Before starting the experiments with the mentioned drugs, the stability of the nanoemulsions is verified by obtaining their size distributions and zeta potentials using dynamic light scattering (see details in Section 2.6). The suspensions were diluted by 10x before the measurements, to protect the electrodes of the zeta potential cells.

All the size distributions of the nanodroplets were predominantly centered at approximately 50 nm (Figure 1A) and maintained their sizes for at least 15 days, which was enough time to perform the experiments (see Figure 1B). The negative zeta potential is the reason why coalescence is negligible (the droplets repel each other by an electrostatic force) (see Figure 1C). To explain the origin of the negative charge of the droplets, let us remember that olive oil is a composition of fatty acids such as linoleic, stearic, palmitic, and linolenic acids (Boskou et al., 2006). Since the pK of the carboxyl groups contained in their structure is very low (approximately 3.5), and the continuum phase of the nanoemulsion is deionized water (pH approximately 6), deprotonation occurs, being the cause of the negative z potential (Yang et al., 2013).

By evaluating the effect of the nanoemulsions on the model membrane of PE/PG as a E. coli-like-membrane and PG/CL as a S. aureus-like-membrane, it was noted that, compared to the calorimetric profiles of the control, the olive oil nanocarriers either produced a small shift to higher temperatures (due to the inclusion of the fatty acids in the membrane that increases molecular packing and therefore stiffness) (Figure 1D, or did not so, Figure 1E, see Table 1). Therefore, it can be presumed that they are harmless to the profiles and good choices for preparing the nanoemulsions. Indeed, the effects of PFL-OO and CBD-OO (1% v/v), and CBD-OO (3% v/v) are clearly opposite: The profiles stretch, deform, and shift at lower temperatures, indicating strong fluidization and loss of membrane cooperativity (Figures 1D, E). The shoulders in the profiles are associated with a phase separation due to the presence of different domains, regions where the drugs have not been distributed homogeneously (Korkmaz and Severcan, 2005). Although both drugs strongly affect the thermodynamic properties of the vesicles, and therefore compromised the integrity of the membranes, it is remarkably that PFL-OO 1% v/v has a stronger effect than CBD-OO 3% v/v. In fact, PFL-OO 1% v/v moved the T_m 5°C more than CBD-OO at 3% v/v (see Table 1). The widths of the profiles shown in Figures 1D, E were obtained by the standard deviations of three repetitions.

To illustrate the calorimetric changes described above using a simplistic scheme, the cartoon depicted in Figure 2 may be useful. Indeed, the observed changes in the cohesion enthalpies produced by the drugs give rise to a general disorder in the membranes.

It is important to remark that the effects generated by CBD and PFL (see their molecular structures in Figure 1F) on both PE/PG and PG/CL are driven by enthalpic and entropic forces. This indicates that the intrusion of external molecules into the bilayer affects the cohesive energy. Since the 1% v/v and 3% v/v nanoemulsions in the vesicle suspensions corresponded to 2.75 mg/mL (1% v/v) or 8.25 mg/mL (3% v/v), respectively, propofol has a higher number of molecules owing to its lower molecular weight, which could explain its stronger effect on the membrane integrity. However, it is important to note that these concentrations translate into 15.42 mM for PFL-OO (1% v/v), and 8.74 and 26.22 mM for CBD-OO (1% v/v and 3% v/v), respectively. As a result, the amount of molecules is not the unique reason for its greater effect. In addition, since CBD is much more hydrophobic than PFL, hydrophobicity is not the reason for the effect either.

The disrupting effect of PFL has been widely studied before (Pérez-Isidoro et al., 2014; Paiva et al., 2012). Dynamic molecular simulations revealed that PFL exhibits a preference for localizing near the hydrocarbons tails of the phospholipids (Hansen et al., 2013). There, the hydroxyl group of the propofol interacts with their ester oxygens, generating perturbations that cause the lipids to go through the gel-liquid phase with less energy. Similarly, the CBD molecule contains two hydroxyl groups that can interact with the ester oxygen, and the rest of the molecule is oriented toward the hydrophobic region. Furthermore, it is important to note that the drugs were carried by nanodroplets. Consequently, adding hydrophobic molecules to the dispersed phase (olive oil) prevents them from dissolving in the continuous phase (water). Yet, PFL could diffuse better than CBD to reach the lipid bilayer. In addition, it has been shown that, despite being more soluble in octanol and therefore more likely to diffuse into lipid membranes, molecules with logP greater than 5 have reduced penetration in cells (Lipinski et al., 2012).

If a simple membrane model like the one discussed above undergoes a significant change in its structural condition when interacting with PFL and CBD, which completely modify the transition temperature and enthalpy of cohesion, as shown in Figures 1D, E, a biological membrane may also suffer similar modifications in its membrane integrity, compromising the organism viability (see Figure 2).

