1 Modulating photochemical reactions in Langmuir monolayers of Escherichia coli lipid extract with the binding mechanisms of eosin decyl ester and toluidine blue-O photosensitizers Lucas G. Moreira1, Alexandre M. Almeida Jr.1, Tyler Nield1,2, Sabrina A. Camacho1,3 and Pedro H. B. Aoki1 1 School of Sciences, Humanities and Languages, São Paulo State University (UNESP), Assis, SP, 19806-900, Brazil 2 Faculty of Engineering, University of Victoria, Victoria, BC, V8P 5C2, Canada 3 IFSC, São Carlos Institute of Physics, University of São Paulo (USP), São Carlos, SP 13566-590, Brazil 2 Abstract Photodynamic damage to the cell envelope can inactivate microorganisms and may be applied to combat super-resistance phenomenon, empowered by the indiscriminate use of antibiotics. Efficiency in microbial inactivation is dependent on the incorporation of photosensitizers (PS) into the bacterial membranes to trigger oxidation reactions under illumination. Herein, Langmuir monolayers of Escherichia coli lipid extract were built to determine the binding mechanisms and oxidation outcomes induced by eosin decyl ester (EosDEC) and toluidine blue-O (TBO) PSs. Surface-pressure isotherms of the E. coli monolayers were expanded upon EosDEC and TBO, suggesting incorporation of both PSs. Fourier-transform infrared spectroscopy (FTIR) of Langmuir-Schaefer (LS) films reveled that the EosDEC and TBO binding mechanisms are dominated by electrostatic interactions with the anionic polar groups, with limited penetration into the chains. Light- irradiation reduced the relative area of E. coli monolayer on TBO, indicating an increased loss of material to the subphase owing to the chain cleavage, generated by contact- dependent reactions with excited states of TBO. In contrast, the increased relative area of E. coli monolayers containing EosDEC suggests lipid hydroperoxidation, which is PS contact-independent. Even considering a small chain penetration, the saturated EosDEC may have partitioned towards saturated reach domains, avoiding direct contact with membrane unsaturations. 3 Introduction Bacterial super-resistance is an issue of global public health [1,2], which has been exacerbated by the excessive use and inadequate discarding of antibiotics [3,4]. This resistance is heightened in Gram-negative bacteria such as Escherichia coli owing to their complex cell envelope composed by an internal and external membranes of phospholipids and lipopolysaccharides, respectively [5–9]. Porin proteins are also anchored in the external membrane, working as protective barrier and responsible for regulation of nutrient transport [8,10,11]. The antibiotic activity of many pharmaceuticals depends on the cell envelop internalization throughout the porins, which can be affected by mutations of the microorganism, leading to an acquired resistance [12]. The generation of new antibiotics do not keep up such mutation rate, demanding new strategies to combat the rise of super-resistant bacteria. In this scenario, the photodynamic inactivation of microorganisms (PDIM) [13–18] emerges as a non-invasive therapeutic modality capable to reduce or completely eliminate bacterial strains. The therapy does not rely on the internalization [9,19–21] but on the photodynamic damage caused to the cell envelope, reason why it is considered a promising alternative to avoid the super-resistance phenomena [22–24]. In simple terms, reactive oxygen species (ROS) are generated by excited states of photosensitizers (PS), activated by light irradiation [25–28]. Oxidation is the most likely result of the ROS interaction with cell membranes triggered by contact-independent and contact-dependent reactions. Hydroperoxides are the main products generated by contact-independent reaction due to the singlet oxygen (1O2) attacking the unsaturated lipids [13,29–31]. Additionally, truncated lipid aldehydes are created by contact-dependent reaction between the excited triplet state (3PS*) of the PS and unsaturated lipids or previously formed hydroperoxides, leading to membrane permeabilization [32]. Therefore, PS positioning into the cell envelope plays a central role in the photooxidative processes, 4 which has prompted a search for PSs with the ability to interact with lipid membranes. In this quest, phenothiazine [33,34] and xanthene [35–38] derivatives stand out for being efficiently incorporated into lipid membranes and for the efficiency in PDIM applications. Langmuir, Langmuir-Blodgett (LB) and Langmuir-Schaefer (LS) films are powerful strategies for designing mimetic systems of cell membranes with molecular level control [39]. Langmuir monolayers are considered a nanoarchitectonic approach, confining systems within a two-dimensional plane to drastically reduce translational motion, in which molecular interactions can be investigated [40]. LB and LS are extension techniques to assemble in solid substrates two-dimensional materials, supramolecular structures of organic and inorganic nanoparticles, and biomaterials such as lipids, DNA and enzymes [39,41]. Taken together, these techniques have proven useful to determine mechanisms of action of many biologically relevant materials. Indeed, using simplified models of bacterial membranes based on Langmuir films, we have recently shown that the molecular-level interactions with the hydrophilic toluidine blue-O (TBO) [42] and hydrophobic eosin decyl ester (EosDEC) [43] can modulate the photochemical outcome under illumination. Chain cleavage by contact-dependent reactions was probed in phospholipids monolayers containing EosDEC, which efficiently penetrated the lipid chains owing