Maintenance of bacterial outer membrane lipid asymmetry: insight into MlaA.

The outer membrane (OM) of Gram-negative bacteria acts as an effective barrier to protect against toxic compounds. By nature, the OM is asymmetric with the highly packed lipopolysaccharide (LPS) at the outer leaflet and glycerophospholipids at the inner leaflet. OM asymmetry is maintained by the Mla system, in which is responsible for the retrograde transport of glycerophospholipids from the OM to the inner membrane. This system is comprised of six Mla proteins, including MlaA, an OM lipoprotein involved in the removal of glycerophospholipids that are mis-localized at the outer leaflet of the OM. Interestingly, MlaA was initially identified - and called VacJ - based on its role in the intracellular spreading of Shigella flexneri.Many open questions remain with respect to the Mla system and the mechanism involved in the translocation of mislocated glycerophospholipids at the outer leaflet of the OM, by MlaA. After summarizing the current knowledge on MlaA, we focus on the impact of mlaA deletion on OM lipid composition and biophysical properties of the OM. How changes in OM lipid composition and biophysical properties can impact the generation of membrane vesicles and membrane permeability is discussed. Finally, we explore whether and how MlaA might be a candidate for improving the activity of antibiotics and as a vaccine candidate.Efforts dedicated to understanding the relationship between the OM lipid composition and the mechanical strength of the bacterial envelope and, in turn, how such properties act against external stress, are needed for the design of new targets or drugs for Gram-negative infections.

The molecular mechanism of Nystatin action is dependent on the membrane biophysical properties and lipid composition.

Nystatin (Nys) is a pore forming broad-spectrum and efficient antifungal drug with significant toxicity in mammalian organisms. In order to develop a non-toxic and more effective Nys formulation, its molecular mechanism of action at the cell membrane needs to be better understood. It is widely accepted that Nys activity and toxicity depend on the presence and type of membrane sterols. Taking advantage of multiple biophysical methodologies, we now show that the formation and stabilization of Nys aqueous pores, which are associated with Nys cytotoxicity, occur in the absence of membrane sterols. Our results suggest that the Nys mechanism of action is driven by the presence of highly ordered membrane domains capable of stabilizing the Nys oligomers. Moreover, Nys pore formation is accompanied by strong Nys-induced membrane reorganization that depends on membrane lipid composition and seems to underlie the Nys cytotoxic effect. Accordingly, in membranes enriched in a gel-phase forming phospholipid, Nys incorporates within the phospholipid-enriched gel domains, where it forms pores able to expand the gel domains. In contrast, in membranes enriched in gel domain forming sphingolipids, Nys-induced pore formation occurs through the destabilization of the gel phase. These results show that the Nys mechanism of action is complex and not only dependent on membrane sterols, and provide further insight into the molecular details governing Nys activity and toxicity.

Surfactins modulate the lateral organization of fluorescent membrane polar lipids: a new tool to study drug:membrane interaction and assessment of the role of cholesterol and drug acyl chain length.

The lipopeptide surfactin exhibits promising antimicrobial activities which are hampered by haemolytic toxicity. Rational design of new surfactin molecules, based on a better understanding of membrane:surfactin interaction, is thus crucial. We here performed bioimaging of lateral membrane lipid heterogeneity in adherent living human red blood cells (RBCs), as a new relevant bioassay, and explored its potential to better understand membrane:surfactin interactions. RBCs show (sub)micrometric membrane domains upon insertion of BODIPY analogs of glucosylceramide (GlcCer), sphingomyelin (SM) and phosphatidylcholine (PC). These domains exhibit increasing sensitivity to cholesterol depletion by methyl-ß-cyclodextrin. At concentrations well below critical micellar concentration, natural cyclic surfactin increased the formation of PC and SM, but not GlcCer, domains, suggesting preferential interaction with lipid assemblies with the highest vulnerability to methyl-ß-cyclodextrin. Surfactin not only reversed disappearance of SM domains upon cholesterol depletion but further increased PC domain abundance over control RBCs, indicating that surfactin can substitute cholesterol to promote micrometric domains. Surfactin sensitized excimer formation from PC and SM domains, suggesting increased lipid recruitment and/or diffusion within domains. Comparison of surfactin congeners differing by geometry, charge and acyl chain length indicated a strong dependence on acyl chain length. Thus, bioimaging of micrometric lipid domains is a visual powerful tool, revealing that intrinsic lipid domain organization, cholesterol abundance and drug acyl chain length are key parameters for membrane:surfactin interaction. Implications for surfactin preferential location in domains or at their boundaries are discussed and may be useful for rational design of better surfactin molecules.