To test this hypothesis, the effect of the nanoemulsion on both bacteria was evaluated by assessing the growth behavior of colonies in agar plates. This assay consisted in proving first the harmless effect of the vehicle and then performing independent test of the bactericide effects of the drugs (see Figures 3A, B). To see whether the incubation time was an important factor, the CFU was measured at different periods of time. It is clear that propofol, prepared at 1% v/v concentration, has a fully annihilation effect in E. coli after 10 h and in S. aureus after 8 h. On the other hand, CBD at 1% v/v produced only a mild inhibition effect on the growth of bacteria compared to the control, but at 3% v/v reduced a bit more than two decades in a log scale, which means a drop in viability better than 99%.

Our findings contrast with previous studies on propofol emulsions, where it has been reported that they promote bacterial growth when dissolved in soybean oil. Hence, it seems that nanoemulsions are susceptible to contamination (Baker et al., 2005). For such a reason, efforts to reformulate the colloidal suspensions have been undertaken, for example, the use of other oily vehicles and/or the addition of ethylenediaminetetraacetic acid (EDTA) (Sakuragi et al., 1999; Wang et al., 2007; Shevalkar et al., 2019). In contrast, our results showed that, carried by olive oil nanodroplets, PFL at 1% v/v strongly inhibits the growth of E. coli and S. aureus bacteria.

Some action mechanisms have been studied for the bactericidal action of hydrophobic molecules (Blaskovich et al., 2021). It is proposed that one of the potential unexplored targets could be the composition of the outer lipid bilayer because the membrane is the first component found between the extracellular medium and the cells. Indeed, to understand our findings beyond the usual conception behind antibiotic strategies, which are mostly based on attacking specific biochemical targets, or in creating membrane pores that shortcut the membrane potential, it is important to discuss a recent paper by some of us (Salinas-Almaguer et al., 2022). It was studied the effect produced by pentanol as a membrane softening by bilayer disruption. Similar to the damage produced by fluidization (Royce et al., 2013), or lethal permeability upsurge (Vaara, 1992), it is shown that, after a long induction phase in which the microorganisms are striving against a bacteriostatic action, the cells cannot proliferate due to lipid packing distortion which results in further delocalization of membrane proteins impairing multiple essential processes including respiration (Strahl and Hamoen, 2010), nucleotide synthesis (Wenzel et al., 2014), FtsZ assembly (Silber et al., 2020; Mileykovskaya et al., 1998), and nucleoid segregation (Mileykovskaya et al., 1998; Chai et al., 2014), with the collateral effect to harm the cytokinetic apparatus so as to prevent and/or delay the membrane constriction need for normal cell division and offspring generation. In summary, a membrane disruptor such as pentanol, or propofol, compromises the crucial role of membrane rigidity as the key regulator of bacterial proliferation. Other studies of essential oil nanoemulsions have found toxic effects on bacteria (Moghimi et al., 2016; Donśı et al., 2012; Terjung et al., 2012), although the mechanisms involved are attributed to perturbation and cellular leakage caused by the hydrophobic components of the nanodroplets (Seow et al., 2014).

As above mentioned, since PFL is much less hydrophobic than CBD, it can be more easily solubilized in the medium. After the exogenous molecules are internalized in the membrane, they alter its correct functioning. The greater their ability to diffuse in the membrane, the greater the cellular damage inflicted.

Since it is crucial to investigate whether a lower dose could give similar results in terms of bacterial inhibition, experiments to find the minimal inhibitory concentration (MIC) were performed. First, the results show that the most effective carrier was PFL-OO, compared to PFL-DMSO (see Figures 3C, D). Second, the MIC sensitivity in both bacteria was 300 μg after 12 h of incubation, much lower than the concentration used above in the CFU assay (2.75 mg/mL). Let us remark though that MIC assays are performed in liquid media Luria-Bertani (LB) Broth and the CFUs on LB agar.

As a crucial probe to confirm that the morphology of bacteria after the treatment is similar to the one depicted in Figure 2, AFM images were captured as shown in Figure 4. Let us remark that we focused only on the effect produced by propofol, which was the drug that produced the greatest effects on the viability of the organisms.

In Figures 4A, E, it is observed the typical rod-shaped and spherical structures of E. coli and S. aureus, whose average sizes are shown in Figures 4D, H, which is in agreement with previous reports (Meincken et al., 2005; Yao et al., 2012; Harris and Theriot, 2016). Likewise, the cells with DMSO did not exhibit great deformations on their membranes, indicating that DMSO (0.5% v/v) was not harmful (see Figures 5B, F). It is important to remark that the growth of the colonies in the experiments reported in Figure 3 was not assessed, not only because so little concentration of DMSO is innocuous (Lim et al., 2012; Summer et al., 2022), but also because the evolution of the optical density for both bacteria, with and without DMSO, was the same.