to its hydrophobic nature. Meanwhile, the hydrophilic TBO was driven to the monolayers by electrostatic interactions with the lipid heads, favoring hydroperoxidation by contact-independent reactions. Chain cleavage could only be observed on monolayers with net negative charge, whose stronger electrostatic interaction favored TBO insertion up to the lipid chains. Although the important contributions on the correlation between the PS’s binding site and photooxidation outcome, the experiments were performed on rather simple Langmuir monolayers, based on neat unsaturated phospholipids. A step forward will be given here unraveling the incorporation and photosensitization mechanisms induced by TBO and EosDEC on Langmuir monolayers 5 of E. coli lipid extract, employed as a complex model of Gram-negative bacterial membranes. The surface pressure isotherms (π-A) allowed evaluation of the TBO and EosDEC membrane incorporation and further photooxidation effects. The molecular- level binding sites were determined by Fourier-transform infrared spectroscopy (FTIR) on LS films deposited from Langmuir monolayers of the E. coli lipid extract containing the PSs. As we shall demonstrate, the complexity of the lipid extract impacts not only on the mechanisms of TBO and EosDEC incorporation but also on the photooxidation effects expected from neat phospholipid monolayers. Experimental Materials Escherichia coli total lipid extract (which is a combination of 16, 17 and 18 carbon length tails containing 57.5% of phosphatidylethanolamines - PE, 15.1% of phosphatidylglycerols - PG, 9.8% of cardiolipin - CLP and 17.6% of unknows substances) was purchased from Avanti Polar Lipids (100500P). The extract comes from an E. coli B (ATCC 11303), containing a variety of fatty acids in their membrane, as detailed in the literature [44–49]. Chloroform (CHCl3, 99,0 ~ 99,4%) and the phenothiazine toluidine blue-O (TBO, log Kp = -0.50, ~80%) were acquired from Sigma- Aldrich (32211 and T3260, respectively). The xanthene eosin decyl ester (EosDEC, log Kp = 1.84, ~90%) was prepared as previously reported [50]. The chemicals were employed as received, without further purification. A Milli-Q system (model Direct-Q® 3UV) provided the ultrapure water (resistivity = 18.2 MΩ·cm) used as aqueous solvent and subphase in the Langmuir and Langmuir-Schaefer (LS) films. 6 Langmuir and Langmuir-Schaefer (LS) films Langmuir film of Escherichia coli lipid extract was produced in a Langmuir trough KSV-NIMA (model KN 2002) by spreading 20 µL of 1.0 mg/mL chloroform solution over the water subphase at room temperature (23oC). The solvent was allowed to evaporate for 10 min before the symmetrical compression of the barriers at 5 mm/min. Surface pressure (π) versus area (cm2/mg of extract) isotherms were built using the Wilhelmy method with a platinum sensor [51]. The same methodology was applied to fabricate mixed Langmuir films of E. coli:EosDEC (10:1 and 5:1 volumetric proportion) and E. coli on TBO (10-5 mol/L) subphase. No precaution was taken to avoid uncontrolled oxidation from the air [52,53], but we ensured the reproducibility by performing experiments in triplicate with surface pressure variation below ± 2 mN/m for a certain area. The E. coli on TBO subphase and E. coli:EosDEC mixed monolayer were kept at 30 mN/m and irradiated using red (IP66−50W) and green (BRIWAX FFG-50 W) LED sources, respectively. The LEDs were positioned 20 cm above the interface, ensuring a homogeneous illumination over the entire trough. Langmuir monolayers of E. coli lipid extract, EosDEC, E. coli mixed with EosDEC (E. coli:EosDEC 10:1 v/v) and E. coli on TBO solution 10-5 mol/L (E. coli + TBO subphase) were transferred to solid substrates by the Langmuir-Schaefer (LS) protocol [54]. The surface pressure of the Langmuir monolayers was kept constant at 30 mN/m while the surface of the films was touched by a solid substrate horizontally approached. The substrate was then lifted-off and LS multilayers were deposited by repeating this procedure. The growth of the LS films on quartz substrate was monitored by UV-Vis absorption spectroscopy, from 190 to 1100 nm, using an Agilent spectrometer (model Cary 60). Fourier-transform infrared spectroscopy (FTIR) were performed for LS films on Ge substrates using a Bruker spectrometer (model Alpha II) with spectral resolution of 4 cm-1 and 256 scans. The FTIR spectra were taken in transmission mode. A schematic 7 representation of the Langmuir and LS film fabrication as well as the main performed characterization techniques is given in Figure 1a. The molecular structures of EosDEC and TBO are presented in Figure 1b. Figure 1. (a) Langmuir and LS film scheme of E. coli lipid extract and the performed characterization techniques. (i) The chloroform solution is spread at the interface of air/aqueous subphase and the chloroform is allowed to evaporate for 10 minutes; (ii) symmetrical compression until 30 mN/m of pressure; (iii) Ge substrate touches horizontally the interface (iv) transferring the monolayer from the interface to the substrate, completing the LS deposition. Steps (iii) and (iv) are repeated several times until the substrate reaches 60 deposited layers. During the deposition, the growth of LS film is characterized by UV-Vis absorption spectroscopy and after the 60 depositions by FTIR. (b) EosDEC and TBO molecular structures. Results and discussion EosDEC and TBO incorporation into E. coli lipid extract monolayers The π-A isotherms of E. coli:EosDEC mixed monolayers (10:1 and 5:1, v/v) on ultrapure water and E. coli monolayer on TBO solution (10-5 mol/L) are displayed in