Insight into the outer membrane asymmetry of P. aeruginosa and the role of MlaA in modulating the lipidic composition, mechanical, biophysical, and functional membrane properties of the cell envelope.

In Gram-negative bacteria, the outer membrane (OM) is asymmetric, with lipopolysaccharides (LPS) in the outer leaflet and glycerophospholipids (GPLs) in the inner leaflet. The asymmetry is maintained by the Mla system (MlaA-MlaBCDEF), which contributes to lipid homeostasis by removing mislocalized GPLs from the outer leaflet of the OM. Here, we ascribed how Pseudomonas aeruginosa ATCC 27853 coordinately regulates pathways to provide defense against the threats posed by the deletion of mlaA. Especially, we explored (i) the effects on membrane lipid composition including LPS, GPLs, and lysophospholipids, (ii) the biophysical properties of the OM such as stiffness and fluidity, and (iii) the impact of these changes on permeability, antibiotic susceptibility, and membrane vesicles (MVs) generation. Deletion of mlaA induced an increase in total GPLs and a decrease in LPS level while also triggering alterations in lipid A structures (arabinosylation and palmitoylation), likely to be induced by a two-component system (PhoPQ-PmrAB). Altered lipid composition may serve a physiological purpose in regulating the mechanobiological and functional properties of P. aeruginosa. We demonstrated an increase in cell stiffness without alteration of turgor pressure and inner membrane (IM) fluidity in ∆mlaA. In addition, membrane vesiculation increased without any change in OM/IM permeability. An amphiphilic aminoglycoside derivative (3',6-dinonyl neamine) that targets P. aeruginosa membranes induced an opposite effect on ∆mlaA strain with a trend toward a return to the situation observed for the WT strain. Efforts dedicated to understanding the crosstalk between the OM lipid composition, and the mechanical behavior of bacterial envelope, is one needed step for designing new targets or new drugs to fight P. aeruginosa infections.IMPORTANCEPseudomonas aeruginosa is a Gram-negative bacterium responsible for severe hospital-acquired infections. The outer membrane (OM) of Gram-negative bacteria acts as an effective barrier against toxic compounds, and therefore, compromising this structure could increase sensitivity to antibiotics. The OM is asymmetric with the highly packed lipopolysaccharide monolayer at the outer leaflet and glycerophospholipids at the inner leaflet. OM asymmetry is maintained by the Mla pathway resulting in the retrograde transport of glycerophospholipids from the OM to the inner membrane. In this study, we show that deleting mlaA, the membrane component of Mla system located at the OM, affects the mechanical and functional properties of P. aeruginosa cell envelope. Our results provide insights into the role of MlaA, involved in the Mla transport pathway in P. aeruginosa.

Deficient Pseudomonas aeruginosa in MlaA/VacJ outer membrane lipoprotein shows decrease in rhamnolipids secretion, motility, and biofilm formation, and increase in fluoroquinolones susceptibility and innate immune response.

Pseudomonas aeruginosa, a Gram-negative bacterium that causes severe hospital acquired infections poses threat by its ability for adaptation to various growth modes and environmental conditions and by its intrinsic resistance to antibiotics. The latter is mainly due to the outer membrane (OM) asymmetry which is maintained by the Mla pathway resulting in the retrograde transport of glycerophospholipids from the OM to the inner membrane. It comprises six Mla proteins, including MlaA, an OM lipoprotein involved in the removal of glycerophospholipids mislocalized at the outer leaflet of OM. To investigate the role of P. aeruginosa OM asymmetry especially MlaA, this study investigated the effect of mlaA deletion on (i) the susceptibility to antibiotics, (ii) the secretion of virulence factors, the motility, biofilm formation, and (iii) the inflammatory response. mlaA deletion in P. aeruginosa ATCC27853 results in phenotypic changes including, an increase in fluoroquinolones susceptibility and in PQS (Pseudomonas Quinolone Signal) and TNF-α release and a decrease in rhamnolipids secretion, motility and biofilm formation. Investigating how the mlaA knockout impacts on antibiotic susceptibility, bacterial virulence and innate immune response will help to elucidate the biological significance of the Mla system and contribute to the understanding of MlaA in P. aeruginosa OM asymmetry.

