture [41]. Treatments with uliB led to additional structural changes, including increased cell wall thickness and surface depressions, similar to those observed in S. aureus mutants lacking lcpC, a gene essential for peptidoglycan assembly [42]. Together, these results suggest that heHb and its constituent uliB compromise the integrity of the S. aureus cell envelope, potentially by interfering with cell wall synthesis and remodelling.

Consistent with the observed morphological alterations, previous studies have shown that plant-derived compounds can affect S. aureus cell integrity through distinct mechanisms. For instance, Zeng et al. [43] and Dai et al. [44] reported extensive membrane permeabilization in S. aureus following treatment with Polygonum chinense and citral, with more than 60 to 90 % of cells staining positive for propidium iodide. In contrast, the low membrane labelling observed for heHb and uliB suggests a milder or indirect impact on the plasma membrane, indicating that their antimicrobial activity likely involves other cellular targets. In this sense, comparable effects on the metabolic activity of S. aureus have been reported for several plant-derived products that interfere with the bacterial electron transport chain. Al-Bakri and Afifi [45] observed that extracts rich in alkaloids, terpenoids and phenolics reduced cellular metabolism, suggesting that these secondary metabolites can impair bacterial energy production. The reduction in metabolic activity observed for heHb and uliB may therefore be associated with a similar mechanism, potentially involving interference with the electron transport chain or related metabolic pathways rather than direct damage to the cell membrane.

S. aureus possesses efficient defence systems against oxidative stress, mainly through catalase, superoxide dismutase and the carotenoid pigment staphyloxanthin, which confers antioxidant protection and contributes to bacterial survival during infection [20,46]. The induction of ROS observed after exposure to heHb extract and uliB suggests that these treatments may disrupt this protective balance, promoting oxidative damage. A similar mechanism has been described for bactericidal drugs, which trigger hydroxyl radical formation through the Fenton reaction and alter metabolic redox processes [47], supporting the hypothesis that oxidative stress plays a key role in the antimicrobial action of these compounds. Consistently, the partial restoration of bacterial growth in the presence of the antioxidant N-acetylcysteine observed in our study supports the involvement of ROS in the antimicrobial effect and indicates that oxidative stress is not the sole mechanism underlying the compounds' activity.

Biofilm-associated infections pose a major therapeutic challenge because the extracellular matrix and altered metabolic states reduce antimicrobial penetration and efficacy [48]. In our study, heHb and uliB both inhibited biofilm formation, yet sub-MIC exposure promoted biofilm induction. This response is consistent with adaptive mechanisms described under sub-inhibitory antimicrobial pressure, including increased exopolysaccharide synthesis and enhanced adhesiveness, and may also be linked to stress-response pathways triggered by metabolic imbalance and oxidative stress [49,50]. Importantly, mature biofilms were susceptible to disruption only at supra-MIC concentrations, consistent with literature showing that higher drug levels are generally required to disassemble established biofilm architecture [51]. These biofilm dynamics align with our additional findings: the compounds induced a pronounced, dose-independent suppression of metabolic activity and increased ROS production, while causing minimal membrane permeabilization. Together, this pattern supports a model in which interference with cellular energy metabolism and redox homeostasis, rather than primary membrane disruption, contributes to both the inhibition of biofilm formation, while ensuring sufficiently high drug concentrations to effectively disrupt established biofilms and the susceptibility of mature biofilms to higher drug pressure. Clinically, these results reinforce the need for dosing strategies that avoid prolonged sub-inhibitory exposure, which may promote biofilm formation, while achieving sufficiently high concentrations to disrupt established biofilms.

Given experimental evidence that uliB affects the energy metabolism of S. aureus, molecular docking was performed to explore its potential interaction with two key enzymes of the bacterial electron transport chain: NDH-2 and CydA. Since S. aureus relies on NDH-2 and cytochrome bd oxidase for respiration, both absent in mammalian cells, these enzymes represent selective targets for antimicrobial development [52,53]. The results indicated that uliB binds to NDH-2 in a mode like ubiquinone and the known inhibitor myricetin [54], suggesting competitive inhibition at the quinone-binding site. Likewise, the compound showed favourable binding to CydA, forming interactions comparable to those of reported inhibitors such as auraquine D [55], although with lower affinity than quinestrol. Taken together, these findings are consistent with NDH-2 as a putative molecular target of uliB and are in line with the experimental evidence of disrupted energy metabolism and reduced bacterial viability.

The toxicity of test compounds on in vitro and in vivo models was investigated. The haemolytic activity of heHb and uliB was evaluated since erythrocytes are a suitable model for initial cytotoxicity screening [56]. Both showed less than 10 % haemolysis after 24 h at 156.2 and 78.12 μg mL-1, respectively, indicating low toxicity, similar to findings for other plant extracts such as Chamaemelum nobile and Mentha pulegium [57]. Although plant metabolites can occasionally induce erythrocyte rupture [58], the heHb effect appeared to be mild and possibly time-dependent [59]. To complement these results, cytotoxicity was assessed in Vero and HaCaT cells. UliB showed higher CC₅₀ values than the extract (97.6 and 212.26 μg mL-1, respectively), and the SI indicated greater antibacterial selectivity toward S. aureus than cytotoxicity toward epithelial cells. These values exceeded those reported by Akinboye et al. [60] for Erythrina caffra (SI = 3.23-8.55) and by Fontanay et al. [61] for triterpenes with low SI (<10), reinforcing the safety and selectivity of both natural products. Considering the ethical principles of the 3Rs (replacement, reduction and refinement) in toxicity testing [62], Galleria mellonella larvae were used as an in vivo alternative model. Neither heHb (625 μg mL-1) nor uliB (312.5 μg mL-1) caused mortality within 168 h, supporting their low toxicity, in agreement with results for A. colubrina [63] and R. officinalis [64]. Together, these findings indicate that both the extract and uliB exhibit promising biological activity with minimal toxicity in in vitro mammalian cells and in vivo in G. mellonella larvae.

Given that pharmacokinetic limitations and toxicity issues are among the primary causes of drug development failure [65], the pharmacokinetic and toxicity profile of uliB was investigated in silico approaches as a preliminary assessment of its potential as an antimicrobial candidate. The predictions suggest that uliB may have properties consistent with oral bioavailability, which could be advantageous for treatment adherence and cost reduction [66], although these findings require experimental validation. Among the transporters analysed, P-glycoprotein (PgP), a mammalian efflux pump, plays a key role in modulating the pharmacokinetic properties of many compounds, including antibiotics. PgP can limit drug accumulation at the target site by decreasing cellular availability and exposure time, thereby reducing antimicrobial efficacy [67]. Based on in silico predictions, uliB may interact with PgP; however, the functional relevance of this interaction remains to be experimentally determined. Additionally, considering that mammalian and bacterial efflux pumps may share substrates, and that some natural products exhibit dual inhibitory activity, uliB may potentially display similar behaviour. Nevertheless, this hypothesis remains speculative and warrants further investigation, particularly in the context of neuroinfections.