Figures 2a and 2b, respectively. E. coli lipid extract is surface active and presents π-A isotherm with collapse pressure of ca. 45 mN/m. The ca. 130 cm2 of extrapolated area from 30 to 0 mN/m of pressure (or 6.5 x 10³ cm²/mg of extract) is comparable to the ca. 136 cm2 reported by Sandrino et al. [55]. Surface activity was also observed for neat EosDEC (Figure S1) while forming highly packed monolayers, as previously reported [43]. The increase of EosDEC volume in E. coli:EosDEC mixed film shifted the π-A 8 isotherms to larger areas owing to the increased insertion of EosDEC into the monolayers. The displacements in relative area ([(𝐴 − 𝐴𝑜)/𝐴𝑜]) of E. coli monolayers upon PS incorporation is summarized in Table 1. A comparison with previous findings [43] is not straightforward since the number of E. coli lipid molecules cannot be estimated and neither the molar ratio with EosDEC. Nevertheless, incorporation of similar amounts of EosDEC significantly expanded E. coli more than neat 1,2-dioleoyl-sn-glycero-3- phosphoethanolamine (DOPE) monolayers, suggesting that phosphatidylethanolamines (PE) may not govern the interactions with the lipid extract. Incorporation of the water soluble TBO into the E. coli lipid film (Figure 2b) is also suggested by the area expansion of the π-A isotherms. In contrast to EosDEC, the relative area of E. coli (11.9%) is comparable to the observed for neat DOPE monolayers on 10-5 mol/L TBO subphase (13%) [42], reveling a major role of PE in the mechanisms of interactions with the cationic TBO. Additional data on the in-plane elasticity of the monolayers are provided in Figure S2. Figure 2. Surface pressure (mN/m) versus area (cm2/mg of lipid extract) isotherms of (a) E. coli:EosDEC (10:1 and 5:1 v/v) mixed monolayers and (b) E. coli monolayer on TBO solution at 10-5 mol/L. The π-A isotherm of neat E. coli lipid extract monolayer is given as reference. 9 Table 1. Relative shifts in area per E. coli lipid extract [( 𝐴 − 𝐴𝑜 𝐴𝑜 ) 𝑥 100)]. A0 and A are the extrapolated areas at 30 mN/m for the π-A isotherms of neat E. coli and E. coli containing EosDEC or TBO, respectively. Relative shift E. coli:EosDEC E. coli + TBO subphase 10:1 (v/v) 12.5 % ± 4.7 % - 5:1 (v/v) 33.9 % ± 4.1 % - 10-5 mol/L - 11.9 % ± 4.5 % The molecular interactions of EosDEC and TBO in E. coli lipid film were accessed by Fourier-transform infrared spectroscopy (FTIR) and controlled by UV-Vis absorption spectroscopy (Figure S3). E. coli:EosDEC (10:1 v/v) and E. coli on TBO (10-5 mol/L) Langmuir monolayers were transferred to solid substrates forming Langmuir-Schaefer (LS) films. It is expected that the interactions established on the Langmuir films are preserved at high extent upon LS deposition [56]. The FTIR spectra of neat E. coli lipid extract, neat EosDEC and E. coli:EosDEC (10:1 v/v) LS films are shown in Figure 3. The assignments of the main vibrational modes are displayed in Table 2 along with the shifts induced by the EosDEC and TBO incorporation. The E. coli polar groups (Figure 3a) were significantly more affected than the aliphatic region (Figure 3b) upon EosDEC incorporation. For instance, the υs(PO2 ─) shifted from 1075 cm-1 to 1079 cm-1 and the υ(C–O–PO2 ─) from 1031 cm-1 to 1027 cm-1, with an increased intensity. The υas(C–O–C) shifted from 1180 cm-1 to 1170 cm-1 and the C–O vibrational mode at 1260 cm-1 had the intensity increased. The nonhydrated υ(C=O) [55] at 1739 cm-1 was displaced to 1734 cm-1 and had the intensity decreased. These modifications reveal that EosDEC incorporation affects the phosphate groups up to the carbonyl region of the E. coli lipid film. The EosDEC bands at 1558 cm-1, 1507 cm-1 and 1457 cm-1, corresponding to C=C and aromatic groups, appear in the E. coli:EosDEC mixed film without displacements, as well as the vibrational mode at 978 cm-1, assigned to the tetraquinonoid form of xanthene 10 dyes [57]. Vibrational modes of amide I and amide II are also observed in the FTIR spectrum of both E. coli and E. coli:EosDEC (10:1 v/v) LS films, supporting the PM- IRRAs data reported by Sandrino et al. [55]. These bands are likely to arise from residual proteins that could be part of the 17.6% of unknown substances present in the E. coli lipid extract. Regarding the alkyl chains (Figure 3b), only the HC=CH stretching was affected, displacing from 3012 cm-1 to 3003 cm-1. The low intensity of HC=CH vibrational mode relative to the υ(CH2) suggests highly organized chains, characteristic of saturated lipids, which is the major (> 70 %) component in the E. coli lipid membrane [44]. In fact, no changes were observed in the vibrational modes of υas(CH2) and υs(CH2), reveling that the ordering state of the aliphatic tails was not affected by EosDEC incorporation [42,43,53,58]. Therefore, the main changes noticed in the LS film of E. coli:EosDEC (10:1 v/v), indicate that the EosDEC molecules are closer to the E. coli polar region, with limited penetration into the tails, as depicted in Figure 3c. It is likely that the presence of saturated lipids may have increased the packing of the E. coli monolayers hampering deep penetration of EosDEC. Such behavior differs from what was previous observed for fully unsaturated monolayers, which have shown to be less packed allowing EosDEC incorporation up to the chain regions [43]. 