The cells with PFL were notoriously different, as observed in Figures 4C, G. It is obvious that these changes in the membrane compromise the viability of the cells, explaining the results shown in Figure 3. Therefore, the mechanistic reason for bacterial death is simply that this hydrophobic molecule completely affects the entire membrane, inflicting drastic damages in the regulation of bacterial proliferation.

It is enthralling that S. aureus bacteria treated with DMSO/PFL not only lose the spherical morphology but also give rise to small vesicles not observed in the control sample, probably due to homeostatic and stressful conditions. Indeed, to combat stressors and survive, pathogens have established various defensive mechanisms, and one of them is the production of membrane vesicles (Mozaheb and Mingeot-Leclercq, 2020).

Figure 5 shows topographic images corresponding to Figure 4. It is clear the expected shape of both bacteria in A, B, D, and E (Chang and Liu, 2018; Salinas-Almaguer et al., 2015; Whitehead et al., 2006). Next, it is shown that PFL produces height irregularities in E. coli (C) and S. aureus (F). The side profile along the length of the bacterium, as shown in the second column, is consistent with such 3D images. The height (Z) profiles from the organisms are seen in the last column, which indicates a non-random behavior. Note the dramatic change in the topography of the cell membrane with PFL treatment along the cross section, which modifies shape and height. It is worth mentioning that the very low concentration found in our MIC experiments could produce damage to the bacteria membranes in the same way as those shown in Figure 4, although not necessarily observable with the sensitivity of our AFM.

As a final test to evaluate bacterial membrane damage caused by PFL, it was performed fluorescence microscopy measurements, using a phosphoethanolamine dye (TR-DHPE), which is employed as an indicator for liquid disordered membranes (Krivic et al., 2022; Skaug et al., 2011; Baumgart et al., 2007; Veatch and Keller, 2005). We evaluated the effect in both culture strains (see Figure 6). Brightfield and fluorescence images are depicted. It is clear that the dye diffuses in both bacteria only when they are treated with PFL, although in the case of S. aureus, the dye penetration is stronger. Even in the absence of PFL (empty DMSO carrier), the dye diffuses a little. This observation is consistent with the fact that S. aureus does not have an additional barrier. Furthermore, this also explains the greater inhibition produced by PFL for the Gram-positive bacteria (see Figures 3A, B).

It is important to note before closing that the research we report here is not the first to explore the promising idea that hydrophobic drugs can be used as a treatment against bacteria. In fact, growth inhibition has been reported to occur for S. epidermidis, S. pyogenes, and S. pneumoniae with 2.5 mg/mL of the local anesthetic bupivacaine (Rosenberg and Renkonen, 1985). Furthermore, it has been shown that nanoemulsions can significantly improve the efficacy of the hydrophobic drugs delivery, as demonstrated with the carbapenem-resistant K. pneumoniae (Tayeb et al., 2022). However, to the best of our knowledge, our study is the first to explore pathogen membrane destabilization upon insertion of hydrophobic drugs. We found that PFL has a remarkable effect in this regard.

To put our findings in perspective, we claim that PFL, and to a lesser extent CBD, could have enormous potential in skin infections where antimicrobial resistance is already a medical problem. In fact, despite recent advances in wound treatment, very few topical therapies have proven effective in promoting wound healing, especially because when recalcitrant bacteria invade wounds, they create a cytotoxic environment that often promotes very serious injuries (Kawano et al., 2020). Furthermore, it is worth mentioning that the method proposed here could be combined with other strategies to improve effectiveness. For example, since essential oils have been found to enhance the antimicrobial activity of drugs and also inhibit the transmission of resistance to other populations (Bueno, 2016), our nanoemulsions could be used in combination with antibiotics or antimicrobial peptides (AMP). Clearly, they could act as adjuvants to increase the uptake of the antibiotic through the bacterial membrane in topical applications. The absorption of PFL through the skin, if it occurs, would not be a problem, since the concentration used here (2.75 mg/mL) is less than the injectable dose normally administered during anesthesia (Schüttler and Ihmsen, 2000). In fact, for a person who weighs approximately 70 kg, the propofol infusion rate is between 2 and 4 mg/mL per minute, for not less than 60 min (time for rapid surgery). This is equivalent to at least 120 mg. Furthermore, recent investigations on the anesthetic effects of propofol through skin absorption found that this drug may have a potential use in clinical practice, especially in pediatric applications (Zhang et al., 2021), where the concentrations used are approximately 10 mg/mL. Concomitantly, this transdermal way of administering propofol gels implies safety for cutaneous/topical applications.

Two limitations of this study are the unexpected poor outcome of CBD and the absence of evidence on the possible medical application of our findings. As we expected the more hydrophobic CBD to give better results but it did not, we would like to provide a positive electrical charge to the nanodroplets and make the release of the drug into the bacterial membranes, which have a negative charge, more efficient. In addition, we would like to test PFL in model animal infections. Such a strategy could provide us with important information about medical applications in humans, to try to contribute to the imminent health emergency that is approaching.