Decrease of elastic moduli of DOPC bilayers induced by a macrolide antibiotic, azithromycin.

The elastic properties of membrane bilayers are key parameters that control its deformation and can be affected by pharmacological agents. Our previous atomic force microscopy studies revealed that the macrolide antibiotic, azithromycin, leads to erosion of DPPC domains in a fluid DOPC matrix [A. Berquand, M. P. Mingeot-Leclercq, Y. F. Dufrene, Real-time imaging of drug-membrane interactions by atomic force microscopy, Biochim. Biophys. Acta 1664 (2004) 198-205.]. Since this observation could be due to an effect on DOPC cohesion, we investigated the effect of azithromycin on elastic properties of DOPC giant unilamellar vesicles (GUVs). Microcinematographic and morphometric analyses revealed that azithromycin addition enhanced lipid membranes fluctuations, leading to eventual disruption of the largest GUVs. These effects were related to change of elastic moduli of DOPC, quantified by the micropipette aspiration technique. Azithromycin decreased both the bending modulus (k(c), from 23.1+/-3.5 to 10.6+/-4.5 k(B)T) and the apparent area compressibility modulus (K(app), from 176+/-35 to 113+/-25 mN/m). These data suggested that insertion of azithromycin into the DOPC bilayer reduced the requirement level of both the energy for thermal fluctuations and the stress to stretch the bilayer. Computer modeling of azithromycin interaction with DOPC bilayer, based on minimal energy, independently predicted that azithromycin (i) inserts at the interface of phospholipid bilayers, (ii) decreases the energy of interaction between DOPC molecules, and (iii) increases the mean surface occupied by each phospholipid molecule. We conclude that azithromycin inserts into the DOPC lipid bilayer, so as to decrease its cohesion and to facilitate the merging of DPPC into the DOPC fluid matrix, as observed by atomic force microscopy. These investigations, based on three complementary approaches, provide the first biophysical evidence for the ability of an amphiphilic antibiotic to alter lipid elastic moduli. This may be an important determinant for drug: lipid interactions and cellular pharmacology.

Contribution of plasma membrane lipid domains to red blood cell (re)shaping.

Although lipid domains have been evidenced in several living cell plasma membranes, their roles remain largely unclear. We here investigated whether they could contribute to function-associated cell (re)shaping. To address this question, we used erythrocytes as cellular model since they (i) exhibit a specific biconcave shape, allowing for reversible deformation in blood circulation, which is lost by membrane vesiculation upon aging; and (ii) display at their outer plasma membrane leaflet two types of submicrometric domains differently enriched in cholesterol and sphingomyelin. We here reveal the specific association of cholesterol- and sphingomyelin-enriched domains with distinct curvature areas of the erythrocyte biconcave membrane. Upon erythrocyte deformation, cholesterol-enriched domains gathered in high curvature areas. In contrast, sphingomyelin-enriched domains increased in abundance upon calcium efflux during shape restoration. Upon erythrocyte storage at 4 °C (to mimick aging), lipid domains appeared as specific vesiculation sites. Altogether, our data indicate that lipid domains could contribute to erythrocyte function-associated (re)shaping.

Effect of the antibiotic azithromycin on thermotropic behavior of DOPC or DPPC bilayers.

Azithromycin is a macrolide antibiotic known to bind to lipids and to affect endocytosis probably by interacting with lipid membranes [Tyteca, D., Schanck, A., Dufrene, Y.F., Deleu, M., Courtoy, P.J., Tulkens, P.M., Mingeot-Leclercq, M.P., 2003. The macrolide antibiotic azithromycin interacts with lipids and affects membrane organization and fluidity: studies on Langmuir-Blodgett monolayers, liposomes and J774 macrophages. J. Membr. Biol. 192, 203-215]. In this work, we investigate the effect of azithromycin on lipid model membranes made of 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) or 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC). Thermal transitions of both lipids in contact with azithromycin are studied by (31)P NMR and DSC on multilamellar vesicles. Concerning the DPPC, azithromycin induces a suppression of the pretransition whereas a phase separation between the DOPC and the antibiotic is observed. For both lipids, the enthalpy associated with the phase transition is strongly decreased with azithromycin. Such effects may be due to an increase of the available space between hydrophobic chains after insertion of azithromycin in lipids. The findings provide a molecular insight of the phase merging of DPPC gel in DOPC fluid matrix induced by azithromycin [Berquand, A., Mingeot-Leclercq, M.P., Dufrene, Y.F., 2004. Real-time imaging of drug-membrane interactions by atomic force microscopy. Biochim. Biophys. Acta 1664, 198-205] and could help to a better understanding of azithromycin-cell interaction.