11 Figure 3. FTIR spectra of the E. coli lipid extract (black), EosDEC (orange) and E. coli:EosDEC 10:1 v/v (yellow) LS films (60 layers) deposited on germanium substrates. The spectra were split into (a) polar and (b) nonpolar region. (c) Schematic illustration of the EosDEC interaction with E. coli monolayers. The extract is composed by unsaturated and saturated phospholipids (57.5% PE, 15.1% PG, 9.8% CLP and 17.6% unknown substances). The negative sign in the polar head of some phospholipids were used to represent the negatively charged phospholipids (PG and CLP). The inset shows the EosDEC approaching phospholipids mainly by the polar heads (1) and in less extent by the tails (2). The phospholipids can have different sizes (R2) and different polar groups (R1). The FTIR spectra of E. coli lipid extract and E. coli + TBO (10-5 mol/L) LS films are shown in Figure 4. The spectrum of TBO powder is provided as reference. The effects of TBO molecules into E. coli monolayers were quite significant on the phosphate groups (Figure 4a). For instance, the υ(C–O–PO2 ─) at 1031 cm-1 and υs(PO2 ─) at 1075 cm-1 shifted to 1025 cm-1 and 1070 cm-1, respectively. In addition, the υas(PO2 ─) was displaced from 1234 cm-1 to 1219 cm-1 and had the intensity increased. It is known that phosphates are susceptible to H-bonding with surrounding water molecules and amine groups of 12 adjacent lipids in the membrane [59]. TBO incorporation appears to have disrupted such H-bonds owing to electrostatic interactions between the cationic amino groups of TBO and the anionic phosphates of E. coli extract, similar to what was previously observed for neat DOPE and DOPG Langmuir films on TBO subphase [42]. More evidence of the strong incorporation of TBO molecules into the E. coli film is the presence of the phenothiazine bands at 1441 cm-1 and 1600 cm-1 [42]. As for the tails (Figure 4b), the υ(CH2) also dominates the spectrum and the only modification is observed in the υ(HC=CH), which shifted from 3012 cm-1 to 3004 cm-1. Although incorporation of TBO in the vicinity of the lipid unsaturations might be suggested, such a slight modification do not afford massive penetration into the chain region, as observed for E. coli:EosDEC (10:1 v/v) LS film and depicted in Figure 4c. Figure 4. FTIR spectra of the E. coli lipid extract (black), and E. coli + TBO 10-5 mol/L (green) LS films (60 layers) deposited on germanium substrates. The spectrum of TBO powder (blue) is also provided as reference. The spectra were split into (a) polar and (b) nonpolar region. (c) Schematic illustration of the TBO interaction with E. coli monolayers. The extract is composed by unsaturated and saturated phospholipids (57.5% PE, 15.1% PG, 9.8% CLP and 17.6% unknown substances). The negative sign in the polar 13 head of some phospholipids were used to represent the negatively charged phospholipids (PG and CLP). The inset shows the TBO approaching phospholipids mainly by the electrostatic attraction between the cationic amine groups and the anionic phosphates, respectively. The phospholipids can have different sizes (R2) and different polar groups (R1). Table 2. Assignment of the main vibrational modes of E. coli, EosDEC, E. coli:EosDEC (10:1 v/v) and E. coli + TBO (10-5 mol/L) LS films. Assignments E. coli total lipid extract EosDEC (cm-1) Water E. coli:EosDEC E. coli + TBO Water υ(HC=CH) 3012 3003 3004 - υas(CH3) 2955 2958 2958 2954 υas(CH2) 2923 2924 2923 2926 υs(CH3) 2873 - - 2874 υs(CH2) 2853 2853 2853 2854 υ(C=O) 1739 1734 1739 1719 Amide I (Protein)a 1690 1699 1692 - Amide II (Protein) 1549 1541 1549 - Phenothiazine group - - 1441 and 1600 - υ(C=C, Aromatic)2 - 1558 - 1553 and 1567 υ(C=C, Aromatic)1 - 1507 - 1504 υ(C6H5) - 1457 - 1455 υs(CO2 -) - 1351 - 1345 υ(C–O) 1260 1260 - 1267 υas(PO2 -) 1234 1231 1219 - υas(C–O–C) 1180 1170 1170 - υs(PO2 -) 1075 1079 1070 - υ(C–O–PO2 -) 1031 1027 1025 - υs(C–O–C) 983 - - - Tetraquinonoid form - 978 - 978 * E. coli:EosDEC = 10:1 (v/v); E. coli + TBO = 10-5 mol/L. a Amide related to the secondary structure. 14 Photoactivation of EosDEC and TBO incorporated into E. coli lipid extract monolayers The photooxidation of E. coli Langmuir monolayers induced by EosDEC (E. coli:EosDEC 10:1 v/v) and TBO (E. coli + TBO 10-5 mol/L) were followed by the surface area evolution at a constant surface pressure of 30 mN/m, as presented in Figure 5a and 5b. The rate of area decrease in nonirradiated monolayers might be related to the uncontrolled oxidation by reactive oxygen species (ROS) present in the air [53], resulting in the loss of material to the subphase. Besides, irradiation did not produce any significant modification on E. coli:EosDEC (5:1 v/v) monolayer, as detailed in Figure S5. A slight increase of 3.9% ± 1.8% in relative area is only observed for E. coli:EosDEC (10:1 v/v) monolayer (Figure 5a). Excited states of EosDEC generate single oxygen (1O2) species that can further react with chain unsaturations and form hydroperoxides [50]. This hydrophilic groups migrate towards the polar moiety of the monolayer [60–62], increasing the area occupied by the lipid molecules, similar to that observed for irradiated monolayers of unsaturated phosphatidylcholines on erythrosin B [63,64] and eosin Y [65]. However, this finding contrasts with the previous reported data on mixed monolayers of DOPE:EosDEC (5:1), DOPG:EosDEC (5:1) and CLP:EosDEC (2:1), whose relative areas decreased 28%, 19% and 17% upon irradiation, respectively [43]. This increased rate of material loss to the subphase was associated with the hydrophobic nature of EosDEC [50] allowing deeper penetration into the monolayers, which has shown to be key for the chain cleavage at the unsaturation site by contact-dependent reactions [32]. Therefore, the complexity of E. coli lipid extract impacted the interaction mechanism with EosDEC, modulating the photochemical outcome under irradiation. In bacterial membranes the phospholipids are segregated into distinct domains that differ in composition, proteo-lipid interaction, and degree of order [66]. Matsumoto et al. [67–69] have explained that saturated phospholipids (more ordered lipid chains than unsaturated phospholipids) present in bacterial membranes tend to be close to each other, 15 forming segregated microdomains. Since EosDEC is attached to saturated carbon chain its partitioning towards the saturated reach domains of E. coli lipid extract is expected owing to its similar degree of chain order. In this case, even considering a small penetration into the chain region of the monolayer, the direct contact with unsaturations is avoided favoring lipid hydroperoxidation. Besides, such small relative area variation is due to the limited amount of unsaturated lipids within the extract [44]. The mechanism of hydroperoxide formation proposed for the contact-independent reactions between 1O2 species, generated from triplet excited states of EosDEC, and unsaturated lipids of E. coli extract are exhibited in the scheme of Figure 5c [32]. Light-irradiation reduced the relative area of E. coli monolayer on TBO (10-5 mol/L) in 2.5% ± 0.5% (Figure 5b), indicating an increased loss of material to the subphase. This is consistent with the 10% decrease in relative area previously reported for irradiated DOPG monolayers on TBO subphase (10-5 mol/L) [42]. The strong attractive interaction with the anionic groups of DOPG allowed deep insertion of the cationic TBO into the monolayer, favoring contact-dependent reactions between excited states of TBO and chain unsaturation or hydroperoxides, resulting in aldehydes formation and membrane permeabilization [32,42]. The FTIR data presented here (Figure 4) also suggest a strong TBO interaction with the anionic phosphates of the E. coli lipid extract, which contains more than 15% of phosphatidylglycerols in the composition. Such interaction is the main driving force allowing the TBO penetration into the monolayers, which may allow the proximity with E. coli lipid unsaturations and further photo-induced permeabilization, similar to neat DOPG monolayers. Figure 5c depicts the proposed mechanism of contact- dependent reactions between triplet excited states of TBO (or radicals) and unsaturated lipids of E. coli extract, leading to the chain scission and formation of aldehydes, favoring membrane permeabilization. 16 Figure 5. Changes in relative area (A/A0) for nonirradiated (black) and irradiated (red) (a) E. coli:EosDEC (10:1 v/v) monolayer and (b) E. coli monolayer on TBO solution (10- 5 mol/L). A0 is the extrapolated area for the E. coli isotherms at 30 mN/m. (c) Proposed mechanisms for contact-independent (type I) and contact-dependent (type II) reactions between E. coli lipid extract monolayer and EosDEC and TBO, respectively. PS ground (S0) and excited triplet (T1) states; ground (3O2) and excited singlet states (1O2) of oxygen; generic radical species (R•); non-oxidized lipid (LH); lipid carbon-centered (L•), peroxyl (LOO•) and alkoxyl radicals (LO•); lipid hydroperoxide (LOOH). Reprint (adapted) with permission from Bacellar et al. [32] Copyright (2018) American Chemistry Society. 17 Conclusions Langmuir monolayers of E. coli lipid extract were built as a rather complex model of Gram-negative bacterial membrane to determine the binding mechanisms and photooxidative effects induced by the irradiation of EosDEC and TBO photosensitizers. The anionic phosphates were affected by the incorporation of both TBO and EosDEC, with stronger effect for TBO owing to attractive electrostatic interactions. On the other hand, the extent of EosDEC interactions reaches the carbonyl region, benefited from its hydrophobic nature. Both PSs presented limited penetration towards the chain region, which were slightly affected upon TBO and EosDEC incorporation. Nevertheless, chains were cleaved as result of the contact-dependent reactions between excited states of TBO and unsaturations or previously formed hydroperoxides, leading to an increased rate of material loss to the subphase under illumination. Contact-independent reactions were triggered by photoactivated EosDEC, resulting in an increased relative area of the monolayer due to the hydroperoxides formed. Surprisingly the hydrophobic nature of EosDEC did not afford strong penetration into the lipid chains, contrasting with the previous data on neat unsaturated lipid monolayers. Even considering a small chain penetration, the saturated tail of EosDEC may have favored the partitioning towards saturated reach domains of E. coli lipid extract, thus avoiding direct contact with membrane unsaturations. Taken together, these results highlight the role played by the molecular interactions and PS binding site to the photochemical outcome. The understand of the molecular interactions between PSs of different lipophilicities and Gram-negative lipid extract may provide control over the photochemical outcome, which is relevant for further developments in photomedicine. Meanwhile, rather efficient photosensitizers might be developed with increased ability of incorporating and partitioning towards domains enriched with unsaturated phospholipids, since membrane damage strongly depended on contact-dependent reactions. 18 Acknowledgments This work was supported by São Paulo Research Foundation (FAPESP 2013/14262- 7, 2018/16713-0, 2018/14692-5 and 2018/22214-6), INEO and National Council for Scientific and Technological Development. We thank Dr. Diogo S. Pellosi, Drª. Bianca M. Estevão and Dr. Wilker Caetano for their assistance with EosDEC. L.G.M. is thankful for his scholarship provided by FAPESP (2018/13021-0). Conflicts of interest There are no conflicts to declare. 