Real-time imaging of drug-membrane interactions by atomic force microscopy.

Understanding drug-biomembrane interactions at high resolution is a key issue in current biophysical and pharmaceutical research. Here we used real-time atomic force microscopy (AFM) imaging to visualize the interaction of the antibiotic azithromycin with lipid domains in model biomembranes. Various supported lipid bilayers were prepared by fusion of unilamellar vesicles on mica and imaged in buffer solution. Phase-separation was observed in the form of domains made of dipalmitoylphosphatidylcholine (DPPC), sphingomyelin (SM), or SM/cholesterol (SM/Chl) surrounded by a fluid matrix of dioleoylphosphatidylcholine (DOPC). Time-lapse images collected following addition of 1 mM azithromycin revealed progressive erosion and disappearance of DPPC gel domains within 60 min. We attribute this effect to the disruption of the tight molecular packing of the DPPC molecules by the drug, in agreement with earlier biophysical experiments. By contrast, SM and SM-Chl domains were not modified by azithromycin. We suggest that the higher membrane stability of SM-containing domains results from stronger intermolecular interactions between SM molecules. This work provides direct evidence that the perturbation of lipid domains by azithromycin strongly depends on the lipid nature and opens the door for developing new applications in membrane biophysics and pharmacology.

Induction of apoptosis in human promyelocytic leukemia cells by a natural trachylobane diterpene.

BACKGROUND: Trachylobane diterpenes are secondary metabolites, quite rare in nature, and their bioactivities are poorly understood. Recently, we have described the cytotoxic activity of ent-trachyloban-3beta-ol isolated from the leaves of Croton zambesicus, a plant used in African folk medicine. MATERIALS AND METHODS: Cell viability on several cell lines, cell morphology, DNA laddering, annexin Vand caspase-3 activation experiments were undertaken in order to analyse the cytotoxicty of trachylobane diterpene and to determine if this compound is able to induce apoptosis. RESULTS: ent-Trachyloban-3beta-ol exerts a dose-dependent cytotoxic effect, which varies between cell lines. Induction of apoptosis in HL-60 cells could be detected at a concentration of 50 microM after 24-h treatment. CONCLUSION: We show here, for the first time, that a trachylobane diterpene is able to induce apoptosis in human promyelocytic leukemia cells via caspase-3 activation in a concentration-dependent manner.

Azithromycin, a lysosomotropic antibiotic, impairs fluid-phase pinocytosis in cultured fibroblasts.

The dicationic macrolide antibiotic azithromycin inhibits the uptake of horseradish peroxidase (HRP) by fluid-phase pinocytosis in fibroblasts in a time- and concentration-dependent fashion without affecting its decay (regurgitation and/or degradation). The azithromycin effect is additive to that of nocodazole, known to impair endocytic uptake and transport of solutes along the endocytic pathway. Cytochemistry (light and electron microscopy) shows a major reduction by azithromycin in the number of HRP-labeled endocytic vesicles at 5 min (endosomes) and 2 h (lysosomes). Within 3 h of exposure, azithromycin also causes the appearance of large and light-lucentlelectron-lucent vacuoles, most of which can be labeled by lucifer yellow when this tracer is added to culture prior to azithromycin exposure. Three days of treatment with azithromycin result in the accumulation of very large vesicles filled with pleiomorphic content, consistent with phospholipidosis. These vesicles are accessible to fluorescein-labeled bovine serum albumin (FITC-BSA) and intensively stained with filipin, indicating a mixed storage with cholesterol. The impairment of HRP pinocytosis directly correlates with the amount of azithromycin accumulated by the cells, but not with the phospholipidosis induced by the drug. The proton ionophore monensin, which completely suppresses azithromycin accumulation, also prevents inhibition of HRP uptake. Erythromycylamine, another dicationic macrolide, also inhibits HRP pinocytosis in direct correlation with its cellular accumulation and is as potent as azithromycin at equimolar cellular concentrations. We suggest that dicationic macrolides inhibit fluid-phase pinocytosis by impairing the formation of pinocytic vacuoles and endosomes.