19 References [1] M.M. Masadeh, N.M. Mhaidat, K.H. Alzoubi, E.I. Hussein, E.I. Al-Trad, In vitro determination of the antibiotic susceptibility of biofilm-forming Pseudomonas aeruginosa and Staphylococcus aureus: Possible role of proteolytic activity and membrane lipopolysaccharide, Infect. Drug Resist. 6 (2013) 27—32. https://doi.org/10.2147/IDR.S41501. [2] S. Basak, P. Singh, M. Rajurkar, Multidrug resistant and extensively drug resistant bacteria: A study, J. Pathog. 2016 (2016) 5. https://doi.org/10.1155/2016/4065603. [3] R.P.N. Mishra, E. Oviedo-Orta, P. Prachi, R. Rappuoli, F. Bagnoli, Vaccines and antibiotic resistance, Curr. Opin. Microbiol. 15 (2012) 596–602. https://doi.org/10.1016/j.mib.2012.08.002. [4] P. Grenni, V. Ancona, A. Barra Caracciolo, Ecological effects of antibiotics on natural ecosystems: A review, Microchem. J. 136 (2018) 25–39. https://doi.org/10.1016/j.microc.2017.02.006. [5] H. Goldfine, Bacterial membranes and lipid packing theory., J. Lipid Res. 25 (1984) 1501–1507. http://www.jlr.org/content/25/13/1501.abstract. [6] H. Goldfine, Lipids of prokaryotes–structure and distribution, Curr. Top. Membr. Transp. 17 (1982) 1–43. https://doi.org/10.1016/S0070-2161(08)60307-X. [7] J.M. Stokes, C.R. Macnair, B. Ilyas, S. French, J.P. Côté, C. Bouwman, M.A. Farha, A.O. Sieron, C. Whitfield, B.K. Coombes, E.D. Brown, Pentamidine sensitizes Gram-negative pathogens to antibiotics and overcomes acquired colistin resistance, Nat. Microbiol. 2 (2017) 1–21. https://doi.org/10.1038/nmicrobiol.2017.28. [8] T.J. Silhavy, D. Kahne, S. Walker, The bacterial cell envelope, Cold Spring Harb. Perspect. Biol. 2 (2010) 1–17. https://doi.org/10.1101/cshperspect.a00414. [9] E. Alves, M.A.F. Faustino, M.G.P.M.S. Neves, A. Cunha, J. Tome, A. Almeida, An insight on bacterial cellular targets of photodynamic inactivation, Future Med. Chem. 6 (2014) 141–164. https://doi.org/10.4155/fmc.13.211. [10] C. Schwechheimer, M.J. Kuehn, Outer-membrane vesicles from Gram-negative bacteria: Biogenesis and functions, Nat. Rev. Microbiol. 13 (2015) 605–619. https://doi.org/10.1038/nrmicro3525. 20 [11] A.H. Delcour, Outer membrane permeability and antibiotic resistance, Biochim. Biophys. Acta - Proteins Proteomics. 1794 (2009) 808–816. https://doi.org/10.1016/j.bbapap.2008.11.005. [12] U. Choi, C.R. Lee, Distinct roles of outer membrane porins in antibiotic resistance and membrane integrity in Escherichia coli, Front. Microbiol. 10 (2019) 1–9. https://doi.org/10.3389/fmicb.2019.00953. [13] S. Gai, G. Yang, P. Yang, F. He, J. Lin, D. Jin, B. Xing, Recent advances in functional nanomaterials for light–triggered cancer therapy, Nano Today. 19 (2018) 146–187. https://doi.org/10.1016/j.nantod.2018.02.010. [14] K. Kairyte, S. Lapinskas, V. Gudelis, Z. Luksiene, Effective inactivation of food pathogens Listeria monocytogenes and Salmonella enterica by combined treatment of hypericin-based photosensitization and high power pulsed light, J. Appl. Microbiol. 112 (2012) 1144–1151. https://doi.org/10.1111/j.1365- 2672.2012.05296.x. [15] I. Buchovec, V. Lukseviciūtė, R. Kokstaite, D. Labeikyte, L. Kaziukonyte, Z. Luksiene, Inactivation of Gram (−) bacteria Salmonella enterica by chlorophyllin- based photosensitization: Mechanism of action and new strategies to enhance the inactivation efficiency, J. Photochem. Photobiol. B Biol. 172 (2017) 1–10. https://doi.org/10.1016/j.jphotobiol.2017.05.008. [16] M.M. Awad, A. Tovmasyan, J.D. Craik, I. Batinic-Haberle, L.T. Benov, Important cellular targets for antimicrobial photodynamic therapy, Appl. Microbiol. Biotechnol. 100 (2016) 7679–7688. https://doi.org/10.1007/s00253-016-7632-3. [17] N. Moslemi, N. Rouzmeh, F. Shakerinia, A. Bahador, P.S. Azar, M.J. Kharazifard, M. Paknejad, R. Fekrazad, Photodynamic inactivation of Porphyromonas gingivalis utilizing radachlorin and toluidine blue-O as photosensitizers: An in vitro study, J. Lasers Med. Sci. 9 (2018) 107–112. https://doi.org/10.15171/jlms.2018.21. [18] P. Delcanale, C. Hally, S. Nonell, S. Bonardi, C. Viappiani, S. Abbruzzetti, Photodynamic action of: Hypericum perforatum hydrophilic extract against Staphylococcus aureus, Photochem. Photobiol. Sci. 19 (2020) 324–331. https://doi.org/10.1039/c9pp00428a. [19] M.R. Hamblin, T. Hasan, Photodynamic therapy: A new antimicrobial approach to 21 infectious disease?, Photochem. Photobiol. Sci. 3 (2004) 436–450. https://doi.org/10.1039/b311900a. [20] T.N. Demidova, M.R. Hamblin, Photodynamic therapy targeted to pathogens, Int. J. Immunopathol. Pharmacol. 17 (2004) 245–254. https://doi.org/10.1177/039463200401700304. [21] N. Topaloglu, M. Gulsoy, S. Yuksel, Antimicrobial photodynamic therapy of resistant bacterial strains by indocyanine green and 809-nm diode laser, Photomed. Laser Surg. 31 (2013) 155–162. https://doi.org/10.1089/pho.2012.3430. [22] D.C.S. Costa, M.C. Gomes, M.A.F. Faustino, M.G.P.M.S. Neves, Â. Cunha, J.A.S. Cavaleiro, A. Almeida, J.P.C. Tomé, Comparative photodynamic inactivation of antibiotic resistant bacteria by first and second generation cationic photosensitizers, Photochem. Photobiol. Sci. 11 (2012) 1905–1913. https://doi.org/10.1039/c2pp25113b. [23] M.R. Hamblin, H. Abrahamse, Can light-based approaches overcome antimicrobial resistance?, Drug Dev. Res. 80 (2019) 48–67. https://doi.org/10.1002/ddr.21453. [24] A. Almeida, M.A. Faustino, J.P. Tomé, Photodynamic inactivation of bacteria: Finding the effective targets, Future Med. Chem. 7 (2015) 1221–1224. https://doi.org/10.4155/fmc.15.59. [25] K. Kim, C.S. Lee, K. Na, Light-controlled reactive oxygen species (ROS)- producible polymeric micelles with simultaneous drug-release triggering and endo/lysosomal escape, Chem. Commun. 52 (2016) 2839–2842. https://doi.org/10.1039/c5cc09239f. [26] J.D. Knoll, C. Turro, Control and utilization of ruthenium and rhodium metal complex excited states for photoactivated cancer therapy, Coord. Chem. Rev. 282– 283 (2015) 110–126. https://doi.org/10.1016/j.ccr.2014.05.018. [27] S. Estalayo-Adrián, S. Blasco, S.A. Bright, G.J. McManus, G. Orellana, D.C. Williams, J.M. Kelly, T. Gunnlaugsson, Water-soluble amphiphilic ruthenium(ii) polypyridyl complexes as potential light-activated therapeutic agents, Chem. Commun. 56 (2020) 9332–9335. https://doi.org/10.1039/D0CC04397D. [28] J. Tian, J. Zhou, Z. Shen, L. Ding, J.S. Yu, H. Ju, A pH-activatable and aniline- 22 substituted photosensitizer for near-infrared cancer theranostics, Chem. Sci. 6 (2015) 5969–5977. https://doi.org/10.1039/c5sc01721a. [29] B.W. Henderson, T.J. Dougherty, How does photodynamic therapy work?, Photochem. Photobiol. 55 (1992) 145–157. https://doi.org/10.1111/j.1751- 1097.1992.tb04222.x. [30] A.P. Castano, T.N. Demidova, M.R. Hamblin, Mechanisms in photodynamic therapy: Part one - Photosensitizers, photochemistry and cellular localization, Photodiagnosis Photodyn. Ther. 1 (2004) 279–293. https://doi.org/10.1016/S1572- 1000(05)00007-4. [31] J.C. Ahn, P.S. Chung, The activity of G-ROS and the predominant role of type II reaction in the photodynamic therapy using 9-hydroxypheophorbide-α for HeLa cell lines, Gen. Physiol. Biophys. 31 (2012) 343–350. https://doi.org/10.4149/gpb- 2012-040. [32] I.O.L. Bacellar, M.C. Oliveira, L.S. Dantas, E.B. Costa, H.C. Junqueira, W.K. Martins, A.M. Durantini, G. Cosa, P. Di Mascio, M. Wainwright, R. Miotto, R.M. Cordeiro, S. Miyamoto, M.S. Baptista, Photosensitized membrane permeabilization requires contact-dependent reactions between photosensitizer and lipids, J. Am. Chem. Soc. 140 (2018) 9606–9615. https://doi.org/10.1021/jacs.8b05014. [33] R. Fekrazad, H. Zare, S.M.S. Vand, Photodynamic therapy effect on cell growth inhibition induced by Radachlorin and toluidine blue O on Staphylococcus aureus and Escherichia coli: An in vitro study, Photodiagnosis Photodyn. Ther. 15 (2016) 213–217. https://doi.org/10.1016/j.pdpdt.2016.07.001. [34] H. Liang, J. Xu, Y. Liu, J. Zhang, W. Peng, J. Yan, Z. Li, Q. Li, Optimization of hydrogel containing toluidine blue-O for photodynamic therapy by response surface methodology, J. Photochem. Photobiol. B Biol. 173 (2017) 389–396. https://doi.org/10.1016/j.jphotobiol.2017.06.019. [35] D.P. Valenzeno, J.P. Pooler, Cell membrane photomodification: Relative effectiveness of halogenated fluoresceins for photohemolosysis, Photochem. Photobiol. 35 (1982) 343–350. https://doi.org/10.1111/j.1751- 1097.1982.tb02572.x. [36] D.P. Valenzeno, Photohemolytic lesions: Stoichiometry of creation by Phloxine B, 23 Photochem. Photobiol. 40 (1984) 681–688. https://doi.org/10.1111/j.1751- 1097.1984.tb04637.x. [37] D.P. Valenzeno, M. Tarr, Membrane photomodification of cardiac myocytes: Potassium and leakage currents, Photochem. Photobiol. 53 (1991) 195–201. https://doi.org/10.1111/j.1751-1097.1991.tb03923.x. [38] F. Killig, L. Kunz, G. Stark, Photomodification of the electrical properties of the plasma membrane: A comparison between 6 different membrane-active photosensitizers, J. Membr. Biol. 181 (2001) 41–46. https://doi.org/10.1007/s0023200100046. [39] K. Ariga, Don’t forget Langmuir–Blodgett films 2020: Interfacial nanoarchitectonics with molecules, materials, and living objects, Langmuir. 36 (2020) 7158–7180. https://doi.org/10.1021/acs.langmuir.0c01044. [40] K. Ariga, T. Mori, J. Li, Langmuir nanoarchitectonics from basic to frontier, Langmuir. 35 (2019) 3585–3599. https://doi.org/10.1021/acs.langmuir.8b01434. [41] T. Yunoki, Y. Kimura, A. Fujimori, Maintenance properties of enzyme molecule stereostructure at high temperature by adsorption on organo-modified magnetic nanoparticle layer template, Bull. Chem. Soc. Jpn. 92 (2019) 1662–1671. https://doi.org/10.1246/bcsj.20190102. [42] A.M. Almeida, O.N. Oliveira, P.H.B. Aoki, Role of toluidine blue-O binding mechanism for photooxidation in bioinspired bacterial membranes, Langmuir. 35 (2019) 16745–16751. https://doi.org/10.1021/acs.langmuir.9b03045. [43] L.G. Moreira, A.M. Almeida, S.A. Camacho, B.M. Estevão, O.N. Oliveira, P.H.B. Aoki, Chain cleavage of bioinspired bacterial membranes photoinduced by eosin decyl ester, Langmuir. 36 (2020) 9578–9585. https://doi.org/10.1021/acs.langmuir.0c01600. [44] M. Ishinaga, R. Kanamoto, M. Kito, Distribution of phospholipid molecular species in outer and cytoplasmic membranes of Escherichia coli, J. Biochem. 86 (1979) 161–165. https://doi.org/10.1093/oxfordjournals.jbchem.a132497. [45] M. Kito, M. Ishinaga, M. Nishihara, M. Kato, S. Sawada, T. Hata, Metabolism of the phosphatidylglycerol molecular species in Escherichia coli, Eur. J. Biochem. 54 (1975) 55–63. https://doi.org/10.1111/j.1432-1033.1975.tb04113.x. 24 [46] M. Kito, S. Aibara, M. Ishinaga, T. Hata, A particulate phospholipid synthesizing system from Escherichia coli B, Agr. Biol. Chem. 36 (1971) 837–844. [47] M. Kito, S. Aibara, M. Kato, M. Ishinaga, T. Hata, Effect of changes in fatty acid composition of phospholipid species on the β-galactoside transport system of Escherichia coli K-12, BBA - Biomembr. 298 (1973) 69–74. https://doi.org/10.1016/0005-2736(73)90010-2. [48] A. Shigeo, K. Michie, I. Masataka, K. Makoto, Changes in positional distribution of fatty acids in the phospholipids of Escherichia coli after shift-down in temperature, Biochim. Biophys. Acta (BBA)/Lipids Lipid Metab. 