New derivatives of kanamycin B obtained by modifications and substitutions in position 6". 1. Synthesis and microbiological evaluation.

The clinical use of the potent, wide-spectrum aminoglycoside antibiotics is limited by oto- and nephrotoxicities. The latter is related to the binding of these polycationic drugs to negatively charged phospholipids and to the subsequent inhibition of lysosomal phospholipases. In order to explore the influence of a modification of the hydrophobic/hydrophilic balance at a specific site of an aminoglycoside, kanamycin B has been chemically modified in position 6" by substitution of the hydroxyl group with a halogen atom (or a pseudohalogen group), or an amino, an amido, a thioalkyl, or an alkoxy group, each series containing increasingly bulkier chains. Examination of the antibacterial activity of the synthesized compounds revealed a negative correlation between the size of the 6"-substituent and the antibacterial activity against kanamycin B sensitive Gram-positive and -negative organisms. Only derivatives with small substituents in position 6", namely chloro, bromo, azido, amino, methylcarbamido, acetamido, methylthio, methylsulfinyl, O-methyl, O-ethyl, and O-isopropyl, showed acceptable activity (geometric mean of minimum inhibitory concentrations for Gram-negative strains less than or equal to 2.5 mg/L; value for kanamycin B, 0.5 mg/L). In vitro toxicological evaluation of all derivatives and computer-aided conformational analysis of selected compounds inserted in a phosphatidylinositol monolayer are presented in the following paper in this issue.

New derivatives of kanamycin B obtained by combined modifications in positions 1 and 6". Synthesis, microbiological properties, and in vitro and computer-aided toxicological evaluation.

Substitution of the C-1 atom in the 2-deoxystreptamine moiety of gentamicin C2, a broad-spectrum aminoglycoside antibiotic, by an axial hydroxymethyl group has been reported to confer protection against most clinically important bacterial enzymes inactivating aminoglycosides, while simultaneously reducing the nephrotoxic potential of this drug. We report here on a similar modification of kanamycin B. Microbiological evaluation, however, revealed no useful protection, as established by the almost complete lack of activity of 1-C-(hydroxymethyl)kanamycin B against an array of organisms producing defined types of aminoglycoside-inactivating enzymes and against which 1-C-(hydroxymethyl)gentamicin C2 and amikacin (1-N-[(S)-2-hydroxy-4-aminobutyryl]kanamycin A) are active. Moreover, toxicological evaluation, based on the in vitro measurement of the drug inhibitory potential toward lysosomal phospholipases, a predictive test of the intrinsic nephrotoxic potential of aminoglycosides, showed not decreased but rather increased toxicity. Comparative conformational analysis of the interactions of the drug with a phosphatidylinositol monolayer explained the lack of protective effect, since no significant change of the mode of insertion of the derivative in this monolayer was detected compared to that of kanamycin B. Combination of a 1-C-(hydroxymethyl) substituent with a 6"-chloro, 6"-acetamido substituent resulted in a partial improvement of the toxicological behavior with no loss of activity for the 6"-chloro and the 6"-azido derivatives, but not to the extent of obtaining better derivatives than kanamycin B itself. We, therefore, suggest that the advantages of an axial hydroxymethyl substituent at C-1 are probably restricted to the gentamicin family and do not extend to kanamycins. It might be concluded that the structural differences between gentamicins and kanamycins play an important, still undescribed role both in their effective recognition by aminoglycoside-inactivating enzymes, which are responsible for most of the clinically important cases of resistance to aminoglycosides, and also in the interactions with phospholipids, which in turn cause nephrotoxicity.

New derivatives of kanamycin B obtained by modifications and substitutions in position 6''. 2. In vitro and computer-aided toxicological evaluation with respect to interactions with phosphatidylinositol.