270 (1972) 301– 306. https://doi.org/10.1016/0005-2760(72)90194-4. [49] M. Kito, S. Aibara, M. Kato, T. Hata, Differences in fatty acid composition among phosphatidylethanolamine, phosphatidylglycerol and cardiolipin of Escherichia coli, Biochim. Biophys. Acta (BBA)/Lipids Lipid Metab. 260 (1972) 475–478. https://doi.org/10.1016/0005-2760(72)90062-8. [50] B.M. Estevão, D.S. Pellosi, C.F. De Freitas, D. Vanzin, D.S. Franciscato, W. Caetano, N. Hioka, Interaction of eosin and its ester derivatives with aqueous biomimetic micelles: Evaluation of photodynamic potentialities, J. Photochem. Photobiol. A Chem. 287 (2014) 30–39. https://doi.org/10.1016/j.jphotochem.2014.04.015. [51] P.B. Welzel, I. Weis, G. Schwarz, Sources of error in Langmuir trough measurements Wilhelmy plate effects and surface curvature, Colloids Surfaces A Physicochem. Eng. Asp. 144 (1998) 229–234. https://doi.org/10.1016/S0927- 7757(98)00649-9. [52] S. Sennato, F. Bordi, C. Cametti, C. Coluzza, A. Desideri, S. Rufini, Evidence of domain formation in cardiolipin−glycerophospholipid mixed monolayers. A thermodynamic and AFM study, J. Phys. Chem. B. 109 (2005) 15950–15957. https://doi.org/10.1021/jp051893q. [53] J.F.D. Liljeblad, V. Bulone, E. Tyrode, M.W. Rutland, C.M. Johnson, Phospholipid monolayers probed by vibrational sum frequency spectroscopy: Instability of unsaturated phospholipids, Biophys. J. 98 (2010) L50–L52. https://doi.org/10.1016/j.bpj.2010.02.009. [54] J. Kurniawan, J.F. Ventrici De Souza, A.T. Dang, G.Y. Liu, T.L. Kuhl, Preparation 25 and characterization of solid-supported lipid bilayers formed by Langmuir- Blodgett deposition: A tutorial, Langmuir. 34 (2018) 15622–15639. https://doi.org/10.1021/acs.langmuir.8b03504. [55] B. Sandrino, J.F.A. De Oliveira, T.M. Nobre, P. Appelt, A. Gupta, M.P. De Araujo, V.M. Rotello, O.N. Oliveira, Challenges in application of Langmuir monolayer studies to determine the mechanisms of bactericidal activity of ruthenium complexes, Langmuir. 33 (2017) 14167–14174. https://doi.org/10.1021/acs.langmuir.7b02247. [56] P.H.B. Aoki, P. Alessio, M.L. Rodríguez-Méndez, J.A. De Saja Saez, C.J.L. Constantino, Taking advantage of electrostatic interactions to grow Langmuir- Blodgett films containing multilayers of the phospholipid dipalmitoylphosphatidylglycerol, Langmuir. 25 (2009) 13062–13070. https://doi.org/10.1021/la901923v. [57] F. Amat-Guerri, M.M.C. López-González, R. Martínez-Utrilla, R. Sastre, Synthesis and spectroscopic properties of new rose bengal and eosin Y derivatives, Dye. Pigment. 12 (1990) 249–272. https://doi.org/10.1016/0143-7208(90)85017- I. [58] I.W. Levin, T.E. Thompson, Y. Barenholz, C. Huang, Two types of hydrocarbon chain interdigitation in sphingomyelin bilayers, Biochemistry. 24 (1985) 6282– 6286. https://doi.org/10.1021/bi00343a036. [59] S.M. Gruner, M.W. Tate, G.L. Kirk, P.T.C. So, D.C. Turner, D.T. Keane, C.P.S. Tilcock, P.R. Cullis, X-ray diffraction study of the polymorphic behavior of N- methylated dioleoylphosphatidylethanolamine, Biochemistry. 27 (1988) 2853– 2866. https://doi.org/10.1021/bi00408a029. [60] G.R. Buettner, The pecking order of free radicals and antioxidants: Lipid peroxidation, α-tocopherol, and ascorbate, Arch. Biochem. Biophys. 300 (1993) 535–543. https://doi.org/https://doi.org/10.1006/abbi.1993.1074. [61] G. van Ginkel, J.M. Muller, F. Siemsen, A.A. van ’t Veld, L.J. Korstanje, M.A.M. van Zandvoort, M. Lou Wratten, A. Sevanian, Impact of oxidized lipids and antioxidants, such as vitamin E and lazaroids, on the structure and dynamics of unsaturated membranes, J. Chem. Soc. Faraday Trans. 88 (1992) 1901–1912. https://doi.org/10.1039/FT9928801901. 26 [62] J. Wong-Ekkabut, Z. Xu, W. Triampo, I.M. Tang, D.P. Tieleman, L. Monticelli, Effect of lipid peroxidation on the properties of lipid bilayers: A molecular dynamics study, Biophys. J. 93 (2007) 4225–4236. https://doi.org/10.1529/biophysj.107.112565. [63] P.H.B. Aoki, L.F.C. Morato, F.J. Pavinatto, T.M. Nobre, C.J.L. Constantino, O.N. Oliveira, Molecular-level modifications induced by photo-oxidation of lipid monolayers interacting with erythrosin, Langmuir. 32 (2016) 3766–3773. https://doi.org/10.1021/acs.langmuir.6b00693. [64] M.J. Bistaffa, M.B. Kobal, P.S.S. Souza, K.A. Toledo, S.A. Camacho, P.H.B. Aoki, Photo-induced necrosis on oropharyngeal carcinoma (HEp-2) cells mediated by the xanthene erythrosine, J. Nanosci. Nanotechnol. 20 (2020) 6180–6190. https://doi.org/10.1166/jnn.2020.18123. [65] L.S.A. Pereira, S.A. Camacho, A.A. Malfatti-Gasperini, K. Jochelavicius, T.M. Nobre, O.N. Oliveira, P.H.B. Aoki, Evidence of photoinduced lipid hydroperoxidation in Langmuir monolayers containing Eosin Y, Colloids Surfaces B Biointerfaces. 171 (2018) 682–689. https://doi.org/https://doi.org/10.1016/j.colsurfb.2018.08.002. [66] S. Vanounou, A.H. Parola, I. Fishov, Phosphatidylethanolamine and phosphatidylglycerol are segregated into different domains in bacterial membrane. A study with pyrene-labelled phospholipids, Mol. Microbiol. 49 (2003) 1067– 1079. https://doi.org/https://doi.org/10.1046/j.1365-2958.2003.03614.x. [67] K. Matsumoto, J. Kusaka, A. Nishibori, H. Hara, Lipid domains in bacterial membranes, Mol. Microbiol. 61 (2006) 1110–1117. https://doi.org/10.1111/j.1365-2958.2006.05317.x. [68] F.R. Maxfield, Plasma membrane microdomains, Curr. Opin. Cell Biol. 14 (2002) 483–487. https://doi.org/10.1016/S0955-0674(02)00351-4. [69] K. Simons, D. Toomre, Lipid rafts and signal transduction, Nat. Rev. Mol. Cell Biol. 1 (2000) 31–39. https://doi.org/10.1038/35036052.