In a companion paper (previous paper in this issue), we report on the synthesis and microbiological evaluation of new derivatives of the aminoglycoside antibiotic kanamycin B carrying substitutions in 6" (halogeno, or amino, amido, thioalkyl, and alkoxy groups, each series with increasingly bulkier chains). These modifications were intended to potentially modulate the interactions of kanamycin B with phospholipids since these are related to inhibition of lysosomal phospholipase activities and lysosomal phospholipidosis, an early and predictive index of the nephrotoxic potential of aminoglycosides. The new derivatives were therefore examined for inhibitory potency in vitro toward lysosomal phospholipase A1 acting on phosphatidylcholine included in negatively charged liposomes. No simple correlation was observed between the nature or the size of the 6''-substituent and the inhibitory potencies of the corresponding derivatives, although certain groups (diethylamino, isopropylthio) caused a significant increase in inhibitory potency, whereas an N-acetyl-N-methylamino substituent had the opposite effect. 6''-Deoxy-6''-chlorokanamycin B, however, was the only derivative showing both a decrease (albeit limited) of inhibitory potency toward phospholipase A1 associated with the maintenance of a satisfactory microbiological activity (actually equal or slightly better than that of kanamycin B). Computer-aided conformational analysis showed that this chloro substituent did not allow the molecule to insert itself very differently compared to kanamycin B or 6''-deoxykanamycin B in a monolayer of phosphatidylinositol, all three drugs adopting an orientation largely parallel to the hydrophobic-hydrophilic interface and being largely "embedded" in the bilayer at that level. In contrast, the N-acetyl-N-methylamino and isopropylthio substituents caused the corresponding derivatives to adopt an orientation largely perpendicular to the interface, because of the attraction of this substituent, and therefore of the 3''-amino sugar moiety of kanamycin B into the hydrophobic domain of the monolayer, whereas the opposite part of the drug (2',6'-diamino sugar) protruded into the aqueous phase. No simple correlation, however, could be drawn between these changes of conformation and the relative inhibitory potencies of the derivatives.

Synthesis and antimicrobial and toxicological studies of amino acid and peptide derivatives of kanamycin A and netilmicin.

Amino acid and peptide derivatives of aminoglycosides have been obtained by substitution of the 1-N or 6'-N amino functions of kanamycin A and netilmicin via the temporary complexation of vicinal and nonvicinal amino and hydroxy functions by copper ion [1-N kanamycin A derivatives: L-Ala (6a), D-Ala (6b), Gly (6c), L-Asp (6d), L-Ala-L-Ala (6e). 6'-N kanamycin A derivatives: L-Ala (3a), D-Ala (3b), Gly (3c), L-Ala-L-Ala (3e), L-Leu (3f). 6'-N netilmicin derivatives: L-Ala (9a), D-Ala (9b), Gly (9c), L-Asp (9d), L-Ala-L-Ala (9e)]. Characterization was made by FAB-MS, IR, 1H-NMR, and 13C-NMR. All derivatives were essentially inactive. The nephrotoxic potential of the derivatives obtained in sufficient quantities (3b,e and 9a-e) was assessed by measuring their inhibitory potential toward the activity of lysosomal phospholipase A1 acting on phosphatidylcholine embedded in negatively-charged membranes. One compound, 6'-N-L-Ala-netilmicin (9a), showed a 2-fold decrease of inhibitory potency compared to its parent drug. A conformational analysis revealed that it adopts two equally probable conformations and orientations when interacting with phosphatidylinositol. The first in which the drug lies parallel to the hydrophobic-hydrophilic interface, is similar to that of netilmicin. The second, in which the drug inserts itself in the bilayer across the hydrophilic/hydrophobic interface, is similar to that described for streptomycin, an almost non-nephrotoxic aminoglycoside.

Effect of acidic phospholipids on the activity of lysosomal phospholipases and on their inhibition by aminoglycoside antibiotics--I. Biochemical analysis.

Aminoglycoside antibiotics accumulate in lysosomes of kidney and cultured cells and cause an impairment of phospholipid catabolism which is considered to be an early and significant step in the development of their toxicity. Using liposomes, wer previously demonstrated that the activity of lysosomal phospholipases A1 and A2 towards phosphatidylcholine was markedly enhanced by the inclusion of phosphatidylinositol in the bilayer, and that gentamicin impaired this activity by binding to phosphatidylinositol. Since gentamicin-induced inhibition was inversely related to the amount of phosphatidylinositol included in the liposomes, we proposed that gentamicin impairs activity of phospholipases by decreasing the quantity of available negative charges carried by the bilayer surface (Mingeot-Leclercq et al., Biochem Pharmacol 37: 591-599, 1988). We now extend these observations to phosphatidylserine and phosphatidic acid, and compare the inhibition caused by gentamicin, amikacin and streptomycin towards lysosomal phospholipases on the hydrolysis of phosphatidylcholine in the presence of each of these acidic phospholipids. Inclusion of phosphatidic acid in liposomes, and, to a lesser extent, phosphatidylserine, caused a larger increase in phospholipases activity than phosphatidylinositol. In parallel, the three aminoglycosides tested were found less inhibitory towards phospholipases activity measured on phosphatidic acid-or phosphatidylserine-containing liposomes than was previously observed with phosphatidylinositol, even though equilibrium dialysis experiments failed to demonstrate significant difference in binding parameters of the drug towards each of these liposomes populations. Yet, as for phosphatidylinositol-containing liposomes, the inhibition was inversely related to the amount of phosphatidic acid or phosphatidylserine included in the bilayer and the inhibitory potency of the three drugs was consistently gentamicin greater than amikacin greater than streptomycin with the three types of negatively-charged liposomes used. We conclude that impairment of lysosomal phospholipases activity towards phosphatidylcholine included in negatively-charged membranes by aminoglycoside antibiotics is dependent upon drug binding to the bilayer, but that it is modulated by the nature of the acidic phospholipid that binds the drug as well as by that of the drug itself. A companion paper (Mingeot-Leclercq et al., Biochem Pharmacol 40: 499-506, 1990) will examine by computer-aided conformational analysis the parameters (drug-phospholipid energy of interaction, position of the drug in a monolayer and its accessibility to the aqueous phase) which may be important for these effects.

Effect of acidic phospholipids on the activity of lysosomal phospholipases and on their inhibition induced by aminoglycoside antibiotics--II. Conformational analysis.

In a companion paper (Mingeot-Leclercq et al. Biochem Pharmacol 40: 489-497, 1990), we showed that the inhibitory potency of gentamicin on the activity of lysosomal phospholipases, measured towards phosphatidylcholine included in negatively-charged liposomes, is markedly influenced by the nature of the acidic phospholipid used (phosphatidylinositol, phosphatidylserine, phosphatidic acid), whereas the binding of the drug to the three types of liposomes is similar. This result challenged previous conclusions pointing to a key role exerted by drug binding to phospholipid membranes and presumably charge neutralization, for phospholipases inhibition (Carlier et al. Antimicrob Agents Chemother, 23: 440-449, 1983; Mingeot-Leclercq et al., Biochem Pharmacol 37:591-599, 1988). Conformational analysis of mixed monolayers of gentamicin and each of the three acid phospholipids shows that gentamicin systematically adopts an orientation largely parallel to the hydrophobic-hydrophilic interface, but that (i) the energies of interaction are largely different (phosphatidylinositol greater than phosphatidylserine greater than phosphatidic acid), and (ii) the apparent accessibility of the bound drug to water varies in an inverse relation with the energies of interaction. Amikacin, a semisynthetic derivative of kanamycin A with a lower inhibitory potential towards phospholipases than gentamicin in the three types of liposomes used, also showed similar differences in energies of interaction and accessibility to water, but constantly exhibited an orientation perpendicular to the hydrophobic-hydrophilic interface. We conclude that impairment of lysosomal phospholipase activities towards phosphatidylcholine included in negatively-charged membranes by aminoglycoside antibiotics is indeed dependent upon drug binding to the bilayer, but is also modulated by (i) the nature of the acidic phospholipid, which influences the energy of interaction and the accessibility of the drug with respect to the hydrophilic phase, and (ii) the orientation of the drug, which it itself related to its chemical structure. Inasmuch as phospholipases inhibition is related to aminoglycoside nephrotoxicity, these findings may help in better defining the molecular determinants and mechanisms responsible for this adverse effect.

Accessibility of aminoglycosides, isolated and in interaction with phosphatidylinositol, to water. A conformational analysis using the concept of molecular hydrophobicity potential.

The mode of interaction between aminoglycosides and negatively charged phospholipids plays a critical role in the inhibition of lysosomal phospholipases induced by these antibiotics and therefore in their nephrotoxicity. Previous works suggested that accessibility of the drug interacting with phospholipids to water could be crucial in this respect. We have used the concept of molecular hydrophobicity potential described by Brasseur [J Med Chem 266: 16120-16127, 1991] to visualize the hydrophobic and hydrophilic envelopes around aminoglycosides assembled with phosphatidylinositol molecules, and to obtain a three-dimensional representation of the complex formed. Using a series of different aminoglycosides, we showed that molecules with a lower inhibitory potential (gentamicin B, amikacin and isepamicin) are surrounded by both hydrophobic and hydrophilic envelopes whereas aminoglycosides which are more inhibitory are enveloped primarily by either hydrophilic (kanamycin A or B) or hydrophobic (gentamicin C1a) envelopes. This approach, which is here for the first time applied to the study of drug-lipid complexes, could help in the better understanding of the molecular mechanism of lysosomal phospholipase inhibition induced by aminoglycosides.

Biochemical mechanism of aminoglycoside-induced inhibition of phosphatidylcholine hydrolysis by lysosomal phospholipases.

Aminoglycosides such as gentamicin are hydrophilic, polycationic drugs which bind to negatively-charged phospholipid bilayers, inhibit the activities of the lysosomal enzymes involved in the degradation of the major phospholipids and cause, in kidney in vivo or in cultured cells, a lysosomal phospholipidosis. In the present study, we show that the hydrolysis of phosphatidylcholine induced in liposomes by lysosomal extracts at pH 5.4 in vitro is critically dependent on the negative charges carried by the bilayer. This hydrolysis, which is predominantly carried on by phospholipases A1 and A2, markedly increases when the phosphatidylinositol content is raised from 10 to 30% of the total phospholipids, i.e. in a range found in natural membranes. Addition of gentamicin decreases the activity of these enzymes in a non-competitive fashion, but the effect is inversely proportional to the amount of phosphatidylinositol present in the bilayer. Gentamicin and bis(beta-diethylaminoethylether)hexestrol (DEH), a cationic amphiphile which also binds to phospholipid bilayers, are equipotent inhibitors when added to negatively-charged liposomes at equinormal concentrations. Although direct aminoglycoside-enzyme interactions cannot be excluded, these results strongly suggest that gentamicin impairs the activities of the lysosomal phospholipases towards phosphatidylcholine by decreasing the available negative charges required for optimal activity.

Ultrastructural, physico-chemical and conformational study of the interactions of gentamicin and bis(beta-diethylaminoethylether) hexestrol with negatively-charged phospholipid layers.

Aminoglycoside antibiotics such as gentamicin, which are fully hydrophilic, and cationic amphiphilic drugs such as bis(beta-diethylaminoethylether)hexestrol (DEH), are both known to inhibit lysosomal phospholipases and induce phospholipidosis. This enzymatic inhibition is probably related to the neutralization of the surface negative charges on which the lysosomal phospholipases A1 and A2 are dependent to express fully their activities (Mingeot-Leclerq et al., Biochem Pharmacol 37: 591-599, 1988). Using negatively charged liposomes, we show by 31P NMR spectroscopy that both gentamicin and DEH cause a significant restriction in the phosphate head mobility and, in sonicated vesicles, the appearance of larger bilayer structures. Both DEH and gentamicin increased the apparent size of sonicated negatively charged liposomes (but not of neutral liposomes) as measured by quasi-elastic light scattering spectroscopy. Examination of replicas from freeze-etched samples, however, revealed that gentamicin caused aggregation of liposomes, whereas DEH induced their fusion and the formation of intramembranous roundly shaped structures. Only DEH caused a significant decrease of the fluorescence polarization of 1,6-diphenyl-1,3,5-hexatriene, a fluorescent lipid-soluble probe. In addition, DEH, but not gentamicin, interfered with the bilayer to hexagonal phase transition occurring in dioleoyl- and dielaidoylphosphatidylethanolamine liposomes upon warming, and caused the appearance of an isotropic signal suggestive of the formation of inverted micelles. In computer-aided conformational analysis of the molecules at a simulated air-water interface, gentamicin was shown to display a largely-open crescent shape. When surrounded by phosphatidylinositol molecules, it remained as such at the interface which it locally mis-shaped, establishing close contact with the negatively charged phospho groups. In contrast, DEH could be oriented perpendicularly to the interface, with its two cationic groups associated with the phospho groups, and its phenyl- and diethylethandiyl moieties deeply inserted between and interacting with the aliphatic chains. Thus, although both agents cause lysosomal phospholipases inhibition, the differences in their interactions with negatively-charged bilayers is likely to result in a different organization of the phospholipids accumulated in vivo, which could lead to different toxicities.